subsurface/core/libdivecomputer.cpp

1616 lines
51 KiB
C++
Raw Normal View History

// SPDX-License-Identifier: GPL-2.0
#ifdef __clang__
// Clang has a bug on zero-initialization of C structs.
#pragma clang diagnostic ignored "-Wmissing-field-initializers"
#endif
#include <fcntl.h>
#include <stdio.h>
#include <stdlib.h>
#include <unistd.h>
#include <inttypes.h>
#include <string.h>
use libdivecomputer 'fingerprint' to avoid downloading extra data This opportunistically uses a cache of 'fingerprints' for already downloaded dives. As we download data from a dive computer, we save the fingerprint and dive ID of the most recent dive in a per-divecopmputer fingerprint cache file. The next time we download from that dive computer, we will load the cache file for that dive computer if it exists, verify that we still have the dive that is referenced in that cachefile, and if so use the fingerprint to let libdivecomputer potentially stop downloading dives early. This doesn't much matter for most dive computers, but some (like the Scubapro G2) are not able to download one dive at a time, and need the fingerprint to avoid doing a full dump. That is particularly noticeable over bluetooth, where a full dump can be very slow. NOTE! The fingerprint cache is a separate entity from the dive log itself. Unlike the dive log, it doesn't synchronize over the cloud, so if you download using different clients (say, your phone and your laptop), the fingerprint cache entries are per device. So you may still end up downloading dives you already have, because the fingerprint code basically only works to avoid duplicate downloads on the same installation. Also, note that we only have a cache of one single entry per dive computer and downloader, so if you download dives and then don't save the end result, the fingerprint will now point to a dive that you don't actually have in your dive list. As a result, next time you download, the fingerprint won't match any existing dive, and we'll resort to the old non-optimized behavior. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-04-24 17:37:31 +00:00
#include <sys/types.h>
#include <sys/stat.h>
#include <chrono>
#include "gettext.h"
#include "divelog.h"
#include "divesite.h"
#include "sample.h"
#include "subsurface-float.h"
#include "subsurface-string.h"
#include "format.h"
Assemble the actual Suunto serial number It turns out that the serial number returned by libdivecomputer isn't really the serial number as interpreted by the vendor. Those tend to be strings, but libdivecomputer gives us a 32bit number. Some experimenting showed that for the Suunto devies tested the serial number is encoded in that 32bit number: It so happens that the Suunto serial number strings are strings that have all numbers, but they aren't *one* number. They are four bytes representing two numbers each, and the "23500027" string is actually the four bytes 23 50 00 27 (0x17 0x32 0x00 0x1b). And libdivecomputer has incorrectly parsed those four bytes as one number, not as the encoded serial number string it is. So the value 389152795 is actually hex 0x1732001b, which is 0x17 0x32 0x00 0x1b, which is - 23 50 00 27. This should be done by libdivecomputer, but hey, in the meantime this at least shows the concept. And helps test the XML save/restore code. It depends on the two patches that create the whole "device.c" infrastructure, of course. With this, my dive file ends up having the settings section look like this: <divecomputerid model='Suunto Vyper Air' deviceid='d4629110' serial='01201094' firmware='1.1.22'/> <divecomputerid model='Suunto HelO2' deviceid='995dd566' serial='23500027' firmware='1.0.4'/> where the format of the firmware version is something I guessed at, but it was the obvious choice (again, it's byte-based, I'm ignoring the high byte that is zero for both of my Suuntos). Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
2013-01-10 00:14:21 +00:00
#include "device.h"
#include "dive.h"
#include "errorhelper.h"
#include "event.h"
#include "sha1.h"
#include "subsurface-time.h"
#include <libdivecomputer/version.h>
#include <libdivecomputer/usbhid.h>
#include <libdivecomputer/usb.h>
#include <libdivecomputer/serial.h>
#include <libdivecomputer/irda.h>
#include <libdivecomputer/bluetooth.h>
#include "libdivecomputer.h"
#include "core/version.h"
use libdivecomputer 'fingerprint' to avoid downloading extra data This opportunistically uses a cache of 'fingerprints' for already downloaded dives. As we download data from a dive computer, we save the fingerprint and dive ID of the most recent dive in a per-divecopmputer fingerprint cache file. The next time we download from that dive computer, we will load the cache file for that dive computer if it exists, verify that we still have the dive that is referenced in that cachefile, and if so use the fingerprint to let libdivecomputer potentially stop downloading dives early. This doesn't much matter for most dive computers, but some (like the Scubapro G2) are not able to download one dive at a time, and need the fingerprint to avoid doing a full dump. That is particularly noticeable over bluetooth, where a full dump can be very slow. NOTE! The fingerprint cache is a separate entity from the dive log itself. Unlike the dive log, it doesn't synchronize over the cloud, so if you download using different clients (say, your phone and your laptop), the fingerprint cache entries are per device. So you may still end up downloading dives you already have, because the fingerprint code basically only works to avoid duplicate downloads on the same installation. Also, note that we only have a cache of one single entry per dive computer and downloader, so if you download dives and then don't save the end result, the fingerprint will now point to a dive that you don't actually have in your dive list. As a result, next time you download, the fingerprint won't match any existing dive, and we'll resort to the old non-optimized behavior. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-04-24 17:37:31 +00:00
#include "core/qthelper.h"
#include "core/file.h"
#include <array>
#include <charconv>
std::string dumpfile_name;
std::string logfile_name;
std::string progress_bar_text;
void (*progress_callback)(const std::string &text) = NULL;
double progress_bar_fraction = 0.0;
static int stoptime, stopdepth, ndl, po2, cns, heartbeat, bearing;
static bool in_deco, first_temp_is_air;
static int current_gas_index;
#define INFO(fmt, ...) report_info("INFO: " fmt, ##__VA_ARGS__)
#define ERROR(fmt, ...) report_info("ERROR: " fmt, ##__VA_ARGS__)
device_data_t::device_data_t()
{
}
device_data_t::~device_data_t()
{
if (descriptor)
dc_descriptor_free(descriptor);
}
/*
* Directly taken from libdivecomputer's examples/common.c to improve
* the error messages resulting from libdc's return codes
*/
const char *errmsg (dc_status_t rc)
{
switch (rc) {
case DC_STATUS_SUCCESS:
return "Success";
case DC_STATUS_UNSUPPORTED:
return "Unsupported operation";
case DC_STATUS_INVALIDARGS:
return "Invalid arguments";
case DC_STATUS_NOMEMORY:
return "Out of memory";
case DC_STATUS_NODEVICE:
return "No device found";
case DC_STATUS_NOACCESS:
return "Access denied";
case DC_STATUS_IO:
return "Input/output error";
case DC_STATUS_TIMEOUT:
return "Timeout";
case DC_STATUS_PROTOCOL:
return "Protocol error";
case DC_STATUS_DATAFORMAT:
return "Data format error";
case DC_STATUS_CANCELLED:
return "Cancelled";
default:
return "Unknown error";
}
}
/**
* @brief get_deeper_gasmix Returns the gas mix with the deeper MOD.
* NOTE: Parameters are passed by value in order to use them as local working
* storage.
* Invalid gas mixes are converted to air for the purpose of this operation.
* The gas mix with the lower MOD is taken as the one with the lower O2 content,
* or, if equal, the one with the higher HE content. No actual MOD calculations
* are performed.
* @param a The first gas mix to compare.
* @param b The second gas mix to compare.
* @return The gas mix with the deeper MOD.
*/
static struct gasmix get_deeper_gasmix(struct gasmix a, struct gasmix b)
{
if (same_gasmix(a, gasmix_invalid)) {
a = gasmix_air;
}
if (same_gasmix(b, gasmix_invalid)) {
b = gasmix_air;
}
if (get_o2(a) < get_o2(b)) {
return a;
}
if (get_o2(a) > get_o2(b)) {
return b;
}
return get_he(a) < get_he(b) ? b : a;
}
/**
* @brief parse_gasmixes matches gas mixes with cylinders
* This function retrieves all tanks and gas mixes reported by libdivecomputer
* and attepmts to match them. The matching logic assigns the mixes to the
* tanks in a 1:1 ordering.
* If there are more gas mixes than tanks, additional tanks are created.
* If there are fewer gas mixes than tanks, the remaining tanks are assigned to
* the gas mix with the lowest (deepest) MOD.
* @param devdata The dive computer data.
* @param dive The dive to which these tanks and gas mixes will be assigned.
* @param parser The libdivecomputer parser data.
* @param ngases The number of gas mixes to process.
* @return DC_STATUS_SUCCESS on success, otherwise an error code.
*/
static dc_status_t parse_gasmixes(device_data_t *devdata, struct dive *dive, dc_parser_t *parser, unsigned int ngases)
{
static bool shown_warning = false;
unsigned int i;
dc_status_t rc;
unsigned int ntanks = 0;
rc = dc_parser_get_field(parser, DC_FIELD_TANK_COUNT, 0, &ntanks);
if (rc == DC_STATUS_SUCCESS) {
if (ntanks && ntanks < ngases) {
shown_warning = true;
report_error("Warning: different number of gases (%d) and cylinders (%d)", ngases, ntanks);
} else if (ntanks > ngases) {
shown_warning = true;
report_error("Warning: smaller number of gases (%d) than cylinders (%d).", ngases, ntanks);
}
}
bool no_volume = true;
struct gasmix bottom_gas = { {1000}, {0} }; /* Default to pure O2, or air if there are no mixes defined */
if (ngases == 0) {
bottom_gas = gasmix_air;
}
dive->cylinders.clear();
for (i = 0; i < std::max(ngases, ntanks); i++) {
cylinder_t cyl;
cyl.cylinder_use = NOT_USED;
if (i < ngases) {
dc_gasmix_t gasmix = { 0 };
int o2, he;
rc = dc_parser_get_field(parser, DC_FIELD_GASMIX, i, &gasmix);
if (rc == DC_STATUS_SUCCESS) {
o2 = lrint(gasmix.oxygen * 1000);
he = lrint(gasmix.helium * 1000);
/* Ignore bogus data - libdivecomputer does some crazy stuff */
if (o2 + he <= O2_IN_AIR || o2 > 1000) {
if (!shown_warning) {
shown_warning = true;
report_error("unlikely dive gas data from libdivecomputer: o2 = %.3f he = %.3f", gasmix.oxygen, gasmix.helium);
}
o2 = 0;
}
if (he < 0 || o2 + he > 1000) {
if (!shown_warning) {
shown_warning = true;
report_error("unlikely dive gas data from libdivecomputer: o2 = %.3f he = %.3f", gasmix.oxygen, gasmix.helium);
}
he = 0;
}
cyl.gasmix.o2.permille = o2;
cyl.gasmix.he.permille = he;
bottom_gas = get_deeper_gasmix(bottom_gas, cyl.gasmix);
switch (gasmix.usage) {
case DC_USAGE_DILUENT:
cyl.cylinder_use = DILUENT;
break;
case DC_USAGE_OXYGEN:
cyl.cylinder_use = OXYGEN;
break;
case DC_USAGE_OPEN_CIRCUIT:
cyl.cylinder_use = OC_GAS;
break;
default:
if (dive->dcs[0].divemode == CCR)
cyl.cylinder_use = DILUENT;
else
cyl.cylinder_use = OC_GAS;
break;
}
}
}
if (i < ntanks) {
// If we've run out of gas mixes, assign this cylinder to bottom
// gas. Note that this can be overridden below if the dive computer
// explicitly reports a gas mix for this tank.
if (i >= ngases) {
cyl.gasmix = bottom_gas;
}
dc_tank_t tank = { 0 };
rc = dc_parser_get_field(parser, DC_FIELD_TANK, i, &tank);
if (rc == DC_STATUS_SUCCESS) {
cyl.type.size.mliter = lrint(tank.volume * 1000);
cyl.type.workingpressure.mbar = lrint(tank.workpressure * 1000);
if (tank.type & DC_TANKVOLUME_IMPERIAL) {
if (devdata->model == "Suunto EON Steel") {
/* Suunto EON Steele gets this wrong. Badly.
* but on the plus side it only supports a few imperial sizes,
* so let's try and guess at least the most common ones.
* First, the pressures are off by a constant factor. WTF?
* Then we can round the wet sizes so we get to multiples of 10
* for cuft sizes (as that's all that you can enter) */
cyl.type.workingpressure.mbar = lrint(
cyl.type.workingpressure.mbar * 206.843 / 206.7 );
char name_buffer[17];
int rounded_size = lrint(ml_to_cuft(cyl.gas_volume(
cyl.type.workingpressure).mliter));
rounded_size = (int)((rounded_size + 5) / 10) * 10;
switch (cyl.type.workingpressure.mbar) {
case 206843:
snprintf(name_buffer, sizeof(name_buffer), "AL%d", rounded_size);
break;
case 234422: /* this is wrong - HP tanks tend to be 3440, but Suunto only allows 3400 */
snprintf(name_buffer, sizeof(name_buffer), "HP%d", rounded_size);
break;
case 179263:
snprintf(name_buffer, sizeof(name_buffer), "LP+%d", rounded_size);
break;
case 165474:
snprintf(name_buffer, sizeof(name_buffer), "LP%d", rounded_size);
break;
default:
snprintf(name_buffer, sizeof(name_buffer), "%d cuft", rounded_size);
break;
}
cyl.type.description = name_buffer;
cyl.type.size.mliter = lrint(cuft_to_l(rounded_size) * 1000 /
mbar_to_atm(cyl.type.workingpressure.mbar));
}
}
if (tank.gasmix != DC_GASMIX_UNKNOWN && tank.gasmix != i) { // we don't handle this, yet
shown_warning = true;
report_error("gasmix %d for tank %d doesn't match", tank.gasmix, i);
}
}
if (!nearly_0(tank.volume))
no_volume = false;
// this new API also gives us the beginning and end pressure for the tank
// normally 0 is not a valid pressure, but for some Uwatec dive computers we
// don't get the actual start and end pressure, but instead a start pressure
// that matches the consumption and an end pressure of always 0
// In order to make this work, we arbitrary shift this up by 30bar so the
// rest of the code treats this as if they were valid values
if (!nearly_0(tank.beginpressure)) {
if (!nearly_0(tank.endpressure)) {
cyl.start.mbar = lrint(tank.beginpressure * 1000);
cyl.end.mbar = lrint(tank.endpressure * 1000);
} else if (devdata->vendor == "Uwatec") {
cyl.start.mbar = lrint(tank.beginpressure * 1000 + 30000);
cyl.end.mbar = 30000;
}
}
}
if (no_volume) {
/* for the first tank, if there is no tanksize available from the
* dive computer, fill in the default tank information (if set) */
fill_default_cylinder(dive, &cyl);
}
/* whatever happens, make sure there is a name for the cylinder */
if (cyl.type.description.empty())
cyl.type.description = translate("gettextFromC", "unknown");
dive->cylinders.push_back(std::move(cyl));
}
return DC_STATUS_SUCCESS;
}
static void handle_event(struct divecomputer *dc, const struct sample &sample, dc_sample_value_t value)
{
int type, time;
struct event *ev;
/* we mark these for translation here, but we store the untranslated strings
* and only translate them when they are displayed on screen */
static const char *events[] = {
[SAMPLE_EVENT_NONE] = QT_TRANSLATE_NOOP("gettextFromC", "none"),
[SAMPLE_EVENT_DECOSTOP] = QT_TRANSLATE_NOOP("gettextFromC", "deco stop"),
[SAMPLE_EVENT_RBT] = QT_TRANSLATE_NOOP("gettextFromC", "rbt"),
[SAMPLE_EVENT_ASCENT] = QT_TRANSLATE_NOOP("gettextFromC", "ascent"),
[SAMPLE_EVENT_CEILING] = QT_TRANSLATE_NOOP("gettextFromC", "ceiling"),
[SAMPLE_EVENT_WORKLOAD] = QT_TRANSLATE_NOOP("gettextFromC", "workload"),
[SAMPLE_EVENT_TRANSMITTER] = QT_TRANSLATE_NOOP("gettextFromC", "transmitter"),
[SAMPLE_EVENT_VIOLATION] = QT_TRANSLATE_NOOP("gettextFromC", "violation"),
[SAMPLE_EVENT_BOOKMARK] = QT_TRANSLATE_NOOP("gettextFromC", "bookmark"),
[SAMPLE_EVENT_SURFACE] = QT_TRANSLATE_NOOP("gettextFromC", "surface"),
[SAMPLE_EVENT_SAFETYSTOP] = QT_TRANSLATE_NOOP("gettextFromC", "safety stop"),
[SAMPLE_EVENT_GASCHANGE] = QT_TRANSLATE_NOOP("gettextFromC", "gaschange"),
[SAMPLE_EVENT_SAFETYSTOP_VOLUNTARY] = QT_TRANSLATE_NOOP("gettextFromC", "safety stop (voluntary)"),
[SAMPLE_EVENT_SAFETYSTOP_MANDATORY] = QT_TRANSLATE_NOOP("gettextFromC", "safety stop (mandatory)"),
[SAMPLE_EVENT_DEEPSTOP] = QT_TRANSLATE_NOOP("gettextFromC", "deepstop"),
[SAMPLE_EVENT_CEILING_SAFETYSTOP] = QT_TRANSLATE_NOOP("gettextFromC", "ceiling (safety stop)"),
[SAMPLE_EVENT_FLOOR] = std::array<const char *, 2>{QT_TRANSLATE_NOOP3("gettextFromC", "below floor", "event showing dive is below deco floor and adding deco time")}[1],
[SAMPLE_EVENT_DIVETIME] = QT_TRANSLATE_NOOP("gettextFromC", "divetime"),
[SAMPLE_EVENT_MAXDEPTH] = QT_TRANSLATE_NOOP("gettextFromC", "maxdepth"),
[SAMPLE_EVENT_OLF] = QT_TRANSLATE_NOOP("gettextFromC", "OLF"),
[SAMPLE_EVENT_PO2] = QT_TRANSLATE_NOOP("gettextFromC", "pO₂"),
[SAMPLE_EVENT_AIRTIME] = QT_TRANSLATE_NOOP("gettextFromC", "airtime"),
[SAMPLE_EVENT_RGBM] = QT_TRANSLATE_NOOP("gettextFromC", "rgbm"),
[SAMPLE_EVENT_HEADING] = QT_TRANSLATE_NOOP("gettextFromC", "heading"),
[SAMPLE_EVENT_TISSUELEVEL] = QT_TRANSLATE_NOOP("gettextFromC", "tissue level warning"),
[SAMPLE_EVENT_GASCHANGE2] = QT_TRANSLATE_NOOP("gettextFromC", "gaschange"),
};
const int nr_events = sizeof(events) / sizeof(const char *);
const char *name;
/*
* Other evens might be more interesting, but for now we just print them out.
*/
type = value.event.type;
name = QT_TRANSLATE_NOOP("gettextFromC", "invalid event number");
if (type < nr_events && events[type])
name = events[type];
#ifdef SAMPLE_EVENT_STRING
if (type == SAMPLE_EVENT_STRING)
name = value.event.name;
#endif
time = value.event.time;
time += sample.time.seconds;
ev = add_event(dc, time, type, value.event.flags, value.event.value, name);
if (ev->is_gaschange() && ev->gas.index >= 0)
current_gas_index = ev->gas.index;
}
static void handle_gasmix(struct divecomputer *dc, const struct sample &sample, int idx)
Add support for libdivecomputer using DC_SAMPLE_GASMIX New libdivecomputer versions use DC_SAMPLE_GASMIX to indicate a gas change (which contains the cylinder index we're changing to) rather than SAMPLE_EVENT_GASCHANGE*. Unlike the old GASCHANGE model, and despite the name, DC_SAMPLE_GASMIX does not actually say what the mix is, it only specifies a cylinder index. We had already extended SAMPLE_EVENT_GASCHANGE2 to have the cylinder index in the otherwise unused "flags" field, so this is not all that different from what we used to do. And subsurface internally already had the logic that "if we know what the cylinder index is, take the gas mix from the cylinder data", so we've already been able to transparently use _either_ the actual gas mix or the cylinder index to show the event. But we do want to make it an event rather than some sample data, because we want to show it as such in the profile. But because we are happy with just the cylinder index, we'll just translate the DC_SAMPLE_GASMIX thing to the SAMPLE_EVENT_GASCHANGE2 event, and nothing really changes for subsurface. libdivecomputer has made other changes, like indicating the initial cylinder index with an early DC_SAMPLE_GASMIX report, but we've seen that before too (in the form of early SAMPLE_EVENT_GASCHANGE events), so that doesn't really end up changing anything for us either. HOWEVER, one thing that is worth noticing: do *not* apply this patch and then use an old libdivecomputer library that sends both the DC_SAMPLE_GASMIX samples _and_ the deprecated SAMPLE_EVENT_GASCHANGE events. It will all *work*, but since subsurface will take either, you'll then get duplicate gas mix events. It's not like that is in any way fatal, but it might be a bit confusing. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-01-03 23:18:03 +00:00
{
/* TODO: Verify that index is not higher than the number of cylinders */
if (idx < 0)
Add support for libdivecomputer using DC_SAMPLE_GASMIX New libdivecomputer versions use DC_SAMPLE_GASMIX to indicate a gas change (which contains the cylinder index we're changing to) rather than SAMPLE_EVENT_GASCHANGE*. Unlike the old GASCHANGE model, and despite the name, DC_SAMPLE_GASMIX does not actually say what the mix is, it only specifies a cylinder index. We had already extended SAMPLE_EVENT_GASCHANGE2 to have the cylinder index in the otherwise unused "flags" field, so this is not all that different from what we used to do. And subsurface internally already had the logic that "if we know what the cylinder index is, take the gas mix from the cylinder data", so we've already been able to transparently use _either_ the actual gas mix or the cylinder index to show the event. But we do want to make it an event rather than some sample data, because we want to show it as such in the profile. But because we are happy with just the cylinder index, we'll just translate the DC_SAMPLE_GASMIX thing to the SAMPLE_EVENT_GASCHANGE2 event, and nothing really changes for subsurface. libdivecomputer has made other changes, like indicating the initial cylinder index with an early DC_SAMPLE_GASMIX report, but we've seen that before too (in the form of early SAMPLE_EVENT_GASCHANGE events), so that doesn't really end up changing anything for us either. HOWEVER, one thing that is worth noticing: do *not* apply this patch and then use an old libdivecomputer library that sends both the DC_SAMPLE_GASMIX samples _and_ the deprecated SAMPLE_EVENT_GASCHANGE events. It will all *work*, but since subsurface will take either, you'll then get duplicate gas mix events. It's not like that is in any way fatal, but it might be a bit confusing. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-01-03 23:18:03 +00:00
return;
add_event(dc, sample.time.seconds, SAMPLE_EVENT_GASCHANGE2, idx+1, 0, "gaschange");
Add support for libdivecomputer using DC_SAMPLE_GASMIX New libdivecomputer versions use DC_SAMPLE_GASMIX to indicate a gas change (which contains the cylinder index we're changing to) rather than SAMPLE_EVENT_GASCHANGE*. Unlike the old GASCHANGE model, and despite the name, DC_SAMPLE_GASMIX does not actually say what the mix is, it only specifies a cylinder index. We had already extended SAMPLE_EVENT_GASCHANGE2 to have the cylinder index in the otherwise unused "flags" field, so this is not all that different from what we used to do. And subsurface internally already had the logic that "if we know what the cylinder index is, take the gas mix from the cylinder data", so we've already been able to transparently use _either_ the actual gas mix or the cylinder index to show the event. But we do want to make it an event rather than some sample data, because we want to show it as such in the profile. But because we are happy with just the cylinder index, we'll just translate the DC_SAMPLE_GASMIX thing to the SAMPLE_EVENT_GASCHANGE2 event, and nothing really changes for subsurface. libdivecomputer has made other changes, like indicating the initial cylinder index with an early DC_SAMPLE_GASMIX report, but we've seen that before too (in the form of early SAMPLE_EVENT_GASCHANGE events), so that doesn't really end up changing anything for us either. HOWEVER, one thing that is worth noticing: do *not* apply this patch and then use an old libdivecomputer library that sends both the DC_SAMPLE_GASMIX samples _and_ the deprecated SAMPLE_EVENT_GASCHANGE events. It will all *work*, but since subsurface will take either, you'll then get duplicate gas mix events. It's not like that is in any way fatal, but it might be a bit confusing. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-01-03 23:18:03 +00:00
current_gas_index = idx;
}
void
sample_cb(dc_sample_type_t type, const dc_sample_value_t *pvalue, void *userdata)
{
static unsigned int nsensor = 0;
struct divecomputer *dc = (divecomputer *)userdata;
dc_sample_value_t value = *pvalue;
/*
* DC_SAMPLE_TIME is special: it creates a new sample.
* Other types fill in an existing sample.
*/
if (type == DC_SAMPLE_TIME) {
nsensor = 0;
// Create a new sample.
// Mark depth as negative
struct sample *sample = prepare_sample(dc);
sample->time.seconds = value.time / 1000;
sample->depth.mm = -1;
// The current sample gets some sticky values
// that may have been around from before, these
// values will be overwritten by new data if available
sample->in_deco = in_deco;
sample->ndl.seconds = ndl;
sample->stoptime.seconds = stoptime;
sample->stopdepth.mm = stopdepth;
sample->setpoint.mbar = po2;
sample->cns = cns;
sample->heartbeat = heartbeat;
sample->bearing.degrees = bearing;
return;
}
if (dc->samples.empty())
prepare_sample(dc);
struct sample &sample = dc->samples.back();
switch (type) {
case DC_SAMPLE_DEPTH:
sample.depth.mm = lrint(value.depth * 1000);
break;
Try to sanely download multiple concurrent cylinder pressures This tries to sanely handle the case of a dive computer reporting multiple cylinder pressures concurrently. NOTE! There are various "interesting" situations that this whole issue brings up: - some dive computers may report more cylinder pressures than we have slots for. Currently we will drop such pressures on the floor if they come for the same sample, but if they end up being spread across multiple samples we will end up re-using the slots with different sensor indexes. That kind of slot re-use may or may not end up confusing other subsurface logic - for example, make things believe there was a cylidner change event. - some dive computers might send only one sample at a time, but switch *which* sample they send on a gas switch event. If they also report the correct sensor number, we'll now start reporting that pressure in the second slot. This should all be fine, and is the RightThing(tm) to do, but is different from what we used to do when we only ever used a single slot. - When people actually use multiple sensors, our old save format will start to need fixing. Right now our save format comes from the CCR model where the second sensor was always the Oxygen sensor. We save that pressure fine (except we save it as "o2pressure" - just an odd historical naming artifact), but we do *not* save the actual sensor index, because in our traditional format that was always implicit in the data ("it's the oxygen cylinder"). so while this code hopefully makes our libdivecomputer download do the right thing, there *will* be further fallout from having multiple cylinder pressure sensors. We're not done yet. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
2017-07-24 18:55:47 +00:00
case DC_SAMPLE_PRESSURE:
add_sample_pressure(&sample, value.pressure.tank, lrint(value.pressure.value * 1000));
break;
Add support for libdivecomputer using DC_SAMPLE_GASMIX New libdivecomputer versions use DC_SAMPLE_GASMIX to indicate a gas change (which contains the cylinder index we're changing to) rather than SAMPLE_EVENT_GASCHANGE*. Unlike the old GASCHANGE model, and despite the name, DC_SAMPLE_GASMIX does not actually say what the mix is, it only specifies a cylinder index. We had already extended SAMPLE_EVENT_GASCHANGE2 to have the cylinder index in the otherwise unused "flags" field, so this is not all that different from what we used to do. And subsurface internally already had the logic that "if we know what the cylinder index is, take the gas mix from the cylinder data", so we've already been able to transparently use _either_ the actual gas mix or the cylinder index to show the event. But we do want to make it an event rather than some sample data, because we want to show it as such in the profile. But because we are happy with just the cylinder index, we'll just translate the DC_SAMPLE_GASMIX thing to the SAMPLE_EVENT_GASCHANGE2 event, and nothing really changes for subsurface. libdivecomputer has made other changes, like indicating the initial cylinder index with an early DC_SAMPLE_GASMIX report, but we've seen that before too (in the form of early SAMPLE_EVENT_GASCHANGE events), so that doesn't really end up changing anything for us either. HOWEVER, one thing that is worth noticing: do *not* apply this patch and then use an old libdivecomputer library that sends both the DC_SAMPLE_GASMIX samples _and_ the deprecated SAMPLE_EVENT_GASCHANGE events. It will all *work*, but since subsurface will take either, you'll then get duplicate gas mix events. It's not like that is in any way fatal, but it might be a bit confusing. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-01-03 23:18:03 +00:00
case DC_SAMPLE_GASMIX:
handle_gasmix(dc, sample, value.gasmix);
break;
case DC_SAMPLE_TEMPERATURE:
sample.temperature.mkelvin = C_to_mkelvin(value.temperature);
break;
case DC_SAMPLE_EVENT:
handle_event(dc, sample, value);
break;
case DC_SAMPLE_RBT:
sample.rbt.seconds = (!strncasecmp(dc->model.c_str(), "suunto", 6)) ? value.rbt : value.rbt * 60;
break;
#ifdef DC_SAMPLE_TTS
case DC_SAMPLE_TTS:
sample.tts.seconds = value.time;
break;
#endif
case DC_SAMPLE_HEARTBEAT:
sample.heartbeat = heartbeat = value.heartbeat;
break;
case DC_SAMPLE_BEARING:
sample.bearing.degrees = bearing = value.bearing;
break;
#ifdef DEBUG_DC_VENDOR
case DC_SAMPLE_VENDOR:
printf(" <vendor time='%u:%02u' type=\"%u\" size=\"%u\">", FRACTION_TUPLE(sample.time.seconds, 60),
value.vendor.type, value.vendor.size);
for (int i = 0; i < value.vendor.size; ++i)
printf("%02X", ((unsigned char *)value.vendor.data)[i]);
printf("</vendor>\n");
break;
#endif
case DC_SAMPLE_SETPOINT:
/* for us a setpoint means constant pO2 from here */
sample.setpoint.mbar = po2 = lrint(value.setpoint * 1000);
break;
case DC_SAMPLE_PPO2:
if (nsensor < MAX_O2_SENSORS)
sample.o2sensor[nsensor].mbar = lrint(value.ppo2.value * 1000);
else
report_error("%d is more o2 sensors than we can handle", nsensor);
nsensor++;
// Set the amount of detected o2 sensors
if (nsensor > dc->no_o2sensors)
dc->no_o2sensors = nsensor;
break;
case DC_SAMPLE_CNS:
sample.cns = cns = lrint(value.cns * 100);
break;
case DC_SAMPLE_DECO:
if (value.deco.type == DC_DECO_NDL) {
sample.ndl.seconds = ndl = value.deco.time;
sample.stopdepth.mm = stopdepth = lrint(value.deco.depth * 1000.0);
sample.in_deco = in_deco = false;
} else if (value.deco.type == DC_DECO_DECOSTOP ||
value.deco.type == DC_DECO_DEEPSTOP) {
sample.stopdepth.mm = stopdepth = lrint(value.deco.depth * 1000.0);
sample.stoptime.seconds = stoptime = value.deco.time;
sample.in_deco = in_deco = stopdepth > 0;
ndl = 0;
} else if (value.deco.type == DC_DECO_SAFETYSTOP) {
sample.in_deco = in_deco = false;
sample.stopdepth.mm = stopdepth = lrint(value.deco.depth * 1000.0);
sample.stoptime.seconds = stoptime = value.deco.time;
}
sample.tts.seconds = value.deco.tts;
default:
break;
}
}
static void dev_info(const char *fmt, ...)
{
va_list ap;
va_start(ap, fmt);
progress_bar_text = vformat_string_std(fmt, ap);
va_end(ap);
if (verbose)
INFO("dev_info: %s", progress_bar_text.c_str());
if (progress_callback)
(*progress_callback)(progress_bar_text);
}
static int import_dive_number = 0;
static void download_error(const char *fmt, ...)
{
va_list ap;
va_start(ap, fmt);
std::string buffer = vformat_string_std(fmt, ap);
va_end(ap);
report_error("Dive %d: %s", import_dive_number, buffer.c_str());
}
static dc_status_t parse_samples(device_data_t *, struct divecomputer *dc, dc_parser_t *parser)
{
// Parse the sample data.
return dc_parser_samples_foreach(parser, sample_cb, dc);
}
static int might_be_same_dc(const struct divecomputer &a, const struct divecomputer &b)
{
if (a.model.empty() || b.model.empty())
return 1;
if (strcasecmp(a.model.c_str(), b.model.c_str()))
return 0;
if (!a.deviceid || !b.deviceid)
return 1;
return a.deviceid == b.deviceid;
}
static bool match_one_dive(const struct divecomputer &a, const struct dive &dive)
{
/*
* Walk the existing dive computer data,
* see if we have a match (or an anti-match:
* the same dive computer but a different
* dive ID).
*/
for (auto &b: dive.dcs) {
if (match_one_dc(a, b) > 0)
return true;
}
/* Ok, no exact dive computer match. Does the date match? */
for (auto &b: dive.dcs) {
if (a.when == b.when && might_be_same_dc(a, b))
return true;
}
return false;
}
/*
* Check if this dive already existed before the import
*/
static bool find_dive(const struct divecomputer &match)
{
return std::any_of(divelog.dives.rbegin(), divelog.dives.rend(),
[&match] (auto &old) { return match_one_dive(match, *old);} );
}
/*
* The dive ID for libdivecomputer dives is the first word of the
* SHA1 of the fingerprint, if it exists.
*
* NOTE! This is byte-order dependent, and I don't care.
*/
static uint32_t calculate_diveid(const unsigned char *fingerprint, unsigned int fsize)
{
if (!fingerprint || !fsize)
return 0;
return SHA1_uint32(fingerprint, fsize);
}
Clean up divecomputer 'device' handling We have this odd legacy notion of a divecomputer 'device', that was originally just basically the libdivecomputer 'EVENT_DEVINFO' report that was associated with each dive. So it had firmware version, deviceid, and serial number. It had also gotten extended to do 'nickname' handling, and it was all confusing, ugly and bad. It was particularly bad because it wasn't actually a 'per device' thing at all: due to the firmware field, a dive computer that got a firmware update forced a new 'device'. To make matters worse, the 'deviceid' was also almost random, because we've calculated it a couple of different ways, and libdivecomputer itself has changed how the legacy 32-bit 'serial number' is expressed. Finally, because of all these issues, we didn't even try to make the thing unique, so it really ended up being a random snapshot of the state of the dive computer at the time of a dive, and sometimes we'd pick one, and sometimes another, since they weren't really well-defined. So get rid of all this confusion. The new rules: - the actual random dive computer state at the time of a dive is kept in the dive data. So if you want to know the firmware version, it should be in the 'extra data' - the only serial number that matters is the string one in the extra data, because that's the one that actually matches what the dive computer reports, and isn't some random 32-bit integer with ambiguous formatting. - the 'device id' - the thing we match with (together with the model name, eg "Suunto EON Steel") is purely a hash of the real serial number. The device ID that libdivecomputer reports in EVENT_DEVINFO is ignored, as is the device ID we've saved in the XML or git files. If we have a serial number, the device ID will be uniquely associated with that serial number, and if we don't have one, the device ID will be zero (for 'match anything'). So now 'deviceid' is literally just a shorthand for the serial number string, and the two are joined at the hip. - the 'device' managament is _only_ used to track devices that have serial numbers _and_ nicknames. So no more different device structures just because one had a nickname and the other didn't etc. Without a serial number, the device is 'anonymous' and fundamentally cannot be distinguished from other devices of the same model, so a nickname is meaningless. And without a nickname, there is no point in creating a device data structure, since all the data is in the dive itself and the device structure wouldn't add any value.. These rules mean that we no longer have ambiguous 'device' structures, and we can never have duplicates that can confuse us. This does mean that you can't give a nickname to a device that cannot be uniquely identified with a serial number, but those are happily fairly rare (and mostly older ones). Dirk said he'd look at what it takes to give more dive computers proper serial numbers, and I already did it for the Garmin Descent family yesterday. (Honesty in advertizing: right now you can't add a nickname to a dive computer that doesn't already have one, because such a dive computer will not have a device structure. But that's a UI issue, and I'll sort that out separately) Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2021-08-16 22:50:11 +00:00
uint32_t calculate_string_hash(const char *str)
{
return calculate_diveid((const unsigned char *)str, strlen(str));
}
static void parse_string_field(device_data_t *devdata, struct dive *dive, dc_field_string_t *str)
{
// Our dive ID is the string hash of the "Dive ID" string
if (!strcmp(str->desc, "Dive ID")) {
if (!dive->dcs[0].diveid)
dive->dcs[0].diveid = calculate_string_hash(str->value);
return;
}
// This will pick up serial number and firmware data
add_extra_data(&dive->dcs[0], str->desc, str->value);
/* GPS data? */
if (!strncmp(str->desc, "GPS", 3)) {
char *line = (char *) str->value;
location_t location;
/* Do we already have a divesite? */
if (dive->dive_site) {
/*
* "GPS1" always takes precedence, anything else
* we'll just pick the first "GPS*" that matches.
*/
if (strcmp(str->desc, "GPS1") != 0)
return;
}
parse_location(line, &location);
if (location.lat.udeg && location.lon.udeg) {
unregister_dive_from_dive_site(dive);
devdata->log->sites.create(std::string(str->value), location)->add_dive(dive);
}
}
}
static dc_status_t libdc_header_parser(dc_parser_t *parser, device_data_t *devdata, struct dive *dive)
{
dc_status_t rc = static_cast<dc_status_t>(0);
dc_datetime_t dt = { 0 };
struct tm tm;
rc = dc_parser_get_datetime(parser, &dt);
if (rc != DC_STATUS_SUCCESS && rc != DC_STATUS_UNSUPPORTED) {
download_error(translate("gettextFromC", "Error parsing the datetime"));
return rc;
}
Re-do the libdivecomputer fingerprint save/load code This tries to make our fingerprinting code work better, by avoiding using the "deviceid" field that has always been unreliable because we've calculated it multiple different ways, and even for the same version of subsurface, it ends up changing in the middle (ie we calculate one value initially, then re-calculate it when we have a proper serial number string). So instead, the fingerprinting code will look up and save the fingerprint file using purely "stable" information that is available early during the download: - the device model name (which is a string with vendor and product name separated by a space) - the DC_EVENT_DEVINFO 32-bit 'serial' number (which is not necessarily a real serial number at all, but hopefully at least a unique number for the particular product) but because the model name is not necessarily a good filename (think slashes and other possibly invalid characters), we hash that model name and use the resulting hex number in the fingerprint file name. This way the fingerprint file is unambiguous at load and save time, and depends purely on libdivecomputer data. But because we also need to verify that we have the actual _dive_ associated with that fingerprint, we also need to save the final deviceid and diveid when saving the fingerprint file, so that when we load it again we can look up the dive and verify that we have it before we use the fingerprint data. To do that, the fingerprint file itself contains not just the fingerprint data from libdivecomputer, but the last 8 bytes of the file are the (subsurface) deviceid and the diveid of the dive that is associated with the fingerprint. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2021-09-19 01:52:50 +00:00
// Our deviceid is the hash of the serial number
dive->dcs[0].deviceid = 0;
if (rc == DC_STATUS_SUCCESS) {
tm.tm_year = dt.year;
tm.tm_mon = dt.month - 1;
tm.tm_mday = dt.day;
tm.tm_hour = dt.hour;
tm.tm_min = dt.minute;
tm.tm_sec = dt.second;
dive->when = dive->dcs[0].when = utc_mktime(&tm);
}
// Parse the divetime.
std::string date_string = get_dive_date_c_string(dive->when);
dev_info(translate("gettextFromC", "Dive %d: %s"), import_dive_number, date_string.c_str());
unsigned int divetime = 0;
rc = dc_parser_get_field(parser, DC_FIELD_DIVETIME, 0, &divetime);
if (rc != DC_STATUS_SUCCESS && rc != DC_STATUS_UNSUPPORTED) {
download_error(translate("gettextFromC", "Error parsing the divetime"));
return rc;
}
if (rc == DC_STATUS_SUCCESS)
dive->dcs[0].duration.seconds = divetime;
// Parse the maxdepth.
double maxdepth = 0.0;
rc = dc_parser_get_field(parser, DC_FIELD_MAXDEPTH, 0, &maxdepth);
if (rc != DC_STATUS_SUCCESS && rc != DC_STATUS_UNSUPPORTED) {
download_error(translate("gettextFromC", "Error parsing the maxdepth"));
return rc;
}
if (rc == DC_STATUS_SUCCESS)
dive->dcs[0].maxdepth.mm = lrint(maxdepth * 1000);
// Parse temperatures
double temperature;
dc_field_type_t temp_fields[] = {DC_FIELD_TEMPERATURE_SURFACE,
DC_FIELD_TEMPERATURE_MAXIMUM,
DC_FIELD_TEMPERATURE_MINIMUM};
for (int i = 0; i < 3; i++) {
rc = dc_parser_get_field(parser, temp_fields[i], 0, &temperature);
if (rc != DC_STATUS_SUCCESS && rc != DC_STATUS_UNSUPPORTED) {
download_error(translate("gettextFromC", "Error parsing temperature"));
return rc;
}
if (rc == DC_STATUS_SUCCESS)
switch(i) {
case 0:
dive->dcs[0].airtemp.mkelvin = C_to_mkelvin(temperature);
break;
case 1: // we don't distinguish min and max water temp here, so take min if given, max otherwise
case 2:
dive->dcs[0].watertemp.mkelvin = C_to_mkelvin(temperature);
break;
}
}
// Parse the gas mixes.
unsigned int ngases = 0;
rc = dc_parser_get_field(parser, DC_FIELD_GASMIX_COUNT, 0, &ngases);
if (rc != DC_STATUS_SUCCESS && rc != DC_STATUS_UNSUPPORTED) {
download_error(translate("gettextFromC", "Error parsing the gas mix count"));
return rc;
}
// Check if the libdivecomputer version already supports salinity & atmospheric
dc_salinity_t salinity = {
.type = DC_WATER_SALT,
.density = SEAWATER_SALINITY / 10.0
};
rc = dc_parser_get_field(parser, DC_FIELD_SALINITY, 0, &salinity);
if (rc != DC_STATUS_SUCCESS && rc != DC_STATUS_UNSUPPORTED) {
download_error(translate("gettextFromC", "Error obtaining water salinity"));
return rc;
}
if (rc == DC_STATUS_SUCCESS) {
dive->dcs[0].salinity = lrint(salinity.density * 10.0);
if (dive->dcs[0].salinity == 0) {
// sometimes libdivecomputer gives us density values, sometimes just
// a water type and a density of zero; let's make this work as best as we can
switch (salinity.type) {
case DC_WATER_FRESH:
dive->dcs[0].salinity = FRESHWATER_SALINITY;
break;
default:
dive->dcs[0].salinity = SEAWATER_SALINITY;
break;
}
}
}
double surface_pressure = 0;
rc = dc_parser_get_field(parser, DC_FIELD_ATMOSPHERIC, 0, &surface_pressure);
if (rc != DC_STATUS_SUCCESS && rc != DC_STATUS_UNSUPPORTED) {
download_error(translate("gettextFromC", "Error obtaining surface pressure"));
return rc;
}
if (rc == DC_STATUS_SUCCESS)
dive->dcs[0].surface_pressure.mbar = lrint(surface_pressure * 1000.0);
// The dive parsing may give us more device information
int idx;
for (idx = 0; idx < 100; idx++) {
dc_field_string_t str = { NULL };
rc = dc_parser_get_field(parser, DC_FIELD_STRING, idx, &str);
if (rc != DC_STATUS_SUCCESS)
break;
if (!str.desc || !str.value)
break;
parse_string_field(devdata, dive, &str);
free((void *)str.value); // libdc gives us copies of the value-string.
}
dc_divemode_t divemode;
rc = dc_parser_get_field(parser, DC_FIELD_DIVEMODE, 0, &divemode);
if (rc != DC_STATUS_SUCCESS && rc != DC_STATUS_UNSUPPORTED) {
download_error(translate("gettextFromC", "Error obtaining dive mode"));
return rc;
}
if (rc == DC_STATUS_SUCCESS)
switch(divemode) {
case DC_DIVEMODE_FREEDIVE:
dive->dcs[0].divemode = FREEDIVE;
break;
case DC_DIVEMODE_GAUGE:
case DC_DIVEMODE_OC: /* Open circuit */
dive->dcs[0].divemode = OC;
break;
case DC_DIVEMODE_CCR: /* Closed circuit rebreather*/
dive->dcs[0].divemode = CCR;
break;
case DC_DIVEMODE_SCR: /* Semi-closed circuit rebreather */
dive->dcs[0].divemode = PSCR;
break;
}
rc = parse_gasmixes(devdata, dive, parser, ngases);
if (rc != DC_STATUS_SUCCESS && rc != DC_STATUS_UNSUPPORTED) {
download_error(translate("gettextFromC", "Error parsing the gas mix"));
return rc;
}
return DC_STATUS_SUCCESS;
}
/* returns true if we want libdivecomputer's dc_device_foreach() to continue,
* false otherwise */
static int dive_cb(const unsigned char *data, unsigned int size,
const unsigned char *fingerprint, unsigned int fsize,
void *userdata)
{
dc_status_t rc;
dc_parser_t *parser = NULL;
device_data_t *devdata = (device_data_t *)userdata;
/* reset static data, that is only valid per dive */
stoptime = stopdepth = po2 = cns = heartbeat = 0;
ndl = bearing = -1;
in_deco = false;
current_gas_index = -1;
import_dive_number++;
rc = dc_parser_new(&parser, devdata->device, data, size);
if (rc != DC_STATUS_SUCCESS) {
download_error(translate("gettextFromC", "Unable to create parser for %s %s: %d"), devdata->vendor.c_str(), devdata->product.c_str(), errmsg(rc));
Keep parsing dives even if one dive parse failed Eric Charbonnier reported a problem downloading the dives from his OSTC2, and Jef debugged the libdivecomputer log and says: "Your ostc has 75 dives, but subsurface downloaded only one, and then stopped the download. That's because that first dive appears to be corrupt and fails to parse: ERROR: Buffer overflow detected! [in /win/subsurface/libdivecomputer/src/hw_ostc_parser.c:981 (hw_ostc_parser_samples_foreach)] Subsurface (incorrectly) considers that a fatal error and stops the entire download. From a user point of view, it would be much better to ignore the problematic dive, and continue downloading the remaining" Subsurface used to just stop downloading if there were parsing errors, but Jef further says: "How parser errors are handled is up to the application. Aborting the download is probably the worst option here. If a dive fails to parse (because the dive data is corrupt, the parser contains a bug, etc), that does not necessary mean the remaining dives can't be downloaded" so let's change the logic to just continue downloading, and hope other dives work better. We might want to do better error reporting, right now the errors tend to just cause "dev_info()" reports, which just set the progress bar text. So you'll see it in the progress bar as it happens, but it won't get really ever noted as an error, and it's easy to miss. But that error reporting is a separate issue, and this just does the "continue to the next dive" part. Reported-by: Eric Charbonnier <eric.charbonnier69@gmail.com> Suggested-by: Jef Driesen <jef@libdivecomputer.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-09-11 14:27:29 +00:00
return true;
}
auto dive = std::make_unique<struct dive>();
Preferentially use existing device ID when setting serial number We have two different models for setting the deviceid associated with a dive computer: either take the value from the libdivecomputer 'devinfo' field (from the DC_EVENT_DEVINFO event), or generate the device ID by just hashing the serial number string. The one thing we do *not* want to have, is to use both methods, so that the same device generates different device IDs. Because then we'll think we have two different dive computers even though they are one and the same. Usually, this is not an issue, because libdivecomputer either sends the DEVINFO event or gives us the serial number string, and we'll always just pick one or the other. However, in the case of at least the Suunto EON Steel, I intentionally did *not* send the DC_EVENT_DEVINFO event, because it gives no useful information. We used the serial number string to generate a device ID, and everything was fine. However, in commit d40cdb4755ee ("Add the devinfo event") in the libdivecomputer tree, Jeff started generating those DC_EVENT_DEVINFO events for the EON Steel too, and suddenly subsurface would start using a device ID based on that instead. The situation is inherently ambiguous - for the EON Steel, we want to use the hash of the serial number (because that is what we've historically done), but other dive computers might want to use the DEVINFO data (because that is what _those_ backends have historically done, even if they might also implement the new serial string model). This commit makes subsurface resolve this ambiguity by simply preferring whatever previous device ID it has associated with that particular serial number string. If you have no previous device IDs, it doesn't matter which one you pick. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
2016-06-21 00:59:26 +00:00
// Fill in basic fields
dive->dcs[0].model = devdata->model;
dive->dcs[0].diveid = calculate_diveid(fingerprint, fsize);
Preferentially use existing device ID when setting serial number We have two different models for setting the deviceid associated with a dive computer: either take the value from the libdivecomputer 'devinfo' field (from the DC_EVENT_DEVINFO event), or generate the device ID by just hashing the serial number string. The one thing we do *not* want to have, is to use both methods, so that the same device generates different device IDs. Because then we'll think we have two different dive computers even though they are one and the same. Usually, this is not an issue, because libdivecomputer either sends the DEVINFO event or gives us the serial number string, and we'll always just pick one or the other. However, in the case of at least the Suunto EON Steel, I intentionally did *not* send the DC_EVENT_DEVINFO event, because it gives no useful information. We used the serial number string to generate a device ID, and everything was fine. However, in commit d40cdb4755ee ("Add the devinfo event") in the libdivecomputer tree, Jeff started generating those DC_EVENT_DEVINFO events for the EON Steel too, and suddenly subsurface would start using a device ID based on that instead. The situation is inherently ambiguous - for the EON Steel, we want to use the hash of the serial number (because that is what we've historically done), but other dive computers might want to use the DEVINFO data (because that is what _those_ backends have historically done, even if they might also implement the new serial string model). This commit makes subsurface resolve this ambiguity by simply preferring whatever previous device ID it has associated with that particular serial number string. If you have no previous device IDs, it doesn't matter which one you pick. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
2016-06-21 00:59:26 +00:00
// Parse the dive's header data
rc = libdc_header_parser (parser, devdata, dive.get());
if (rc != DC_STATUS_SUCCESS) {
download_error(translate("getextFromC", "Error parsing the header: %s"), errmsg(rc));
goto error_exit;
}
// Initialize the sample data.
rc = parse_samples(devdata, &dive->dcs[0], parser);
if (rc != DC_STATUS_SUCCESS) {
download_error(translate("gettextFromC", "Error parsing the samples: %s"), errmsg(rc));
goto error_exit;
}
Keep parsing dives even if one dive parse failed Eric Charbonnier reported a problem downloading the dives from his OSTC2, and Jef debugged the libdivecomputer log and says: "Your ostc has 75 dives, but subsurface downloaded only one, and then stopped the download. That's because that first dive appears to be corrupt and fails to parse: ERROR: Buffer overflow detected! [in /win/subsurface/libdivecomputer/src/hw_ostc_parser.c:981 (hw_ostc_parser_samples_foreach)] Subsurface (incorrectly) considers that a fatal error and stops the entire download. From a user point of view, it would be much better to ignore the problematic dive, and continue downloading the remaining" Subsurface used to just stop downloading if there were parsing errors, but Jef further says: "How parser errors are handled is up to the application. Aborting the download is probably the worst option here. If a dive fails to parse (because the dive data is corrupt, the parser contains a bug, etc), that does not necessary mean the remaining dives can't be downloaded" so let's change the logic to just continue downloading, and hope other dives work better. We might want to do better error reporting, right now the errors tend to just cause "dev_info()" reports, which just set the progress bar text. So you'll see it in the progress bar as it happens, but it won't get really ever noted as an error, and it's easy to miss. But that error reporting is a separate issue, and this just does the "continue to the next dive" part. Reported-by: Eric Charbonnier <eric.charbonnier69@gmail.com> Suggested-by: Jef Driesen <jef@libdivecomputer.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-09-11 14:27:29 +00:00
dc_parser_destroy(parser);
Re-do the libdivecomputer fingerprint save/load code This tries to make our fingerprinting code work better, by avoiding using the "deviceid" field that has always been unreliable because we've calculated it multiple different ways, and even for the same version of subsurface, it ends up changing in the middle (ie we calculate one value initially, then re-calculate it when we have a proper serial number string). So instead, the fingerprinting code will look up and save the fingerprint file using purely "stable" information that is available early during the download: - the device model name (which is a string with vendor and product name separated by a space) - the DC_EVENT_DEVINFO 32-bit 'serial' number (which is not necessarily a real serial number at all, but hopefully at least a unique number for the particular product) but because the model name is not necessarily a good filename (think slashes and other possibly invalid characters), we hash that model name and use the resulting hex number in the fingerprint file name. This way the fingerprint file is unambiguous at load and save time, and depends purely on libdivecomputer data. But because we also need to verify that we have the actual _dive_ associated with that fingerprint, we also need to save the final deviceid and diveid when saving the fingerprint file, so that when we load it again we can look up the dive and verify that we have it before we use the fingerprint data. To do that, the fingerprint file itself contains not just the fingerprint data from libdivecomputer, but the last 8 bytes of the file are the (subsurface) deviceid and the diveid of the dive that is associated with the fingerprint. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2021-09-19 01:52:50 +00:00
/*
* Save off fingerprint data.
*
* NOTE! We do this after parsing the dive fully, so that
* we have the final deviceid here.
*/
if (fingerprint && fsize && !devdata->fingerprint) {
devdata->fingerprint = (unsigned char *)calloc(fsize, 1);
Re-do the libdivecomputer fingerprint save/load code This tries to make our fingerprinting code work better, by avoiding using the "deviceid" field that has always been unreliable because we've calculated it multiple different ways, and even for the same version of subsurface, it ends up changing in the middle (ie we calculate one value initially, then re-calculate it when we have a proper serial number string). So instead, the fingerprinting code will look up and save the fingerprint file using purely "stable" information that is available early during the download: - the device model name (which is a string with vendor and product name separated by a space) - the DC_EVENT_DEVINFO 32-bit 'serial' number (which is not necessarily a real serial number at all, but hopefully at least a unique number for the particular product) but because the model name is not necessarily a good filename (think slashes and other possibly invalid characters), we hash that model name and use the resulting hex number in the fingerprint file name. This way the fingerprint file is unambiguous at load and save time, and depends purely on libdivecomputer data. But because we also need to verify that we have the actual _dive_ associated with that fingerprint, we also need to save the final deviceid and diveid when saving the fingerprint file, so that when we load it again we can look up the dive and verify that we have it before we use the fingerprint data. To do that, the fingerprint file itself contains not just the fingerprint data from libdivecomputer, but the last 8 bytes of the file are the (subsurface) deviceid and the diveid of the dive that is associated with the fingerprint. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2021-09-19 01:52:50 +00:00
if (devdata->fingerprint) {
devdata->fsize = fsize;
devdata->fdeviceid = dive->dcs[0].deviceid;
devdata->fdiveid = dive->dcs[0].diveid;
Re-do the libdivecomputer fingerprint save/load code This tries to make our fingerprinting code work better, by avoiding using the "deviceid" field that has always been unreliable because we've calculated it multiple different ways, and even for the same version of subsurface, it ends up changing in the middle (ie we calculate one value initially, then re-calculate it when we have a proper serial number string). So instead, the fingerprinting code will look up and save the fingerprint file using purely "stable" information that is available early during the download: - the device model name (which is a string with vendor and product name separated by a space) - the DC_EVENT_DEVINFO 32-bit 'serial' number (which is not necessarily a real serial number at all, but hopefully at least a unique number for the particular product) but because the model name is not necessarily a good filename (think slashes and other possibly invalid characters), we hash that model name and use the resulting hex number in the fingerprint file name. This way the fingerprint file is unambiguous at load and save time, and depends purely on libdivecomputer data. But because we also need to verify that we have the actual _dive_ associated with that fingerprint, we also need to save the final deviceid and diveid when saving the fingerprint file, so that when we load it again we can look up the dive and verify that we have it before we use the fingerprint data. To do that, the fingerprint file itself contains not just the fingerprint data from libdivecomputer, but the last 8 bytes of the file are the (subsurface) deviceid and the diveid of the dive that is associated with the fingerprint. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2021-09-19 01:52:50 +00:00
memcpy(devdata->fingerprint, fingerprint, fsize);
}
}
/* If we already saw this dive, abort. */
if (!devdata->force_download && find_dive(dive->dcs[0])) {
std::string date_string = get_dive_date_c_string(dive->when);
dev_info(translate("gettextFromC", "Already downloaded dive at %s"), date_string.c_str());
Keep parsing dives even if one dive parse failed Eric Charbonnier reported a problem downloading the dives from his OSTC2, and Jef debugged the libdivecomputer log and says: "Your ostc has 75 dives, but subsurface downloaded only one, and then stopped the download. That's because that first dive appears to be corrupt and fails to parse: ERROR: Buffer overflow detected! [in /win/subsurface/libdivecomputer/src/hw_ostc_parser.c:981 (hw_ostc_parser_samples_foreach)] Subsurface (incorrectly) considers that a fatal error and stops the entire download. From a user point of view, it would be much better to ignore the problematic dive, and continue downloading the remaining" Subsurface used to just stop downloading if there were parsing errors, but Jef further says: "How parser errors are handled is up to the application. Aborting the download is probably the worst option here. If a dive fails to parse (because the dive data is corrupt, the parser contains a bug, etc), that does not necessary mean the remaining dives can't be downloaded" so let's change the logic to just continue downloading, and hope other dives work better. We might want to do better error reporting, right now the errors tend to just cause "dev_info()" reports, which just set the progress bar text. So you'll see it in the progress bar as it happens, but it won't get really ever noted as an error, and it's easy to miss. But that error reporting is a separate issue, and this just does the "continue to the next dive" part. Reported-by: Eric Charbonnier <eric.charbonnier69@gmail.com> Suggested-by: Jef Driesen <jef@libdivecomputer.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-09-11 14:27:29 +00:00
return false;
}
/* Various libdivecomputer interface fixups */
if (dive->dcs[0].airtemp.mkelvin == 0 && first_temp_is_air && !dive->dcs[0].samples.empty()) {
dive->dcs[0].airtemp = dive->dcs[0].samples[0].temperature;
dive->dcs[0].samples[0].temperature.mkelvin = 0;
}
/* special case for bug in Tecdiving DiveComputer.eu
* often the first sample has a water temperature of 0C, followed by the correct
* temperature in the next sample */
if (dive->dcs[0].model == "Tecdiving DiveComputer.eu" && !dive->dcs[0].samples.empty() &&
dive->dcs[0].samples[0].temperature.mkelvin == ZERO_C_IN_MKELVIN &&
dive->dcs[0].samples[1].temperature.mkelvin > dive->dcs[0].samples[0].temperature.mkelvin)
dive->dcs[0].samples[0].temperature.mkelvin = dive->dcs[0].samples[1].temperature.mkelvin;
devdata->log->dives.record_dive(std::move(dive));
return true;
error_exit:
dc_parser_destroy(parser);
Keep parsing dives even if one dive parse failed Eric Charbonnier reported a problem downloading the dives from his OSTC2, and Jef debugged the libdivecomputer log and says: "Your ostc has 75 dives, but subsurface downloaded only one, and then stopped the download. That's because that first dive appears to be corrupt and fails to parse: ERROR: Buffer overflow detected! [in /win/subsurface/libdivecomputer/src/hw_ostc_parser.c:981 (hw_ostc_parser_samples_foreach)] Subsurface (incorrectly) considers that a fatal error and stops the entire download. From a user point of view, it would be much better to ignore the problematic dive, and continue downloading the remaining" Subsurface used to just stop downloading if there were parsing errors, but Jef further says: "How parser errors are handled is up to the application. Aborting the download is probably the worst option here. If a dive fails to parse (because the dive data is corrupt, the parser contains a bug, etc), that does not necessary mean the remaining dives can't be downloaded" so let's change the logic to just continue downloading, and hope other dives work better. We might want to do better error reporting, right now the errors tend to just cause "dev_info()" reports, which just set the progress bar text. So you'll see it in the progress bar as it happens, but it won't get really ever noted as an error, and it's easy to miss. But that error reporting is a separate issue, and this just does the "continue to the next dive" part. Reported-by: Eric Charbonnier <eric.charbonnier69@gmail.com> Suggested-by: Jef Driesen <jef@libdivecomputer.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-09-11 14:27:29 +00:00
return true;
}
use libdivecomputer 'fingerprint' to avoid downloading extra data This opportunistically uses a cache of 'fingerprints' for already downloaded dives. As we download data from a dive computer, we save the fingerprint and dive ID of the most recent dive in a per-divecopmputer fingerprint cache file. The next time we download from that dive computer, we will load the cache file for that dive computer if it exists, verify that we still have the dive that is referenced in that cachefile, and if so use the fingerprint to let libdivecomputer potentially stop downloading dives early. This doesn't much matter for most dive computers, but some (like the Scubapro G2) are not able to download one dive at a time, and need the fingerprint to avoid doing a full dump. That is particularly noticeable over bluetooth, where a full dump can be very slow. NOTE! The fingerprint cache is a separate entity from the dive log itself. Unlike the dive log, it doesn't synchronize over the cloud, so if you download using different clients (say, your phone and your laptop), the fingerprint cache entries are per device. So you may still end up downloading dives you already have, because the fingerprint code basically only works to avoid duplicate downloads on the same installation. Also, note that we only have a cache of one single entry per dive computer and downloader, so if you download dives and then don't save the end result, the fingerprint will now point to a dive that you don't actually have in your dive list. As a result, next time you download, the fingerprint won't match any existing dive, and we'll resort to the old non-optimized behavior. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-04-24 17:37:31 +00:00
#ifndef O_BINARY
#define O_BINARY 0
#endif
static void do_save_fingerprint(device_data_t *devdata, const char *tmp, const char *final)
{
int fd, written = -1;
fd = subsurface_open(tmp, O_WRONLY | O_BINARY | O_CREAT | O_TRUNC, 0666);
if (fd < 0)
return;
if (verbose)
dev_info("Saving fingerprint for %08x:%08x to '%s'",
devdata->fdeviceid, devdata->fdiveid, final);
Re-do the libdivecomputer fingerprint save/load code This tries to make our fingerprinting code work better, by avoiding using the "deviceid" field that has always been unreliable because we've calculated it multiple different ways, and even for the same version of subsurface, it ends up changing in the middle (ie we calculate one value initially, then re-calculate it when we have a proper serial number string). So instead, the fingerprinting code will look up and save the fingerprint file using purely "stable" information that is available early during the download: - the device model name (which is a string with vendor and product name separated by a space) - the DC_EVENT_DEVINFO 32-bit 'serial' number (which is not necessarily a real serial number at all, but hopefully at least a unique number for the particular product) but because the model name is not necessarily a good filename (think slashes and other possibly invalid characters), we hash that model name and use the resulting hex number in the fingerprint file name. This way the fingerprint file is unambiguous at load and save time, and depends purely on libdivecomputer data. But because we also need to verify that we have the actual _dive_ associated with that fingerprint, we also need to save the final deviceid and diveid when saving the fingerprint file, so that when we load it again we can look up the dive and verify that we have it before we use the fingerprint data. To do that, the fingerprint file itself contains not just the fingerprint data from libdivecomputer, but the last 8 bytes of the file are the (subsurface) deviceid and the diveid of the dive that is associated with the fingerprint. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2021-09-19 01:52:50 +00:00
use libdivecomputer 'fingerprint' to avoid downloading extra data This opportunistically uses a cache of 'fingerprints' for already downloaded dives. As we download data from a dive computer, we save the fingerprint and dive ID of the most recent dive in a per-divecopmputer fingerprint cache file. The next time we download from that dive computer, we will load the cache file for that dive computer if it exists, verify that we still have the dive that is referenced in that cachefile, and if so use the fingerprint to let libdivecomputer potentially stop downloading dives early. This doesn't much matter for most dive computers, but some (like the Scubapro G2) are not able to download one dive at a time, and need the fingerprint to avoid doing a full dump. That is particularly noticeable over bluetooth, where a full dump can be very slow. NOTE! The fingerprint cache is a separate entity from the dive log itself. Unlike the dive log, it doesn't synchronize over the cloud, so if you download using different clients (say, your phone and your laptop), the fingerprint cache entries are per device. So you may still end up downloading dives you already have, because the fingerprint code basically only works to avoid duplicate downloads on the same installation. Also, note that we only have a cache of one single entry per dive computer and downloader, so if you download dives and then don't save the end result, the fingerprint will now point to a dive that you don't actually have in your dive list. As a result, next time you download, the fingerprint won't match any existing dive, and we'll resort to the old non-optimized behavior. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-04-24 17:37:31 +00:00
/* The fingerprint itself.. */
written = write(fd, devdata->fingerprint, devdata->fsize);
Re-do the libdivecomputer fingerprint save/load code This tries to make our fingerprinting code work better, by avoiding using the "deviceid" field that has always been unreliable because we've calculated it multiple different ways, and even for the same version of subsurface, it ends up changing in the middle (ie we calculate one value initially, then re-calculate it when we have a proper serial number string). So instead, the fingerprinting code will look up and save the fingerprint file using purely "stable" information that is available early during the download: - the device model name (which is a string with vendor and product name separated by a space) - the DC_EVENT_DEVINFO 32-bit 'serial' number (which is not necessarily a real serial number at all, but hopefully at least a unique number for the particular product) but because the model name is not necessarily a good filename (think slashes and other possibly invalid characters), we hash that model name and use the resulting hex number in the fingerprint file name. This way the fingerprint file is unambiguous at load and save time, and depends purely on libdivecomputer data. But because we also need to verify that we have the actual _dive_ associated with that fingerprint, we also need to save the final deviceid and diveid when saving the fingerprint file, so that when we load it again we can look up the dive and verify that we have it before we use the fingerprint data. To do that, the fingerprint file itself contains not just the fingerprint data from libdivecomputer, but the last 8 bytes of the file are the (subsurface) deviceid and the diveid of the dive that is associated with the fingerprint. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2021-09-19 01:52:50 +00:00
/* ..followed by the device ID and dive ID of the fingerprinted dive */
if (write(fd, &devdata->fdeviceid, 4) != 4 || write(fd, &devdata->fdiveid, 4) != 4)
use libdivecomputer 'fingerprint' to avoid downloading extra data This opportunistically uses a cache of 'fingerprints' for already downloaded dives. As we download data from a dive computer, we save the fingerprint and dive ID of the most recent dive in a per-divecopmputer fingerprint cache file. The next time we download from that dive computer, we will load the cache file for that dive computer if it exists, verify that we still have the dive that is referenced in that cachefile, and if so use the fingerprint to let libdivecomputer potentially stop downloading dives early. This doesn't much matter for most dive computers, but some (like the Scubapro G2) are not able to download one dive at a time, and need the fingerprint to avoid doing a full dump. That is particularly noticeable over bluetooth, where a full dump can be very slow. NOTE! The fingerprint cache is a separate entity from the dive log itself. Unlike the dive log, it doesn't synchronize over the cloud, so if you download using different clients (say, your phone and your laptop), the fingerprint cache entries are per device. So you may still end up downloading dives you already have, because the fingerprint code basically only works to avoid duplicate downloads on the same installation. Also, note that we only have a cache of one single entry per dive computer and downloader, so if you download dives and then don't save the end result, the fingerprint will now point to a dive that you don't actually have in your dive list. As a result, next time you download, the fingerprint won't match any existing dive, and we'll resort to the old non-optimized behavior. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-04-24 17:37:31 +00:00
written = -1;
/* I'd like to do fsync() here too, but does Windows support it? */
if (close(fd) < 0)
written = -1;
if (written == (int)devdata->fsize) {
if (!subsurface_rename(tmp, final))
use libdivecomputer 'fingerprint' to avoid downloading extra data This opportunistically uses a cache of 'fingerprints' for already downloaded dives. As we download data from a dive computer, we save the fingerprint and dive ID of the most recent dive in a per-divecopmputer fingerprint cache file. The next time we download from that dive computer, we will load the cache file for that dive computer if it exists, verify that we still have the dive that is referenced in that cachefile, and if so use the fingerprint to let libdivecomputer potentially stop downloading dives early. This doesn't much matter for most dive computers, but some (like the Scubapro G2) are not able to download one dive at a time, and need the fingerprint to avoid doing a full dump. That is particularly noticeable over bluetooth, where a full dump can be very slow. NOTE! The fingerprint cache is a separate entity from the dive log itself. Unlike the dive log, it doesn't synchronize over the cloud, so if you download using different clients (say, your phone and your laptop), the fingerprint cache entries are per device. So you may still end up downloading dives you already have, because the fingerprint code basically only works to avoid duplicate downloads on the same installation. Also, note that we only have a cache of one single entry per dive computer and downloader, so if you download dives and then don't save the end result, the fingerprint will now point to a dive that you don't actually have in your dive list. As a result, next time you download, the fingerprint won't match any existing dive, and we'll resort to the old non-optimized behavior. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-04-24 17:37:31 +00:00
return;
}
unlink(tmp);
}
static std::string fingerprint_file(device_data_t *devdata)
Re-do the libdivecomputer fingerprint save/load code This tries to make our fingerprinting code work better, by avoiding using the "deviceid" field that has always been unreliable because we've calculated it multiple different ways, and even for the same version of subsurface, it ends up changing in the middle (ie we calculate one value initially, then re-calculate it when we have a proper serial number string). So instead, the fingerprinting code will look up and save the fingerprint file using purely "stable" information that is available early during the download: - the device model name (which is a string with vendor and product name separated by a space) - the DC_EVENT_DEVINFO 32-bit 'serial' number (which is not necessarily a real serial number at all, but hopefully at least a unique number for the particular product) but because the model name is not necessarily a good filename (think slashes and other possibly invalid characters), we hash that model name and use the resulting hex number in the fingerprint file name. This way the fingerprint file is unambiguous at load and save time, and depends purely on libdivecomputer data. But because we also need to verify that we have the actual _dive_ associated with that fingerprint, we also need to save the final deviceid and diveid when saving the fingerprint file, so that when we load it again we can look up the dive and verify that we have it before we use the fingerprint data. To do that, the fingerprint file itself contains not just the fingerprint data from libdivecomputer, but the last 8 bytes of the file are the (subsurface) deviceid and the diveid of the dive that is associated with the fingerprint. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2021-09-19 01:52:50 +00:00
{
uint32_t model, serial;
// Model hash and libdivecomputer 32-bit 'serial number' for the file name
model = calculate_string_hash(devdata->model.c_str());
Re-do the libdivecomputer fingerprint save/load code This tries to make our fingerprinting code work better, by avoiding using the "deviceid" field that has always been unreliable because we've calculated it multiple different ways, and even for the same version of subsurface, it ends up changing in the middle (ie we calculate one value initially, then re-calculate it when we have a proper serial number string). So instead, the fingerprinting code will look up and save the fingerprint file using purely "stable" information that is available early during the download: - the device model name (which is a string with vendor and product name separated by a space) - the DC_EVENT_DEVINFO 32-bit 'serial' number (which is not necessarily a real serial number at all, but hopefully at least a unique number for the particular product) but because the model name is not necessarily a good filename (think slashes and other possibly invalid characters), we hash that model name and use the resulting hex number in the fingerprint file name. This way the fingerprint file is unambiguous at load and save time, and depends purely on libdivecomputer data. But because we also need to verify that we have the actual _dive_ associated with that fingerprint, we also need to save the final deviceid and diveid when saving the fingerprint file, so that when we load it again we can look up the dive and verify that we have it before we use the fingerprint data. To do that, the fingerprint file itself contains not just the fingerprint data from libdivecomputer, but the last 8 bytes of the file are the (subsurface) deviceid and the diveid of the dive that is associated with the fingerprint. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2021-09-19 01:52:50 +00:00
serial = devdata->devinfo.serial;
return format_string_std("%s/fingerprints/%04x.%u",
system_default_directory().c_str(),
Re-do the libdivecomputer fingerprint save/load code This tries to make our fingerprinting code work better, by avoiding using the "deviceid" field that has always been unreliable because we've calculated it multiple different ways, and even for the same version of subsurface, it ends up changing in the middle (ie we calculate one value initially, then re-calculate it when we have a proper serial number string). So instead, the fingerprinting code will look up and save the fingerprint file using purely "stable" information that is available early during the download: - the device model name (which is a string with vendor and product name separated by a space) - the DC_EVENT_DEVINFO 32-bit 'serial' number (which is not necessarily a real serial number at all, but hopefully at least a unique number for the particular product) but because the model name is not necessarily a good filename (think slashes and other possibly invalid characters), we hash that model name and use the resulting hex number in the fingerprint file name. This way the fingerprint file is unambiguous at load and save time, and depends purely on libdivecomputer data. But because we also need to verify that we have the actual _dive_ associated with that fingerprint, we also need to save the final deviceid and diveid when saving the fingerprint file, so that when we load it again we can look up the dive and verify that we have it before we use the fingerprint data. To do that, the fingerprint file itself contains not just the fingerprint data from libdivecomputer, but the last 8 bytes of the file are the (subsurface) deviceid and the diveid of the dive that is associated with the fingerprint. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2021-09-19 01:52:50 +00:00
model, serial);
}
use libdivecomputer 'fingerprint' to avoid downloading extra data This opportunistically uses a cache of 'fingerprints' for already downloaded dives. As we download data from a dive computer, we save the fingerprint and dive ID of the most recent dive in a per-divecopmputer fingerprint cache file. The next time we download from that dive computer, we will load the cache file for that dive computer if it exists, verify that we still have the dive that is referenced in that cachefile, and if so use the fingerprint to let libdivecomputer potentially stop downloading dives early. This doesn't much matter for most dive computers, but some (like the Scubapro G2) are not able to download one dive at a time, and need the fingerprint to avoid doing a full dump. That is particularly noticeable over bluetooth, where a full dump can be very slow. NOTE! The fingerprint cache is a separate entity from the dive log itself. Unlike the dive log, it doesn't synchronize over the cloud, so if you download using different clients (say, your phone and your laptop), the fingerprint cache entries are per device. So you may still end up downloading dives you already have, because the fingerprint code basically only works to avoid duplicate downloads on the same installation. Also, note that we only have a cache of one single entry per dive computer and downloader, so if you download dives and then don't save the end result, the fingerprint will now point to a dive that you don't actually have in your dive list. As a result, next time you download, the fingerprint won't match any existing dive, and we'll resort to the old non-optimized behavior. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-04-24 17:37:31 +00:00
/*
* Save the fingerprint after a successful download
Re-do the libdivecomputer fingerprint save/load code This tries to make our fingerprinting code work better, by avoiding using the "deviceid" field that has always been unreliable because we've calculated it multiple different ways, and even for the same version of subsurface, it ends up changing in the middle (ie we calculate one value initially, then re-calculate it when we have a proper serial number string). So instead, the fingerprinting code will look up and save the fingerprint file using purely "stable" information that is available early during the download: - the device model name (which is a string with vendor and product name separated by a space) - the DC_EVENT_DEVINFO 32-bit 'serial' number (which is not necessarily a real serial number at all, but hopefully at least a unique number for the particular product) but because the model name is not necessarily a good filename (think slashes and other possibly invalid characters), we hash that model name and use the resulting hex number in the fingerprint file name. This way the fingerprint file is unambiguous at load and save time, and depends purely on libdivecomputer data. But because we also need to verify that we have the actual _dive_ associated with that fingerprint, we also need to save the final deviceid and diveid when saving the fingerprint file, so that when we load it again we can look up the dive and verify that we have it before we use the fingerprint data. To do that, the fingerprint file itself contains not just the fingerprint data from libdivecomputer, but the last 8 bytes of the file are the (subsurface) deviceid and the diveid of the dive that is associated with the fingerprint. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2021-09-19 01:52:50 +00:00
*
* NOTE! At this point, we have the final device ID for the divecomputer
* we downloaded from. But that 'deviceid' is actually not useful, because
* at the point where we want to _load_ this, we only have the libdivecomputer
* DC_EVENT_DEVINFO state (devdata->devinfo).
*
* Now, we do have the devdata->devinfo at save time, but at load time we
* need to verify not only that it's the proper fingerprint file: we also
* need to check that we actually have the particular dive that was
* associated with that fingerprint state.
*
* That means that the fingerprint save file needs to include not only the
* fingerprint data itself, but also enough data to look up a dive unambiguously
* when loading the fingerprint. And the fingerprint data needs to be looked
* up using the DC_EVENT_DEVINFO data.
*
* End result:
*
* - fingerprint filename depends on the model name and 'devinfo.serial'
* so that we can look it up at DC_EVENT_DEVINFO time before the full
* info has been parsed.
*
* - the fingerprint file contains the 'diveid' of the fingerprinted dive,
* which is just a hash of the fingerprint itself.
*
* - we also save the final 'deviceid' in the fingerprint file, so that
* looking up the dive associated with the fingerprint is possible.
use libdivecomputer 'fingerprint' to avoid downloading extra data This opportunistically uses a cache of 'fingerprints' for already downloaded dives. As we download data from a dive computer, we save the fingerprint and dive ID of the most recent dive in a per-divecopmputer fingerprint cache file. The next time we download from that dive computer, we will load the cache file for that dive computer if it exists, verify that we still have the dive that is referenced in that cachefile, and if so use the fingerprint to let libdivecomputer potentially stop downloading dives early. This doesn't much matter for most dive computers, but some (like the Scubapro G2) are not able to download one dive at a time, and need the fingerprint to avoid doing a full dump. That is particularly noticeable over bluetooth, where a full dump can be very slow. NOTE! The fingerprint cache is a separate entity from the dive log itself. Unlike the dive log, it doesn't synchronize over the cloud, so if you download using different clients (say, your phone and your laptop), the fingerprint cache entries are per device. So you may still end up downloading dives you already have, because the fingerprint code basically only works to avoid duplicate downloads on the same installation. Also, note that we only have a cache of one single entry per dive computer and downloader, so if you download dives and then don't save the end result, the fingerprint will now point to a dive that you don't actually have in your dive list. As a result, next time you download, the fingerprint won't match any existing dive, and we'll resort to the old non-optimized behavior. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-04-24 17:37:31 +00:00
*/
static void save_fingerprint(device_data_t *devdata)
{
Re-do the libdivecomputer fingerprint save/load code This tries to make our fingerprinting code work better, by avoiding using the "deviceid" field that has always been unreliable because we've calculated it multiple different ways, and even for the same version of subsurface, it ends up changing in the middle (ie we calculate one value initially, then re-calculate it when we have a proper serial number string). So instead, the fingerprinting code will look up and save the fingerprint file using purely "stable" information that is available early during the download: - the device model name (which is a string with vendor and product name separated by a space) - the DC_EVENT_DEVINFO 32-bit 'serial' number (which is not necessarily a real serial number at all, but hopefully at least a unique number for the particular product) but because the model name is not necessarily a good filename (think slashes and other possibly invalid characters), we hash that model name and use the resulting hex number in the fingerprint file name. This way the fingerprint file is unambiguous at load and save time, and depends purely on libdivecomputer data. But because we also need to verify that we have the actual _dive_ associated with that fingerprint, we also need to save the final deviceid and diveid when saving the fingerprint file, so that when we load it again we can look up the dive and verify that we have it before we use the fingerprint data. To do that, the fingerprint file itself contains not just the fingerprint data from libdivecomputer, but the last 8 bytes of the file are the (subsurface) deviceid and the diveid of the dive that is associated with the fingerprint. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2021-09-19 01:52:50 +00:00
// Don't try to save nonexistent fingerprint data
if (!devdata->fingerprint || !devdata->fdiveid)
use libdivecomputer 'fingerprint' to avoid downloading extra data This opportunistically uses a cache of 'fingerprints' for already downloaded dives. As we download data from a dive computer, we save the fingerprint and dive ID of the most recent dive in a per-divecopmputer fingerprint cache file. The next time we download from that dive computer, we will load the cache file for that dive computer if it exists, verify that we still have the dive that is referenced in that cachefile, and if so use the fingerprint to let libdivecomputer potentially stop downloading dives early. This doesn't much matter for most dive computers, but some (like the Scubapro G2) are not able to download one dive at a time, and need the fingerprint to avoid doing a full dump. That is particularly noticeable over bluetooth, where a full dump can be very slow. NOTE! The fingerprint cache is a separate entity from the dive log itself. Unlike the dive log, it doesn't synchronize over the cloud, so if you download using different clients (say, your phone and your laptop), the fingerprint cache entries are per device. So you may still end up downloading dives you already have, because the fingerprint code basically only works to avoid duplicate downloads on the same installation. Also, note that we only have a cache of one single entry per dive computer and downloader, so if you download dives and then don't save the end result, the fingerprint will now point to a dive that you don't actually have in your dive list. As a result, next time you download, the fingerprint won't match any existing dive, and we'll resort to the old non-optimized behavior. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-04-24 17:37:31 +00:00
return;
Re-do the libdivecomputer fingerprint save/load code This tries to make our fingerprinting code work better, by avoiding using the "deviceid" field that has always been unreliable because we've calculated it multiple different ways, and even for the same version of subsurface, it ends up changing in the middle (ie we calculate one value initially, then re-calculate it when we have a proper serial number string). So instead, the fingerprinting code will look up and save the fingerprint file using purely "stable" information that is available early during the download: - the device model name (which is a string with vendor and product name separated by a space) - the DC_EVENT_DEVINFO 32-bit 'serial' number (which is not necessarily a real serial number at all, but hopefully at least a unique number for the particular product) but because the model name is not necessarily a good filename (think slashes and other possibly invalid characters), we hash that model name and use the resulting hex number in the fingerprint file name. This way the fingerprint file is unambiguous at load and save time, and depends purely on libdivecomputer data. But because we also need to verify that we have the actual _dive_ associated with that fingerprint, we also need to save the final deviceid and diveid when saving the fingerprint file, so that when we load it again we can look up the dive and verify that we have it before we use the fingerprint data. To do that, the fingerprint file itself contains not just the fingerprint data from libdivecomputer, but the last 8 bytes of the file are the (subsurface) deviceid and the diveid of the dive that is associated with the fingerprint. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2021-09-19 01:52:50 +00:00
// Make sure the fingerprints directory exists
std::string dir = system_default_directory() + "/fingerprints";
subsurface_mkdir(dir.c_str());
Re-do the libdivecomputer fingerprint save/load code This tries to make our fingerprinting code work better, by avoiding using the "deviceid" field that has always been unreliable because we've calculated it multiple different ways, and even for the same version of subsurface, it ends up changing in the middle (ie we calculate one value initially, then re-calculate it when we have a proper serial number string). So instead, the fingerprinting code will look up and save the fingerprint file using purely "stable" information that is available early during the download: - the device model name (which is a string with vendor and product name separated by a space) - the DC_EVENT_DEVINFO 32-bit 'serial' number (which is not necessarily a real serial number at all, but hopefully at least a unique number for the particular product) but because the model name is not necessarily a good filename (think slashes and other possibly invalid characters), we hash that model name and use the resulting hex number in the fingerprint file name. This way the fingerprint file is unambiguous at load and save time, and depends purely on libdivecomputer data. But because we also need to verify that we have the actual _dive_ associated with that fingerprint, we also need to save the final deviceid and diveid when saving the fingerprint file, so that when we load it again we can look up the dive and verify that we have it before we use the fingerprint data. To do that, the fingerprint file itself contains not just the fingerprint data from libdivecomputer, but the last 8 bytes of the file are the (subsurface) deviceid and the diveid of the dive that is associated with the fingerprint. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2021-09-19 01:52:50 +00:00
std::string final = fingerprint_file(devdata);
std::string tmp = final + ".tmp";
use libdivecomputer 'fingerprint' to avoid downloading extra data This opportunistically uses a cache of 'fingerprints' for already downloaded dives. As we download data from a dive computer, we save the fingerprint and dive ID of the most recent dive in a per-divecopmputer fingerprint cache file. The next time we download from that dive computer, we will load the cache file for that dive computer if it exists, verify that we still have the dive that is referenced in that cachefile, and if so use the fingerprint to let libdivecomputer potentially stop downloading dives early. This doesn't much matter for most dive computers, but some (like the Scubapro G2) are not able to download one dive at a time, and need the fingerprint to avoid doing a full dump. That is particularly noticeable over bluetooth, where a full dump can be very slow. NOTE! The fingerprint cache is a separate entity from the dive log itself. Unlike the dive log, it doesn't synchronize over the cloud, so if you download using different clients (say, your phone and your laptop), the fingerprint cache entries are per device. So you may still end up downloading dives you already have, because the fingerprint code basically only works to avoid duplicate downloads on the same installation. Also, note that we only have a cache of one single entry per dive computer and downloader, so if you download dives and then don't save the end result, the fingerprint will now point to a dive that you don't actually have in your dive list. As a result, next time you download, the fingerprint won't match any existing dive, and we'll resort to the old non-optimized behavior. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-04-24 17:37:31 +00:00
do_save_fingerprint(devdata, tmp.c_str(), final.c_str());
use libdivecomputer 'fingerprint' to avoid downloading extra data This opportunistically uses a cache of 'fingerprints' for already downloaded dives. As we download data from a dive computer, we save the fingerprint and dive ID of the most recent dive in a per-divecopmputer fingerprint cache file. The next time we download from that dive computer, we will load the cache file for that dive computer if it exists, verify that we still have the dive that is referenced in that cachefile, and if so use the fingerprint to let libdivecomputer potentially stop downloading dives early. This doesn't much matter for most dive computers, but some (like the Scubapro G2) are not able to download one dive at a time, and need the fingerprint to avoid doing a full dump. That is particularly noticeable over bluetooth, where a full dump can be very slow. NOTE! The fingerprint cache is a separate entity from the dive log itself. Unlike the dive log, it doesn't synchronize over the cloud, so if you download using different clients (say, your phone and your laptop), the fingerprint cache entries are per device. So you may still end up downloading dives you already have, because the fingerprint code basically only works to avoid duplicate downloads on the same installation. Also, note that we only have a cache of one single entry per dive computer and downloader, so if you download dives and then don't save the end result, the fingerprint will now point to a dive that you don't actually have in your dive list. As a result, next time you download, the fingerprint won't match any existing dive, and we'll resort to the old non-optimized behavior. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-04-24 17:37:31 +00:00
}
/*
* The fingerprint cache files contain the actual libdivecomputer
Re-do the libdivecomputer fingerprint save/load code This tries to make our fingerprinting code work better, by avoiding using the "deviceid" field that has always been unreliable because we've calculated it multiple different ways, and even for the same version of subsurface, it ends up changing in the middle (ie we calculate one value initially, then re-calculate it when we have a proper serial number string). So instead, the fingerprinting code will look up and save the fingerprint file using purely "stable" information that is available early during the download: - the device model name (which is a string with vendor and product name separated by a space) - the DC_EVENT_DEVINFO 32-bit 'serial' number (which is not necessarily a real serial number at all, but hopefully at least a unique number for the particular product) but because the model name is not necessarily a good filename (think slashes and other possibly invalid characters), we hash that model name and use the resulting hex number in the fingerprint file name. This way the fingerprint file is unambiguous at load and save time, and depends purely on libdivecomputer data. But because we also need to verify that we have the actual _dive_ associated with that fingerprint, we also need to save the final deviceid and diveid when saving the fingerprint file, so that when we load it again we can look up the dive and verify that we have it before we use the fingerprint data. To do that, the fingerprint file itself contains not just the fingerprint data from libdivecomputer, but the last 8 bytes of the file are the (subsurface) deviceid and the diveid of the dive that is associated with the fingerprint. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2021-09-19 01:52:50 +00:00
* fingerprint, followed by 8 bytes of (deviceid,diveid) data.
*
* Before we use the fingerprint data, verify that we actually
* do have that fingerprinted dive.
use libdivecomputer 'fingerprint' to avoid downloading extra data This opportunistically uses a cache of 'fingerprints' for already downloaded dives. As we download data from a dive computer, we save the fingerprint and dive ID of the most recent dive in a per-divecopmputer fingerprint cache file. The next time we download from that dive computer, we will load the cache file for that dive computer if it exists, verify that we still have the dive that is referenced in that cachefile, and if so use the fingerprint to let libdivecomputer potentially stop downloading dives early. This doesn't much matter for most dive computers, but some (like the Scubapro G2) are not able to download one dive at a time, and need the fingerprint to avoid doing a full dump. That is particularly noticeable over bluetooth, where a full dump can be very slow. NOTE! The fingerprint cache is a separate entity from the dive log itself. Unlike the dive log, it doesn't synchronize over the cloud, so if you download using different clients (say, your phone and your laptop), the fingerprint cache entries are per device. So you may still end up downloading dives you already have, because the fingerprint code basically only works to avoid duplicate downloads on the same installation. Also, note that we only have a cache of one single entry per dive computer and downloader, so if you download dives and then don't save the end result, the fingerprint will now point to a dive that you don't actually have in your dive list. As a result, next time you download, the fingerprint won't match any existing dive, and we'll resort to the old non-optimized behavior. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-04-24 17:37:31 +00:00
*/
static void verify_fingerprint(dc_device_t *device, device_data_t *devdata, const unsigned char *buffer, size_t size)
{
Re-do the libdivecomputer fingerprint save/load code This tries to make our fingerprinting code work better, by avoiding using the "deviceid" field that has always been unreliable because we've calculated it multiple different ways, and even for the same version of subsurface, it ends up changing in the middle (ie we calculate one value initially, then re-calculate it when we have a proper serial number string). So instead, the fingerprinting code will look up and save the fingerprint file using purely "stable" information that is available early during the download: - the device model name (which is a string with vendor and product name separated by a space) - the DC_EVENT_DEVINFO 32-bit 'serial' number (which is not necessarily a real serial number at all, but hopefully at least a unique number for the particular product) but because the model name is not necessarily a good filename (think slashes and other possibly invalid characters), we hash that model name and use the resulting hex number in the fingerprint file name. This way the fingerprint file is unambiguous at load and save time, and depends purely on libdivecomputer data. But because we also need to verify that we have the actual _dive_ associated with that fingerprint, we also need to save the final deviceid and diveid when saving the fingerprint file, so that when we load it again we can look up the dive and verify that we have it before we use the fingerprint data. To do that, the fingerprint file itself contains not just the fingerprint data from libdivecomputer, but the last 8 bytes of the file are the (subsurface) deviceid and the diveid of the dive that is associated with the fingerprint. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2021-09-19 01:52:50 +00:00
uint32_t diveid, deviceid;
use libdivecomputer 'fingerprint' to avoid downloading extra data This opportunistically uses a cache of 'fingerprints' for already downloaded dives. As we download data from a dive computer, we save the fingerprint and dive ID of the most recent dive in a per-divecopmputer fingerprint cache file. The next time we download from that dive computer, we will load the cache file for that dive computer if it exists, verify that we still have the dive that is referenced in that cachefile, and if so use the fingerprint to let libdivecomputer potentially stop downloading dives early. This doesn't much matter for most dive computers, but some (like the Scubapro G2) are not able to download one dive at a time, and need the fingerprint to avoid doing a full dump. That is particularly noticeable over bluetooth, where a full dump can be very slow. NOTE! The fingerprint cache is a separate entity from the dive log itself. Unlike the dive log, it doesn't synchronize over the cloud, so if you download using different clients (say, your phone and your laptop), the fingerprint cache entries are per device. So you may still end up downloading dives you already have, because the fingerprint code basically only works to avoid duplicate downloads on the same installation. Also, note that we only have a cache of one single entry per dive computer and downloader, so if you download dives and then don't save the end result, the fingerprint will now point to a dive that you don't actually have in your dive list. As a result, next time you download, the fingerprint won't match any existing dive, and we'll resort to the old non-optimized behavior. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-04-24 17:37:31 +00:00
Re-do the libdivecomputer fingerprint save/load code This tries to make our fingerprinting code work better, by avoiding using the "deviceid" field that has always been unreliable because we've calculated it multiple different ways, and even for the same version of subsurface, it ends up changing in the middle (ie we calculate one value initially, then re-calculate it when we have a proper serial number string). So instead, the fingerprinting code will look up and save the fingerprint file using purely "stable" information that is available early during the download: - the device model name (which is a string with vendor and product name separated by a space) - the DC_EVENT_DEVINFO 32-bit 'serial' number (which is not necessarily a real serial number at all, but hopefully at least a unique number for the particular product) but because the model name is not necessarily a good filename (think slashes and other possibly invalid characters), we hash that model name and use the resulting hex number in the fingerprint file name. This way the fingerprint file is unambiguous at load and save time, and depends purely on libdivecomputer data. But because we also need to verify that we have the actual _dive_ associated with that fingerprint, we also need to save the final deviceid and diveid when saving the fingerprint file, so that when we load it again we can look up the dive and verify that we have it before we use the fingerprint data. To do that, the fingerprint file itself contains not just the fingerprint data from libdivecomputer, but the last 8 bytes of the file are the (subsurface) deviceid and the diveid of the dive that is associated with the fingerprint. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2021-09-19 01:52:50 +00:00
if (size <= 8)
use libdivecomputer 'fingerprint' to avoid downloading extra data This opportunistically uses a cache of 'fingerprints' for already downloaded dives. As we download data from a dive computer, we save the fingerprint and dive ID of the most recent dive in a per-divecopmputer fingerprint cache file. The next time we download from that dive computer, we will load the cache file for that dive computer if it exists, verify that we still have the dive that is referenced in that cachefile, and if so use the fingerprint to let libdivecomputer potentially stop downloading dives early. This doesn't much matter for most dive computers, but some (like the Scubapro G2) are not able to download one dive at a time, and need the fingerprint to avoid doing a full dump. That is particularly noticeable over bluetooth, where a full dump can be very slow. NOTE! The fingerprint cache is a separate entity from the dive log itself. Unlike the dive log, it doesn't synchronize over the cloud, so if you download using different clients (say, your phone and your laptop), the fingerprint cache entries are per device. So you may still end up downloading dives you already have, because the fingerprint code basically only works to avoid duplicate downloads on the same installation. Also, note that we only have a cache of one single entry per dive computer and downloader, so if you download dives and then don't save the end result, the fingerprint will now point to a dive that you don't actually have in your dive list. As a result, next time you download, the fingerprint won't match any existing dive, and we'll resort to the old non-optimized behavior. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-04-24 17:37:31 +00:00
return;
Re-do the libdivecomputer fingerprint save/load code This tries to make our fingerprinting code work better, by avoiding using the "deviceid" field that has always been unreliable because we've calculated it multiple different ways, and even for the same version of subsurface, it ends up changing in the middle (ie we calculate one value initially, then re-calculate it when we have a proper serial number string). So instead, the fingerprinting code will look up and save the fingerprint file using purely "stable" information that is available early during the download: - the device model name (which is a string with vendor and product name separated by a space) - the DC_EVENT_DEVINFO 32-bit 'serial' number (which is not necessarily a real serial number at all, but hopefully at least a unique number for the particular product) but because the model name is not necessarily a good filename (think slashes and other possibly invalid characters), we hash that model name and use the resulting hex number in the fingerprint file name. This way the fingerprint file is unambiguous at load and save time, and depends purely on libdivecomputer data. But because we also need to verify that we have the actual _dive_ associated with that fingerprint, we also need to save the final deviceid and diveid when saving the fingerprint file, so that when we load it again we can look up the dive and verify that we have it before we use the fingerprint data. To do that, the fingerprint file itself contains not just the fingerprint data from libdivecomputer, but the last 8 bytes of the file are the (subsurface) deviceid and the diveid of the dive that is associated with the fingerprint. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2021-09-19 01:52:50 +00:00
size -= 8;
use libdivecomputer 'fingerprint' to avoid downloading extra data This opportunistically uses a cache of 'fingerprints' for already downloaded dives. As we download data from a dive computer, we save the fingerprint and dive ID of the most recent dive in a per-divecopmputer fingerprint cache file. The next time we download from that dive computer, we will load the cache file for that dive computer if it exists, verify that we still have the dive that is referenced in that cachefile, and if so use the fingerprint to let libdivecomputer potentially stop downloading dives early. This doesn't much matter for most dive computers, but some (like the Scubapro G2) are not able to download one dive at a time, and need the fingerprint to avoid doing a full dump. That is particularly noticeable over bluetooth, where a full dump can be very slow. NOTE! The fingerprint cache is a separate entity from the dive log itself. Unlike the dive log, it doesn't synchronize over the cloud, so if you download using different clients (say, your phone and your laptop), the fingerprint cache entries are per device. So you may still end up downloading dives you already have, because the fingerprint code basically only works to avoid duplicate downloads on the same installation. Also, note that we only have a cache of one single entry per dive computer and downloader, so if you download dives and then don't save the end result, the fingerprint will now point to a dive that you don't actually have in your dive list. As a result, next time you download, the fingerprint won't match any existing dive, and we'll resort to the old non-optimized behavior. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-04-24 17:37:31 +00:00
/* Get the dive ID from the end of the fingerprint cache file.. */
Re-do the libdivecomputer fingerprint save/load code This tries to make our fingerprinting code work better, by avoiding using the "deviceid" field that has always been unreliable because we've calculated it multiple different ways, and even for the same version of subsurface, it ends up changing in the middle (ie we calculate one value initially, then re-calculate it when we have a proper serial number string). So instead, the fingerprinting code will look up and save the fingerprint file using purely "stable" information that is available early during the download: - the device model name (which is a string with vendor and product name separated by a space) - the DC_EVENT_DEVINFO 32-bit 'serial' number (which is not necessarily a real serial number at all, but hopefully at least a unique number for the particular product) but because the model name is not necessarily a good filename (think slashes and other possibly invalid characters), we hash that model name and use the resulting hex number in the fingerprint file name. This way the fingerprint file is unambiguous at load and save time, and depends purely on libdivecomputer data. But because we also need to verify that we have the actual _dive_ associated with that fingerprint, we also need to save the final deviceid and diveid when saving the fingerprint file, so that when we load it again we can look up the dive and verify that we have it before we use the fingerprint data. To do that, the fingerprint file itself contains not just the fingerprint data from libdivecomputer, but the last 8 bytes of the file are the (subsurface) deviceid and the diveid of the dive that is associated with the fingerprint. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2021-09-19 01:52:50 +00:00
memcpy(&deviceid, buffer + size, 4);
memcpy(&diveid, buffer + size + 4, 4);
use libdivecomputer 'fingerprint' to avoid downloading extra data This opportunistically uses a cache of 'fingerprints' for already downloaded dives. As we download data from a dive computer, we save the fingerprint and dive ID of the most recent dive in a per-divecopmputer fingerprint cache file. The next time we download from that dive computer, we will load the cache file for that dive computer if it exists, verify that we still have the dive that is referenced in that cachefile, and if so use the fingerprint to let libdivecomputer potentially stop downloading dives early. This doesn't much matter for most dive computers, but some (like the Scubapro G2) are not able to download one dive at a time, and need the fingerprint to avoid doing a full dump. That is particularly noticeable over bluetooth, where a full dump can be very slow. NOTE! The fingerprint cache is a separate entity from the dive log itself. Unlike the dive log, it doesn't synchronize over the cloud, so if you download using different clients (say, your phone and your laptop), the fingerprint cache entries are per device. So you may still end up downloading dives you already have, because the fingerprint code basically only works to avoid duplicate downloads on the same installation. Also, note that we only have a cache of one single entry per dive computer and downloader, so if you download dives and then don't save the end result, the fingerprint will now point to a dive that you don't actually have in your dive list. As a result, next time you download, the fingerprint won't match any existing dive, and we'll resort to the old non-optimized behavior. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-04-24 17:37:31 +00:00
if (verbose)
dev_info(" ... fingerprinted dive %08x:%08x", deviceid, diveid);
use libdivecomputer 'fingerprint' to avoid downloading extra data This opportunistically uses a cache of 'fingerprints' for already downloaded dives. As we download data from a dive computer, we save the fingerprint and dive ID of the most recent dive in a per-divecopmputer fingerprint cache file. The next time we download from that dive computer, we will load the cache file for that dive computer if it exists, verify that we still have the dive that is referenced in that cachefile, and if so use the fingerprint to let libdivecomputer potentially stop downloading dives early. This doesn't much matter for most dive computers, but some (like the Scubapro G2) are not able to download one dive at a time, and need the fingerprint to avoid doing a full dump. That is particularly noticeable over bluetooth, where a full dump can be very slow. NOTE! The fingerprint cache is a separate entity from the dive log itself. Unlike the dive log, it doesn't synchronize over the cloud, so if you download using different clients (say, your phone and your laptop), the fingerprint cache entries are per device. So you may still end up downloading dives you already have, because the fingerprint code basically only works to avoid duplicate downloads on the same installation. Also, note that we only have a cache of one single entry per dive computer and downloader, so if you download dives and then don't save the end result, the fingerprint will now point to a dive that you don't actually have in your dive list. As a result, next time you download, the fingerprint won't match any existing dive, and we'll resort to the old non-optimized behavior. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-04-24 17:37:31 +00:00
/* Only use it if we *have* that dive! */
if (!divelog.dives.has_dive(deviceid, diveid)) {
if (verbose)
dev_info(" ... dive not found");
Re-do the libdivecomputer fingerprint save/load code This tries to make our fingerprinting code work better, by avoiding using the "deviceid" field that has always been unreliable because we've calculated it multiple different ways, and even for the same version of subsurface, it ends up changing in the middle (ie we calculate one value initially, then re-calculate it when we have a proper serial number string). So instead, the fingerprinting code will look up and save the fingerprint file using purely "stable" information that is available early during the download: - the device model name (which is a string with vendor and product name separated by a space) - the DC_EVENT_DEVINFO 32-bit 'serial' number (which is not necessarily a real serial number at all, but hopefully at least a unique number for the particular product) but because the model name is not necessarily a good filename (think slashes and other possibly invalid characters), we hash that model name and use the resulting hex number in the fingerprint file name. This way the fingerprint file is unambiguous at load and save time, and depends purely on libdivecomputer data. But because we also need to verify that we have the actual _dive_ associated with that fingerprint, we also need to save the final deviceid and diveid when saving the fingerprint file, so that when we load it again we can look up the dive and verify that we have it before we use the fingerprint data. To do that, the fingerprint file itself contains not just the fingerprint data from libdivecomputer, but the last 8 bytes of the file are the (subsurface) deviceid and the diveid of the dive that is associated with the fingerprint. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2021-09-19 01:52:50 +00:00
return;
}
dc_device_set_fingerprint(device, buffer, size);
if (verbose)
dev_info(" ... fingerprint of size %zu", size);
use libdivecomputer 'fingerprint' to avoid downloading extra data This opportunistically uses a cache of 'fingerprints' for already downloaded dives. As we download data from a dive computer, we save the fingerprint and dive ID of the most recent dive in a per-divecopmputer fingerprint cache file. The next time we download from that dive computer, we will load the cache file for that dive computer if it exists, verify that we still have the dive that is referenced in that cachefile, and if so use the fingerprint to let libdivecomputer potentially stop downloading dives early. This doesn't much matter for most dive computers, but some (like the Scubapro G2) are not able to download one dive at a time, and need the fingerprint to avoid doing a full dump. That is particularly noticeable over bluetooth, where a full dump can be very slow. NOTE! The fingerprint cache is a separate entity from the dive log itself. Unlike the dive log, it doesn't synchronize over the cloud, so if you download using different clients (say, your phone and your laptop), the fingerprint cache entries are per device. So you may still end up downloading dives you already have, because the fingerprint code basically only works to avoid duplicate downloads on the same installation. Also, note that we only have a cache of one single entry per dive computer and downloader, so if you download dives and then don't save the end result, the fingerprint will now point to a dive that you don't actually have in your dive list. As a result, next time you download, the fingerprint won't match any existing dive, and we'll resort to the old non-optimized behavior. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-04-24 17:37:31 +00:00
}
/*
* Look up the fingerprint from the fingerprint caches, and
* give it to libdivecomputer to avoid downloading already
* downloaded dives.
*/
static void lookup_fingerprint(dc_device_t *device, device_data_t *devdata)
{
if (devdata->force_download)
return;
Re-do the libdivecomputer fingerprint save/load code This tries to make our fingerprinting code work better, by avoiding using the "deviceid" field that has always been unreliable because we've calculated it multiple different ways, and even for the same version of subsurface, it ends up changing in the middle (ie we calculate one value initially, then re-calculate it when we have a proper serial number string). So instead, the fingerprinting code will look up and save the fingerprint file using purely "stable" information that is available early during the download: - the device model name (which is a string with vendor and product name separated by a space) - the DC_EVENT_DEVINFO 32-bit 'serial' number (which is not necessarily a real serial number at all, but hopefully at least a unique number for the particular product) but because the model name is not necessarily a good filename (think slashes and other possibly invalid characters), we hash that model name and use the resulting hex number in the fingerprint file name. This way the fingerprint file is unambiguous at load and save time, and depends purely on libdivecomputer data. But because we also need to verify that we have the actual _dive_ associated with that fingerprint, we also need to save the final deviceid and diveid when saving the fingerprint file, so that when we load it again we can look up the dive and verify that we have it before we use the fingerprint data. To do that, the fingerprint file itself contains not just the fingerprint data from libdivecomputer, but the last 8 bytes of the file are the (subsurface) deviceid and the diveid of the dive that is associated with the fingerprint. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2021-09-19 01:52:50 +00:00
/* first try our in memory data - raw_data is owned by the table, the dc_device_set_fingerprint function copies the data */
auto [fsize, raw_data] = get_fingerprint_data(fingerprints, calculate_string_hash(devdata->model.c_str()), devdata->devinfo.serial);
if (fsize) {
if (verbose)
dev_info("... found fingerprint in dive table");
dc_device_set_fingerprint(device, raw_data, fsize);
return;
}
/* now check if we have a fingerprint on disk */
std::string cachename = fingerprint_file(devdata);
if (verbose)
dev_info("Looking for fingerprint in '%s'", cachename.c_str());
auto [mem, err] = readfile(cachename.c_str());
if (err > 0) {
if (verbose)
dev_info(" ... got %zu bytes", mem.size());
verify_fingerprint(device, devdata, (unsigned char *)mem.data(), mem.size());
use libdivecomputer 'fingerprint' to avoid downloading extra data This opportunistically uses a cache of 'fingerprints' for already downloaded dives. As we download data from a dive computer, we save the fingerprint and dive ID of the most recent dive in a per-divecopmputer fingerprint cache file. The next time we download from that dive computer, we will load the cache file for that dive computer if it exists, verify that we still have the dive that is referenced in that cachefile, and if so use the fingerprint to let libdivecomputer potentially stop downloading dives early. This doesn't much matter for most dive computers, but some (like the Scubapro G2) are not able to download one dive at a time, and need the fingerprint to avoid doing a full dump. That is particularly noticeable over bluetooth, where a full dump can be very slow. NOTE! The fingerprint cache is a separate entity from the dive log itself. Unlike the dive log, it doesn't synchronize over the cloud, so if you download using different clients (say, your phone and your laptop), the fingerprint cache entries are per device. So you may still end up downloading dives you already have, because the fingerprint code basically only works to avoid duplicate downloads on the same installation. Also, note that we only have a cache of one single entry per dive computer and downloader, so if you download dives and then don't save the end result, the fingerprint will now point to a dive that you don't actually have in your dive list. As a result, next time you download, the fingerprint won't match any existing dive, and we'll resort to the old non-optimized behavior. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-04-24 17:37:31 +00:00
}
}
Assemble the actual Suunto serial number It turns out that the serial number returned by libdivecomputer isn't really the serial number as interpreted by the vendor. Those tend to be strings, but libdivecomputer gives us a 32bit number. Some experimenting showed that for the Suunto devies tested the serial number is encoded in that 32bit number: It so happens that the Suunto serial number strings are strings that have all numbers, but they aren't *one* number. They are four bytes representing two numbers each, and the "23500027" string is actually the four bytes 23 50 00 27 (0x17 0x32 0x00 0x1b). And libdivecomputer has incorrectly parsed those four bytes as one number, not as the encoded serial number string it is. So the value 389152795 is actually hex 0x1732001b, which is 0x17 0x32 0x00 0x1b, which is - 23 50 00 27. This should be done by libdivecomputer, but hey, in the meantime this at least shows the concept. And helps test the XML save/restore code. It depends on the two patches that create the whole "device.c" infrastructure, of course. With this, my dive file ends up having the settings section look like this: <divecomputerid model='Suunto Vyper Air' deviceid='d4629110' serial='01201094' firmware='1.1.22'/> <divecomputerid model='Suunto HelO2' deviceid='995dd566' serial='23500027' firmware='1.0.4'/> where the format of the firmware version is something I guessed at, but it was the obvious choice (again, it's byte-based, I'm ignoring the high byte that is zero for both of my Suuntos). Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
2013-01-10 00:14:21 +00:00
static void event_cb(dc_device_t *device, dc_event_type_t event, const void *data, void *userdata)
{
static unsigned int last = 0;
const dc_event_progress_t *progress = (dc_event_progress_t *)data;
const dc_event_devinfo_t *devinfo = (dc_event_devinfo_t *)data;
const dc_event_clock_t *clock = (dc_event_clock_t *)data;
const dc_event_vendor_t *vendor = (dc_event_vendor_t *)data;
device_data_t *devdata = (device_data_t *)userdata;
switch (event) {
case DC_EVENT_WAITING:
dev_info(translate("gettextFromC", "Event: waiting for user action"));
break;
case DC_EVENT_PROGRESS:
/* this seems really dumb... but having no idea what is happening on long
* downloads makes people think that the app is hung;
* since the progress is in bytes downloaded (usually), simply give updates in 10k increments
*/
if (progress->current < last)
/* this is a new communication with the divecomputer */
last = progress->current;
if (progress->current > last + 10240) {
last = progress->current;
dev_info(translate("gettextFromC", "read %dkb"), progress->current / 1024);
}
if (progress->maximum)
progress_bar_fraction = (double)progress->current / (double)progress->maximum;
break;
case DC_EVENT_DEVINFO:
if (dc_descriptor_get_model(devdata->descriptor) != devinfo->model) {
dc_descriptor_t *better_descriptor = get_descriptor(dc_descriptor_get_type(devdata->descriptor), devinfo->model);
if (better_descriptor != NULL) {
report_info("EVENT_DEVINFO gave us a different detected product (model %d instead of %d), which we are using now.",
devinfo->model, dc_descriptor_get_model(devdata->descriptor));
devdata->descriptor = better_descriptor;
devdata->product = dc_descriptor_get_product(better_descriptor);
devdata->vendor = dc_descriptor_get_vendor(better_descriptor);
devdata->model = devdata->vendor + " " + devdata->product;
} else {
report_info("EVENT_DEVINFO gave us a different detected product (model %d instead of %d), but that one is unknown.",
devinfo->model, dc_descriptor_get_model(devdata->descriptor));
}
}
dev_info(translate("gettextFromC", "model=%s firmware=%u serial=%u"),
devdata->product.c_str(), devinfo->firmware, devinfo->serial);
if (devdata->libdc_logfile) {
fprintf(devdata->libdc_logfile, "Event: model=%u (0x%08x), firmware=%u (0x%08x), serial=%u (0x%08x)\n",
devinfo->model, devinfo->model,
devinfo->firmware, devinfo->firmware,
devinfo->serial, devinfo->serial);
}
use libdivecomputer 'fingerprint' to avoid downloading extra data This opportunistically uses a cache of 'fingerprints' for already downloaded dives. As we download data from a dive computer, we save the fingerprint and dive ID of the most recent dive in a per-divecopmputer fingerprint cache file. The next time we download from that dive computer, we will load the cache file for that dive computer if it exists, verify that we still have the dive that is referenced in that cachefile, and if so use the fingerprint to let libdivecomputer potentially stop downloading dives early. This doesn't much matter for most dive computers, but some (like the Scubapro G2) are not able to download one dive at a time, and need the fingerprint to avoid doing a full dump. That is particularly noticeable over bluetooth, where a full dump can be very slow. NOTE! The fingerprint cache is a separate entity from the dive log itself. Unlike the dive log, it doesn't synchronize over the cloud, so if you download using different clients (say, your phone and your laptop), the fingerprint cache entries are per device. So you may still end up downloading dives you already have, because the fingerprint code basically only works to avoid duplicate downloads on the same installation. Also, note that we only have a cache of one single entry per dive computer and downloader, so if you download dives and then don't save the end result, the fingerprint will now point to a dive that you don't actually have in your dive list. As a result, next time you download, the fingerprint won't match any existing dive, and we'll resort to the old non-optimized behavior. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-04-24 17:37:31 +00:00
Re-do the libdivecomputer fingerprint save/load code This tries to make our fingerprinting code work better, by avoiding using the "deviceid" field that has always been unreliable because we've calculated it multiple different ways, and even for the same version of subsurface, it ends up changing in the middle (ie we calculate one value initially, then re-calculate it when we have a proper serial number string). So instead, the fingerprinting code will look up and save the fingerprint file using purely "stable" information that is available early during the download: - the device model name (which is a string with vendor and product name separated by a space) - the DC_EVENT_DEVINFO 32-bit 'serial' number (which is not necessarily a real serial number at all, but hopefully at least a unique number for the particular product) but because the model name is not necessarily a good filename (think slashes and other possibly invalid characters), we hash that model name and use the resulting hex number in the fingerprint file name. This way the fingerprint file is unambiguous at load and save time, and depends purely on libdivecomputer data. But because we also need to verify that we have the actual _dive_ associated with that fingerprint, we also need to save the final deviceid and diveid when saving the fingerprint file, so that when we load it again we can look up the dive and verify that we have it before we use the fingerprint data. To do that, the fingerprint file itself contains not just the fingerprint data from libdivecomputer, but the last 8 bytes of the file are the (subsurface) deviceid and the diveid of the dive that is associated with the fingerprint. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2021-09-19 01:52:50 +00:00
devdata->devinfo = *devinfo;
use libdivecomputer 'fingerprint' to avoid downloading extra data This opportunistically uses a cache of 'fingerprints' for already downloaded dives. As we download data from a dive computer, we save the fingerprint and dive ID of the most recent dive in a per-divecopmputer fingerprint cache file. The next time we download from that dive computer, we will load the cache file for that dive computer if it exists, verify that we still have the dive that is referenced in that cachefile, and if so use the fingerprint to let libdivecomputer potentially stop downloading dives early. This doesn't much matter for most dive computers, but some (like the Scubapro G2) are not able to download one dive at a time, and need the fingerprint to avoid doing a full dump. That is particularly noticeable over bluetooth, where a full dump can be very slow. NOTE! The fingerprint cache is a separate entity from the dive log itself. Unlike the dive log, it doesn't synchronize over the cloud, so if you download using different clients (say, your phone and your laptop), the fingerprint cache entries are per device. So you may still end up downloading dives you already have, because the fingerprint code basically only works to avoid duplicate downloads on the same installation. Also, note that we only have a cache of one single entry per dive computer and downloader, so if you download dives and then don't save the end result, the fingerprint will now point to a dive that you don't actually have in your dive list. As a result, next time you download, the fingerprint won't match any existing dive, and we'll resort to the old non-optimized behavior. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-04-24 17:37:31 +00:00
lookup_fingerprint(device, devdata);
break;
case DC_EVENT_CLOCK:
dev_info(translate("gettextFromC", "Event: systime=%" PRId64 ", devtime=%u\n"),
(uint64_t)clock->systime, clock->devtime);
if (devdata->libdc_logfile) {
fprintf(devdata->libdc_logfile, "Event: systime=%" PRId64 ", devtime=%u\n",
(uint64_t)clock->systime, clock->devtime);
}
break;
case DC_EVENT_VENDOR:
if (devdata->libdc_logfile) {
fprintf(devdata->libdc_logfile, "Event: vendor=");
for (unsigned int i = 0; i < vendor->size; ++i)
fprintf(devdata->libdc_logfile, "%02X", vendor->data[i]);
fprintf(devdata->libdc_logfile, "\n");
}
break;
default:
break;
}
}
int import_thread_cancelled;
static int cancel_cb(void *)
{
return import_thread_cancelled;
}
static std::string do_device_import(device_data_t *data)
{
dc_status_t rc;
dc_device_t *device = data->device;
data->model = data->vendor + " " + data->product;
Assemble the actual Suunto serial number It turns out that the serial number returned by libdivecomputer isn't really the serial number as interpreted by the vendor. Those tend to be strings, but libdivecomputer gives us a 32bit number. Some experimenting showed that for the Suunto devies tested the serial number is encoded in that 32bit number: It so happens that the Suunto serial number strings are strings that have all numbers, but they aren't *one* number. They are four bytes representing two numbers each, and the "23500027" string is actually the four bytes 23 50 00 27 (0x17 0x32 0x00 0x1b). And libdivecomputer has incorrectly parsed those four bytes as one number, not as the encoded serial number string it is. So the value 389152795 is actually hex 0x1732001b, which is 0x17 0x32 0x00 0x1b, which is - 23 50 00 27. This should be done by libdivecomputer, but hey, in the meantime this at least shows the concept. And helps test the XML save/restore code. It depends on the two patches that create the whole "device.c" infrastructure, of course. With this, my dive file ends up having the settings section look like this: <divecomputerid model='Suunto Vyper Air' deviceid='d4629110' serial='01201094' firmware='1.1.22'/> <divecomputerid model='Suunto HelO2' deviceid='995dd566' serial='23500027' firmware='1.0.4'/> where the format of the firmware version is something I guessed at, but it was the obvious choice (again, it's byte-based, I'm ignoring the high byte that is zero for both of my Suuntos). Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
2013-01-10 00:14:21 +00:00
// Register the event handler.
int events = DC_EVENT_WAITING | DC_EVENT_PROGRESS | DC_EVENT_DEVINFO | DC_EVENT_CLOCK | DC_EVENT_VENDOR;
rc = dc_device_set_events(device, events, event_cb, data);
if (rc != DC_STATUS_SUCCESS) {
dev_info("Import error: %s", errmsg(rc));
return translate("gettextFromC", "Error registering the event handler.");
}
// Register the cancellation handler.
rc = dc_device_set_cancel(device, cancel_cb, data);
if (rc != DC_STATUS_SUCCESS) {
dev_info("Import error: %s", errmsg(rc));
return translate("gettextFromC", "Error registering the cancellation handler.");
}
if (data->libdc_dump) {
dc_buffer_t *buffer = dc_buffer_new(0);
rc = dc_device_dump(device, buffer);
if (rc == DC_STATUS_SUCCESS && !dumpfile_name.empty()) {
FILE *fp = subsurface_fopen(dumpfile_name.c_str(), "wb");
if (fp != NULL) {
fwrite(dc_buffer_get_data(buffer), 1, dc_buffer_get_size(buffer), fp);
fclose(fp);
}
}
dc_buffer_free(buffer);
if (rc != DC_STATUS_SUCCESS) {
progress_bar_fraction = 0.0;
if (rc == DC_STATUS_UNSUPPORTED)
return translate("gettextFromC", "Dumping not supported on this device");
dev_info("Import error: %s", errmsg(rc));
return translate("gettextFromC", "Dive data dumping error");
}
} else {
rc = dc_device_foreach(device, dive_cb, data);
if (rc != DC_STATUS_SUCCESS) {
progress_bar_fraction = 0.0;
dev_info("Import error: %s", errmsg(rc));
return translate("gettextFromC", "Dive data import error");
}
}
/* All good */
return std::string();
}
void logfunc(dc_context_t *, dc_loglevel_t loglevel, const char *file, unsigned int line, const char *function, const char *msg, void *userdata)
{
const char *loglevels[] = { "NONE", "ERROR", "WARNING", "INFO", "DEBUG", "ALL" };
static const auto start(std::chrono::steady_clock::now());
auto now(std::chrono::steady_clock::now());
double elapsed_seconds = std::chrono::duration<double>(now - start).count();
FILE *fp = (FILE *)userdata;
if (loglevel == DC_LOGLEVEL_ERROR || loglevel == DC_LOGLEVEL_WARNING) {
fprintf(fp, "[%.6f] %s: %s [in %s:%d (%s)]\n",
elapsed_seconds,
loglevels[loglevel], msg, file, line, function);
} else {
fprintf(fp, "[%6f] %s: %s\n", elapsed_seconds, loglevels[loglevel], msg);
}
}
/*
* Get the transports supported by us (as opposed to
* the list of transports supported by a particular
* dive computer).
*
* This could have various platform rules too..
*/
unsigned int get_supported_transports(device_data_t *data)
{
#if defined(Q_OS_IOS)
// BLE only - don't bother with being clever.
return DC_TRANSPORT_BLE;
#endif
// start out with the list of transports that libdivecomputer claims to support
// dc_context_get_transports ignores its context argument...
unsigned int supported = dc_context_get_transports(NULL);
// then add the ones that we have our own implementations for
#if defined(BT_SUPPORT)
supported |= DC_TRANSPORT_BLUETOOTH;
#endif
#if defined(BLE_SUPPORT)
supported |= DC_TRANSPORT_BLE;
#endif
#if defined(Q_OS_ANDROID)
// we cannot support transports that need libusb, hid, filesystem access, or IRDA on Android
supported &= ~(DC_TRANSPORT_USB | DC_TRANSPORT_USBHID | DC_TRANSPORT_IRDA | DC_TRANSPORT_USBSTORAGE);
#endif
if (data) {
/*
* If we have device data available, we can refine this:
* We don't support BT or BLE unless bluetooth_mode was set,
* and if it was we won't try any of the other transports.
*/
if (data->bluetooth_mode) {
supported &= (DC_TRANSPORT_BLUETOOTH | DC_TRANSPORT_BLE);
if (starts_with(data->devname, "LE:"))
supported &= DC_TRANSPORT_BLE;
} else {
supported &= ~(DC_TRANSPORT_BLUETOOTH | DC_TRANSPORT_BLE);
}
}
return supported;
}
static dc_status_t usbhid_device_open(dc_iostream_t **iostream, dc_context_t *context, device_data_t *data)
{
dc_status_t rc;
dc_iterator_t *iterator = NULL;
dc_usbhid_device_t *device = NULL;
// Discover the usbhid device.
dc_usbhid_iterator_new (&iterator, context, data->descriptor);
while (dc_iterator_next (iterator, &device) == DC_STATUS_SUCCESS)
break;
dc_iterator_free (iterator);
if (!device) {
ERROR("didn't find HID device");
return DC_STATUS_NODEVICE;
}
dev_info("Opening USB HID device for %04x:%04x",
dc_usbhid_device_get_vid(device),
dc_usbhid_device_get_pid(device));
rc = dc_usbhid_open(iostream, context, device);
dc_usbhid_device_free(device);
return rc;
}
static dc_status_t usb_device_open(dc_iostream_t **iostream, dc_context_t *context, device_data_t *data)
{
dc_status_t rc;
dc_iterator_t *iterator = NULL;
dc_usb_device_t *device = NULL;
// Discover the usb device.
dc_usb_iterator_new (&iterator, context, data->descriptor);
while (dc_iterator_next (iterator, &device) == DC_STATUS_SUCCESS)
break;
dc_iterator_free (iterator);
if (!device)
return DC_STATUS_NODEVICE;
dev_info("Opening USB device for %04x:%04x",
dc_usb_device_get_vid(device),
dc_usb_device_get_pid(device));
rc = dc_usb_open(iostream, context, device);
dc_usb_device_free(device);
return rc;
}
static dc_status_t irda_device_open(dc_iostream_t **iostream, dc_context_t *context, device_data_t *data)
{
unsigned int address = 0;
dc_iterator_t *iterator = NULL;
dc_irda_device_t *device = NULL;
// Try to find the IRDA address
dc_irda_iterator_new (&iterator, context, data->descriptor);
while (dc_iterator_next (iterator, &device) == DC_STATUS_SUCCESS) {
address = dc_irda_device_get_address (device);
dc_irda_device_free (device);
break;
}
dc_iterator_free (iterator);
// If that fails, use the device name. This will
// use address 0 if it's not a number.
if (!address)
std::from_chars(data->devname.c_str(), data->devname.c_str() + data->devname.size(), address);
dev_info("Opening IRDA address %u", address);
return dc_irda_open(&data->iostream, context, address, 1);
}
#if defined(BT_SUPPORT) && !defined(__ANDROID__) && !defined(__APPLE__)
static dc_status_t bluetooth_device_open(dc_context_t *context, device_data_t *data)
{
dc_bluetooth_address_t address = 0;
dc_iterator_t *iterator = NULL;
dc_bluetooth_device_t *device = NULL;
// Try to find the rfcomm device address
dc_bluetooth_iterator_new (&iterator, context, data->descriptor);
while (dc_iterator_next (iterator, &device) == DC_STATUS_SUCCESS) {
address = dc_bluetooth_device_get_address (device);
dc_bluetooth_device_free (device);
break;
}
dc_iterator_free (iterator);
if (!address) {
dev_info("No rfcomm device found");
return DC_STATUS_NODEVICE;
}
dev_info("Opening rfcomm address %llu", address);
return dc_bluetooth_open(&data->iostream, context, address, 0);
}
#endif
dc_status_t divecomputer_device_open(device_data_t *data)
{
dc_status_t rc = DC_STATUS_UNSUPPORTED;
dc_context_t *context = data->context;
unsigned int transports, supported;
transports = dc_descriptor_get_transports(data->descriptor);
supported = get_supported_transports(data);
transports &= supported;
if (!transports) {
dev_info("Dive computer transport not supported");
return DC_STATUS_UNSUPPORTED;
}
#ifdef BT_SUPPORT
if (transports & DC_TRANSPORT_BLUETOOTH) {
dev_info("Opening rfcomm stream %s", data->devname.c_str());
#if defined(__ANDROID__) || defined(__APPLE__)
// we don't have BT on iOS in the first place, so this is for Android and macOS
rc = rfcomm_stream_open(&data->iostream, context, data->devname.c_str());
#else
rc = bluetooth_device_open(context, data);
#endif
if (rc == DC_STATUS_SUCCESS)
return rc;
}
#endif
#ifdef BLE_SUPPORT
if (transports & DC_TRANSPORT_BLE) {
dev_info("Connecting to BLE device %s", data->devname.c_str());
rc = ble_packet_open(&data->iostream, context, data->devname.c_str(), data);
if (rc == DC_STATUS_SUCCESS)
return rc;
}
#endif
if (transports & DC_TRANSPORT_USBHID) {
dev_info("Connecting to USB HID device");
rc = usbhid_device_open(&data->iostream, context, data);
if (rc == DC_STATUS_SUCCESS)
return rc;
}
if (transports & DC_TRANSPORT_USB) {
dev_info("Connecting to native USB device");
rc = usb_device_open(&data->iostream, context, data);
if (rc == DC_STATUS_SUCCESS)
return rc;
}
if (transports & DC_TRANSPORT_SERIAL) {
dev_info("Opening serial device %s", data->devname.c_str());
#ifdef SERIAL_FTDI
if (!strcasecmp(data->devname.c_str(), "ftdi"))
return ftdi_open(&data->iostream, context);
#endif
#ifdef __ANDROID__
if (data->androidUsbDeviceDescriptor)
return serial_usb_android_open(&data->iostream, context, data->androidUsbDeviceDescriptor);
#endif
rc = dc_serial_open(&data->iostream, context, data->devname.c_str());
if (rc == DC_STATUS_SUCCESS)
return rc;
}
if (transports & DC_TRANSPORT_IRDA) {
dev_info("Connecting to IRDA device");
rc = irda_device_open(&data->iostream, context, data);
if (rc == DC_STATUS_SUCCESS)
return rc;
}
if (transports & DC_TRANSPORT_USBSTORAGE) {
dev_info("Opening USB storage at %s", data->devname.c_str());
rc = dc_usb_storage_open(&data->iostream, context, data->devname.c_str());
if (rc == DC_STATUS_SUCCESS)
return rc;
}
return rc;
}
static dc_status_t sync_divecomputer_time(dc_device_t *device)
{
dc_datetime_t now;
dc_datetime_localtime(&now, dc_datetime_now());
return dc_device_timesync(device, &now);
}
std::string do_libdivecomputer_import(device_data_t *data)
{
dc_status_t rc;
FILE *fp = NULL;
import_dive_number = 0;
first_temp_is_air = 0;
data->device = NULL;
data->context = NULL;
data->iostream = NULL;
use libdivecomputer 'fingerprint' to avoid downloading extra data This opportunistically uses a cache of 'fingerprints' for already downloaded dives. As we download data from a dive computer, we save the fingerprint and dive ID of the most recent dive in a per-divecopmputer fingerprint cache file. The next time we download from that dive computer, we will load the cache file for that dive computer if it exists, verify that we still have the dive that is referenced in that cachefile, and if so use the fingerprint to let libdivecomputer potentially stop downloading dives early. This doesn't much matter for most dive computers, but some (like the Scubapro G2) are not able to download one dive at a time, and need the fingerprint to avoid doing a full dump. That is particularly noticeable over bluetooth, where a full dump can be very slow. NOTE! The fingerprint cache is a separate entity from the dive log itself. Unlike the dive log, it doesn't synchronize over the cloud, so if you download using different clients (say, your phone and your laptop), the fingerprint cache entries are per device. So you may still end up downloading dives you already have, because the fingerprint code basically only works to avoid duplicate downloads on the same installation. Also, note that we only have a cache of one single entry per dive computer and downloader, so if you download dives and then don't save the end result, the fingerprint will now point to a dive that you don't actually have in your dive list. As a result, next time you download, the fingerprint won't match any existing dive, and we'll resort to the old non-optimized behavior. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-04-24 17:37:31 +00:00
data->fingerprint = NULL;
data->fsize = 0;
if (data->libdc_log && !logfile_name.empty())
fp = subsurface_fopen(logfile_name.c_str(), "w");
data->libdc_logfile = fp;
rc = dc_context_new(&data->context);
if (rc != DC_STATUS_SUCCESS)
return translate("gettextFromC", "Unable to create libdivecomputer context");
if (fp) {
dc_context_set_loglevel(data->context, DC_LOGLEVEL_ALL);
dc_context_set_logfunc(data->context, logfunc, fp);
fprintf(data->libdc_logfile, "Subsurface: v%s, ", subsurface_git_version());
fprintf(data->libdc_logfile, "built with libdivecomputer v%s\n", dc_version(NULL));
}
std::string err = translate("gettextFromC", "Unable to open %s %s (%s)");
rc = divecomputer_device_open(data);
if (rc != DC_STATUS_SUCCESS) {
dev_info("Import error: %s", errmsg(rc));
} else {
dev_info("Connecting ...");
rc = dc_device_open(&data->device, data->context, data->descriptor, data->iostream);
if (rc != DC_STATUS_SUCCESS) {
INFO("dc_device_open error value of %d", rc);
if (subsurface_access(data->devname.c_str(), R_OK | W_OK) != 0)
#if defined(SUBSURFACE_MOBILE)
err = translate("gettextFromC", "Error opening the device %s %s (%s).\nIn most cases, in order to debug this issue, it is useful to send the developers the log files. You can copy them to the clipboard in the About dialog.");
#else
err = translate("gettextFromC", "Error opening the device %s %s (%s).\nIn most cases, in order to debug this issue, a libdivecomputer logfile will be useful.\nYou can create this logfile by selecting the corresponding checkbox in the download dialog.");
#endif
} else {
dev_info("Starting import ...");
err = do_device_import(data);
/* TODO: Show the logfile to the user on error. */
dev_info("Import complete");
if (err.empty() && data->sync_time) {
dev_info("Syncing dive computer time ...");
rc = sync_divecomputer_time(data->device);
switch (rc) {
case DC_STATUS_SUCCESS:
dev_info("Time sync complete");
break;
case DC_STATUS_UNSUPPORTED:
dev_info("Time sync not supported by dive computer");
break;
default:
dev_info("Time sync failed");
break;
}
}
dc_device_close(data->device);
data->device = NULL;
if (data->log->dives.empty())
dev_info(translate("gettextFromC", "No new dives downloaded from dive computer"));
}
dc_iostream_close(data->iostream);
data->iostream = NULL;
}
dc_context_free(data->context);
data->context = NULL;
if (fp) {
fclose(fp);
}
use libdivecomputer 'fingerprint' to avoid downloading extra data This opportunistically uses a cache of 'fingerprints' for already downloaded dives. As we download data from a dive computer, we save the fingerprint and dive ID of the most recent dive in a per-divecopmputer fingerprint cache file. The next time we download from that dive computer, we will load the cache file for that dive computer if it exists, verify that we still have the dive that is referenced in that cachefile, and if so use the fingerprint to let libdivecomputer potentially stop downloading dives early. This doesn't much matter for most dive computers, but some (like the Scubapro G2) are not able to download one dive at a time, and need the fingerprint to avoid doing a full dump. That is particularly noticeable over bluetooth, where a full dump can be very slow. NOTE! The fingerprint cache is a separate entity from the dive log itself. Unlike the dive log, it doesn't synchronize over the cloud, so if you download using different clients (say, your phone and your laptop), the fingerprint cache entries are per device. So you may still end up downloading dives you already have, because the fingerprint code basically only works to avoid duplicate downloads on the same installation. Also, note that we only have a cache of one single entry per dive computer and downloader, so if you download dives and then don't save the end result, the fingerprint will now point to a dive that you don't actually have in your dive list. As a result, next time you download, the fingerprint won't match any existing dive, and we'll resort to the old non-optimized behavior. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-04-24 17:37:31 +00:00
/*
* Note that we save the fingerprint unconditionally.
* This is ok because we only have fingerprint data if
* we got a dive header, and because we will use the
* dive id to verify that we actually have the dive
* it refers to before we use the fingerprint data.
*
* For now we save the fingerprint both to the local file system
* and to the global fingerprint table (to be then saved out with
* the dive log data).
use libdivecomputer 'fingerprint' to avoid downloading extra data This opportunistically uses a cache of 'fingerprints' for already downloaded dives. As we download data from a dive computer, we save the fingerprint and dive ID of the most recent dive in a per-divecopmputer fingerprint cache file. The next time we download from that dive computer, we will load the cache file for that dive computer if it exists, verify that we still have the dive that is referenced in that cachefile, and if so use the fingerprint to let libdivecomputer potentially stop downloading dives early. This doesn't much matter for most dive computers, but some (like the Scubapro G2) are not able to download one dive at a time, and need the fingerprint to avoid doing a full dump. That is particularly noticeable over bluetooth, where a full dump can be very slow. NOTE! The fingerprint cache is a separate entity from the dive log itself. Unlike the dive log, it doesn't synchronize over the cloud, so if you download using different clients (say, your phone and your laptop), the fingerprint cache entries are per device. So you may still end up downloading dives you already have, because the fingerprint code basically only works to avoid duplicate downloads on the same installation. Also, note that we only have a cache of one single entry per dive computer and downloader, so if you download dives and then don't save the end result, the fingerprint will now point to a dive that you don't actually have in your dive list. As a result, next time you download, the fingerprint won't match any existing dive, and we'll resort to the old non-optimized behavior. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-04-24 17:37:31 +00:00
*/
save_fingerprint(data);
if (data->fingerprint && data->fdiveid)
create_fingerprint_node(fingerprints, calculate_string_hash(data->model.c_str()), data->devinfo.serial,
data->fingerprint, data->fsize, data->fdeviceid, data->fdiveid);
Re-do the libdivecomputer fingerprint save/load code This tries to make our fingerprinting code work better, by avoiding using the "deviceid" field that has always been unreliable because we've calculated it multiple different ways, and even for the same version of subsurface, it ends up changing in the middle (ie we calculate one value initially, then re-calculate it when we have a proper serial number string). So instead, the fingerprinting code will look up and save the fingerprint file using purely "stable" information that is available early during the download: - the device model name (which is a string with vendor and product name separated by a space) - the DC_EVENT_DEVINFO 32-bit 'serial' number (which is not necessarily a real serial number at all, but hopefully at least a unique number for the particular product) but because the model name is not necessarily a good filename (think slashes and other possibly invalid characters), we hash that model name and use the resulting hex number in the fingerprint file name. This way the fingerprint file is unambiguous at load and save time, and depends purely on libdivecomputer data. But because we also need to verify that we have the actual _dive_ associated with that fingerprint, we also need to save the final deviceid and diveid when saving the fingerprint file, so that when we load it again we can look up the dive and verify that we have it before we use the fingerprint data. To do that, the fingerprint file itself contains not just the fingerprint data from libdivecomputer, but the last 8 bytes of the file are the (subsurface) deviceid and the diveid of the dive that is associated with the fingerprint. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2021-09-19 01:52:50 +00:00
free(data->fingerprint);
data->fingerprint = NULL;
use libdivecomputer 'fingerprint' to avoid downloading extra data This opportunistically uses a cache of 'fingerprints' for already downloaded dives. As we download data from a dive computer, we save the fingerprint and dive ID of the most recent dive in a per-divecopmputer fingerprint cache file. The next time we download from that dive computer, we will load the cache file for that dive computer if it exists, verify that we still have the dive that is referenced in that cachefile, and if so use the fingerprint to let libdivecomputer potentially stop downloading dives early. This doesn't much matter for most dive computers, but some (like the Scubapro G2) are not able to download one dive at a time, and need the fingerprint to avoid doing a full dump. That is particularly noticeable over bluetooth, where a full dump can be very slow. NOTE! The fingerprint cache is a separate entity from the dive log itself. Unlike the dive log, it doesn't synchronize over the cloud, so if you download using different clients (say, your phone and your laptop), the fingerprint cache entries are per device. So you may still end up downloading dives you already have, because the fingerprint code basically only works to avoid duplicate downloads on the same installation. Also, note that we only have a cache of one single entry per dive computer and downloader, so if you download dives and then don't save the end result, the fingerprint will now point to a dive that you don't actually have in your dive list. As a result, next time you download, the fingerprint won't match any existing dive, and we'll resort to the old non-optimized behavior. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-04-24 17:37:31 +00:00
return err;
}
/*
* Parse data buffers instead of dc devices downloaded data.
* Intended to be used to parse profile data from binary files during import tasks.
* Actually included Uwatec families because of works on datatrak and smartrak logs
* and OSTC families for OSTCTools logs import.
* For others, simply include them in the switch (check parameters).
* Note that dc_descriptor_t in data *must* have been filled using dc_descriptor_iterator()
* calls.
*/
dc_status_t libdc_buffer_parser(struct dive *dive, device_data_t *data, unsigned char *buffer, int size)
{
dc_status_t rc;
dc_parser_t *parser = NULL;
switch (dc_descriptor_get_type(data->descriptor)) {
case DC_FAMILY_UWATEC_ALADIN:
case DC_FAMILY_UWATEC_MEMOMOUSE:
case DC_FAMILY_UWATEC_SMART:
case DC_FAMILY_UWATEC_MERIDIAN:
case DC_FAMILY_HW_OSTC:
case DC_FAMILY_HW_FROG:
case DC_FAMILY_HW_OSTC3:
rc = dc_parser_new2(&parser, data->context, data->descriptor, buffer, size);
break;
default:
report_error("Device type not handled!");
return DC_STATUS_UNSUPPORTED;
}
if (rc != DC_STATUS_SUCCESS) {
report_error("Error creating parser.");
dc_parser_destroy (parser);
return rc;
}
// Do not parse Aladin/Memomouse headers as they are fakes
// Do not return on error, we can still parse the samples
if (dc_descriptor_get_type(data->descriptor) != DC_FAMILY_UWATEC_ALADIN && dc_descriptor_get_type(data->descriptor) != DC_FAMILY_UWATEC_MEMOMOUSE) {
rc = libdc_header_parser (parser, data, dive);
if (rc != DC_STATUS_SUCCESS) {
report_error("Error parsing the dive header data. Dive # %d: %s", dive->number, errmsg(rc));
}
}
rc = dc_parser_samples_foreach (parser, sample_cb, &dive->dcs[0]);
if (rc != DC_STATUS_SUCCESS) {
report_error("Error parsing the sample data. Dive # %d: %s", dive->number, errmsg(rc));
dc_parser_destroy (parser);
return rc;
}
dc_parser_destroy(parser);
return DC_STATUS_SUCCESS;
}
/*
* Returns a dc_descriptor_t structure based on dc model's number and family.
*
* That dc_descriptor_t needs to be freed with dc_descriptor_free by the reciver.
*/
dc_descriptor_t *get_descriptor(dc_family_t type, unsigned int model)
{
dc_descriptor_t *descriptor = NULL, *needle = NULL;
dc_iterator_t *iterator = NULL;
dc_status_t rc;
rc = dc_descriptor_iterator(&iterator);
if (rc != DC_STATUS_SUCCESS) {
report_info("Error creating the device descriptor iterator: %s", errmsg(rc));
return NULL;
}
while ((dc_iterator_next(iterator, &descriptor)) == DC_STATUS_SUCCESS) {
unsigned int desc_model = dc_descriptor_get_model(descriptor);
dc_family_t desc_type = dc_descriptor_get_type(descriptor);
if (model == desc_model && type == desc_type) {
needle = descriptor;
break;
}
dc_descriptor_free(descriptor);
}
dc_iterator_free(iterator);
return needle;
}