<html> <head> <meta http-equiv="Content-Type" content="text/html; charset=Windows-1252"> <style type="text/css" style="display:none;"> P {margin-top:0;margin-bottom:0;} </style> </head> <body dir="ltr"> <div class="WordSection1"> <div> <a name="_olk_signature">Hi everyone. I have some questions with the memory management with libfprint.</a> I’m creating a program for capture image and retrieve the minutiae from the captured image. I’m using: <ul type="disc" style="margin-bottom: 0in;margin-top:0in"> <li class="MsoListParagraphCxSpFirst" style="margin: 0in 0in 0in 0.5in; font-size: 11pt; font-family: Calibri, sans-serif;margin-left:0in"> fp_device_capture</li><li class="MsoListParagraphCxSpMiddle" style="margin: 0in 0in 0in 0.5in; font-size: 11pt; font-family: Calibri, sans-serif;margin-left:0in"> fp_device_capture_finish</li><li class="MsoListParagraphCxSpMiddle" style="margin: 0in 0in 0in 0.5in; font-size: 11pt; font-family: Calibri, sans-serif;margin-left:0in"> fp_image_detect_minutiae</li><li class="MsoListParagraphCxSpLast" style="margin: 0in 0in 0in 0.5in; font-size: 11pt; font-family: Calibri, sans-serif;margin-left:0in"> fp_image_detect_minutiae_finish</li></ul> I have relied on the img-capture.c example from the version 1.90.7 examples. Here is my code: <div> <div style="color: rgb(212, 212, 212); background-color: rgb(30, 30, 30); font-family: Consolas, "Courier New", monospace; font-weight: normal; font-size: 14px; line-height: 19px;"> <div>///< This function is executed in another thread using pthreads</div> <div>static void * etr_fp_gmain_loop(void *data)</div> <div>{</div> <div> FpContext *context = NULL;</div> <div> tAUTOMATON_DATA *aut = (tAUTOMATON_DATA *) data;</div> <div> aut->ctx.clops = g_cancellable_new();</div> <div> // Create libfprint context</div> <div> if ((context = etr_fp_find_device(aut)) != NULL) // Inside this function open device with fp_device_open_sync</div> <div> {</div> <div> KernelInsertEvent(aut->fsm->aut_id,AUT_EVT_OPENED,NULL,0);</div> <div> aut->ctx.gmloop = g_main_loop_new(NULL,FALSE);</div> <div> g_main_loop_run(aut->ctx.gmloop); // Run until g_main_loop_quit is called.</div> <div> // Free resources.</div> <div> g_main_loop_unref(aut->ctx.gmloop);</div> <div> g_clear_object(&context);</div> <div> aut->ctx.device = NULL;</div> <div> aut->ctx.gmloop = NULL;</div> <div> }</div> <div> g_clear_object(&aut->ctx.clops); // Free GCancellable.</div> <div> return NULL;</div> <div>}</div> <div>//---------------------------------------------------------------------------</div> <div>static void etr_fp_extract_minutiae(FpImage *img, GAsyncResult *res, void *user_data)</div> <div>{</div> <div> GPtrArray *minutiaes = NULL;</div> <div> g_autoptr(GError) error = NULL;</div> <div> tAUTOMATON_DATA *aut = (tAUTOMATON_DATA *) user_data;</div> <div> if (fp_image_detect_minutiae_finish(img,res,&error))</div> <div> {</div> <div> minutiaes = fp_image_get_minutiae(img);</div> <div> if (minutiaes != NULL)</div> <div> {</div> <div> gint img_w = fp_image_get_width(img);</div> <div> gint img_h = fp_image_get_height(img);</div> <div> // Do something</div> <div> }</div> <div> else</div> <div> LogError("Error retrieving minutiae");</div> <div> }</div> <div> else</div> <div> {</div> <div> LogError("Error retrieving minutiae from FingerPrint. %s",(error) ? error->message : "Unknown Error");</div> <div> if (error != NULL && error->code == G_IO_ERROR_CANCELLED)</div> <div> g_cancellable_reset(aut->ctx.clops);</div> <div> }</div> <div> KernelInsertEvent(aut->fsm->aut_id,AUT_EVT_DATA,NULL,0);</div> <div>}</div> <div>//---------------------------------------------------------------------------</div> <div>static void etr_fp_dev_capture_cb(FpDevice *dev, GAsyncResult *res, void *user_data)</div> <div>{</div> <div> int result;</div> <div> FpImage *image = NULL;</div> <div> g_autoptr(GError) error = NULL;</div> <div> tAUTOMATON_DATA *aut = (tAUTOMATON_DATA *) user_data;</div> <div> image = fp_device_capture_finish(dev, res, &error);</div> <div> if (!image)</div> <div> {</div> <div> LogError("Error capturing fingerprint: %s",error->message);</div> <div> if (aut->ctx.cap_cb)</div> <div> aut->ctx.cap_cb(error->code,NULL,aut->ctx.u_data);</div> <div> if (error->code == G_IO_ERROR_CANCELLED) {</div> <div> g_cancellable_reset(aut->ctx.clops);</div> <div> KernelInsertEvent(aut->fsm->aut_id,AUT_EVT_DATA,NULL,0);</div> <div> }</div> <div> else {</div> <div> etr_fp_dev_close(aut);</div> <div> KernelInsertEvent(aut->fsm->aut_id,AUT_EVT_REMOVED,NULL,0);</div> <div> }</div> <div> return;</div> <div> }</div> <div> if (aut->ctx.imgpath != NULL)</div> <div> if ((result = etr_fp_save_image_to_pgm(image, aut->ctx.imgpath)) < 0)</div> <div> LogError("Unable to save the image in specified path: %s. Error code: %d",aut->ctx.imgpath,result);</div> <div> fp_image_detect_minutiae(image,aut->ctx.clops,(GAsyncReadyCallback)etr_fp_extract_minutiae,aut);</div> <div>}</div> <div>//---------------------------------------------------------------------------</div> <div>gboolean etr_fp_start_capture(gpointer data)</div> <div>{</div> <div> tAUTOMATON_DATA *aut = (tAUTOMATON_DATA *) data;</div> <div> fp_device_capture(aut->ctx.device, TRUE, aut->ctx.clops, (GAsyncReadyCallback) etr_fp_dev_capture_cb, aut);</div> <div> return FALSE;</div> <div>}</div> <div>//---------------------------------------------------------------------------</div> <div>void capturing__entry(void* aut_data)</div> <div>{</div> <div> g_idle_add(etr_fp_start_capture,aut_data);</div> <div>}</div> <div>//---------------------------------------------------------------------------</div> <div>static gboolean etr_fp_dev_close(gpointer data)</div> <div>{</div> <div> g_autoptr(GError) gerror = NULL;</div> <div> tAUTOMATON_DATA *aut = (tAUTOMATON_DATA *)data;</div> <div> if (!fp_device_close_sync(aut->ctx.device,NULL,&gerror))</div> <div> LogError("Error closing device: %s",gerror->message);</div> <div> g_main_loop_quit(aut->ctx.gmloop);</div> <div> return FALSE;</div> <div>}</div> <div>//---------------------------------------------------------------------------</div> <div>int etr_fp_init(void *aut)</div> <div>{</div> <div> /*</div> <div> do other things.</div> <div> */</div> <div> if (pthread_create(&aut->ctx.event_t,NULL,&etr_fp_gmain_loop,aut))</div> <div> {</div> <div> LogError("%s:Id%d Unable to create glib event loop thread",K_ETR_FP_API_PREFIX,id);</div> <div> return K_ETR_FP_FAILURE;</div> <div> }</div> <div> return K_ETR_FP_OK;</div> <div>}</div> <div>//---------------------------------------------------------------------------</div> <div>int etr_fp_end(void *aut)</div> <div>{</div> <div> // Free libfrpint, glib resources when gmainloop ends free automaton.;</div> <div> g_idle_add_full(G_PRIORITY_HIGH_IDLE,etr_fp_dev_close,aut,NULL);</div> <div> pthread_join(aut->ctx.event_t,NULL);</div> <div> return 0;</div> <div>}</div> //---------------------------------------------------------------------------</div> <div> </div> </div> I have read some things about valgrind with Glib and possible problems. So, my first question is: Can I use Valgrind with libfprint? When I run valgrind with the following configuration: valgrind --leak-check=full --show-leak-kinds=definite,indirect --track-origins=yes --verbose --log-file=valgrind-out.txt ./myapp The following catches my attention: ==35461== 1,113,600 bytes in 10 blocks are indirectly lost in loss record 910 of 912 ==35461== at 0x483B7F3: malloc (in /usr/lib/x86_64-linux-gnu/valgrind/vgpreload_memcheck-amd64-linux.so) ==35461== by 0x4AECCB8: g_malloc (in /usr/lib/x86_64-linux-gnu/libglib-2.0.so.0.6400.6) ==35461== by 0x5450990: binarize_image_V2 (in /usr/lib/x86_64-linux-gnu/libfprint.so.2.0.0) ==35461== by 0x5450B3C: binarize_V2 (in /usr/lib/x86_64-linux-gnu/libfprint.so.2.0.0) ==35461== by 0x544605A: lfs_detect_minutiae_V2 (in /usr/lib/x86_64-linux-gnu/libfprint.so.2.0.0) ==35461== by 0x543EE8D: get_minutiae (in /usr/lib/x86_64-linux-gnu/libfprint.so.2.0.0) ==35461== by 0x540E7C1: fp_image_detect_minutiae_thread_func (in /usr/lib/x86_64-linux-gnu/libfprint.so.2.0.0) ==35461== by 0x8457C81: ??? (in /usr/lib/x86_64-linux-gnu/libgio-2.0.so.0.6400.6) ==35461== by 0x4B111B3: ??? (in /usr/lib/x86_64-linux-gnu/libglib-2.0.so.0.6400.6) ==35461== by 0x4B10910: ??? (in /usr/lib/x86_64-linux-gnu/libglib-2.0.so.0.6400.6) ==35461== by 0x858C608: start_thread (pthread_create.c:477) ==35461== by 0x4CE2292: clone (clone.S:95) ==35461== ==35461== 1,113,600 bytes in 10 blocks are indirectly lost in loss record 911 of 912 ==35461== at 0x483B7F3: malloc (in /usr/lib/x86_64-linux-gnu/valgrind/vgpreload_memcheck-amd64-linux.so) ==35461== by 0x4AECCB8: g_malloc (in /usr/lib/x86_64-linux-gnu/libglib-2.0.so.0.6400.6) ==35461== by 0x540EF98: fp_image_detect_minutiae (in /usr/lib/x86_64-linux-gnu/libfprint.so.2.0.0) ==35461== by 0x484E10B: etr_fp_dev_capture_cb (fingerprint.c:528) ==35461== by 0x8456F68: ??? (in /usr/lib/x86_64-linux-gnu/libgio-2.0.so.0.6400.6) ==35461== by 0x8457B5C: ??? (in /usr/lib/x86_64-linux-gnu/libgio-2.0.so.0.6400.6) ==35461== by 0x54375AA: fp_device_task_return_in_idle_cb (in /usr/lib/x86_64-linux-gnu/libfprint.so.2.0.0) ==35461== by 0x4AE6E6D: g_main_context_dispatch (in /usr/lib/x86_64-linux-gnu/libglib-2.0.so.0.6400.6) ==35461== by 0x4AE721F: ??? (in /usr/lib/x86_64-linux-gnu/libglib-2.0.so.0.6400.6) ==35461== by 0x4AE7512: g_main_loop_run (in /usr/lib/x86_64-linux-gnu/libglib-2.0.so.0.6400.6) ==35461== by 0x484DAA7: etr_fp_gmain_loop (fingerprint.c:401) ==35461== by 0x858C608: start_thread (pthread_create.c:477) ==35461== ==35461== 2,333,120 (800 direct, 2,332,320 indirect) bytes in 10 blocks are definitely lost in loss record 912 of 912 ==35461== at 0x483B7F3: malloc (in /usr/lib/x86_64-linux-gnu/valgrind/vgpreload_memcheck-amd64-linux.so) ==35461== by 0x4AECCB8: g_malloc (in /usr/lib/x86_64-linux-gnu/libglib-2.0.so.0.6400.6) ==35461== by 0x4B052A5: g_slice_alloc (in /usr/lib/x86_64-linux-gnu/libglib-2.0.so.0.6400.6) ==35461== by 0x4B058CD: g_slice_alloc0 (in /usr/lib/x86_64-linux-gnu/libglib-2.0.so.0.6400.6) ==35461== by 0x837B0CF: g_type_create_instance (in /usr/lib/x86_64-linux-gnu/libgobject-2.0.so.0.6400.6) ==35461== by 0x835A34C: ??? (in /usr/lib/x86_64-linux-gnu/libgobject-2.0.so.0.6400.6) ==35461== by 0x835C377: g_object_new_valist (in /usr/lib/x86_64-linux-gnu/libgobject-2.0.so.0.6400.6) ==35461== by 0x835C6CC: g_object_new (in /usr/lib/x86_64-linux-gnu/libgobject-2.0.so.0.6400.6) ==35461== by 0x5428644: imaging_run_state (in /usr/lib/x86_64-linux-gnu/libfprint.so.2.0.0) ==35461== by 0x543E345: transfer_finish_cb (in /usr/lib/x86_64-linux-gnu/libfprint.so.2.0.0) ==35461== by 0x8456F68: ??? (in /usr/lib/x86_64-linux-gnu/libgio-2.0.so.0.6400.6) ==35461== by 0x8456FAC: ??? (in /usr/lib/x86_64-linux-gnu/libgio-2.0.so.0.6400.6) ==35461== ==35461== LEAK SUMMARY: ==35461== definitely lost: 1,128 bytes in 22 blocks ==35461== indirectly lost: 2,334,793 bytes in 2,604 blocks ==35461== possibly lost: 34,119 bytes in 142 blocks ==35461== still reachable: 190,826 bytes in 1,292 blocks ==35461== of which reachable via heuristic: ==35461== length64 : 1,160 bytes in 29 blocks ==35461== newarray : 1,584 bytes in 19 blocks ==35461== suppressed: 0 bytes in 0 blocks ==35461== Reachable blocks (those to which a pointer was found) are not shown. ==35461== To see them, rerun with: --leak-check=full --show-leak-kinds=all ==35461== ==35461== ERROR SUMMARY: 135 errors from 135 contexts (suppressed: 0 from 0) So fp_image_detect_minutiae have a memory leak? Or maybe is a Valgrind problem. Every time I run a capture the memory usage grows and it doesn't seem to return the resources used. I am executing the g_main_loop_run in another thread and the calls to the capture function are made using g_idle_add Another problem detected is when unplug device, next call capture process. The removed device error is detected correctly but when I reinitialize the library creating a new object of type: <ul> <li>FPContext</li><li> GmainLoop </li></ul> Have the next error: (process:35): libfprint-context-DEBUG: 17:47:21.078: No driver found for USB device 1D6B:0001 ---> Device was removed and not plugged in yet . <div>(process:35): libfprint-image_device-DEBUG: 17:47:21.225: Image device open completed --> Device plugged in again, free up memory and create new objects.</div> <div>(process:35): libfprint-device-DEBUG: 17:47:21.225: Device reported open completion</div> <div>(process:35): libfprint-device-DEBUG: 17:47:21.225: Completing action FPI_DEVICE_ACTION_OPEN in idle!</div> <div>(process:35): GLib-CRITICAL **: 17:47:21.226: g_ptr_array_find_with_equal_func: assertion 'haystack != NULL' failed </div> <div>(process:35): libfprint-context-CRITICAL **: 17:47:21.226: remove_device_idle_cb: assertion 'g_ptr_array_find (priv->devices, data->device, &idx)' failed </div> <div> </div> The library works fine when the device is reconnected, but I want to know about these 2 errors. Thanks for all the support provided. </div> </div> </body> </html>