|
From: Stefan K. <en...@ho...> - 2008-10-30 07:58:00
|
hi, application using fontconfig (and that basically whole of gnome) get several valgrind warnings as fontconfig uses a kind of hackish addressing scheme. Fontconfig uses mmapped caches and for this absolute pointers are useless. Thus they use the lower bit to signal that the pointer is relative (if bit==1). Now I wonder if there is a way to tell valgrind about it (client requests [1]). At least could client-requests be used to blacklist the memory-range from being scanned/monitored? Stefan [1] http://valgrind.org/docs/manual/mc-manual.html#mc-manual.clientreqs |
|
From: Nicholas N. <nj...@cs...> - 2008-10-30 19:32:42
|
On Thu, 30 Oct 2008, Stefan Kost wrote:
> application using fontconfig (and that basically whole of gnome) get
> several valgrind warnings as fontconfig uses a kind of hackish
> addressing scheme. Fontconfig uses mmapped caches and for this absolute
> pointers are useless. Thus they use the lower bit to signal that the
> pointer is relative (if bit==1).
> Now I wonder if there is a way to tell valgrind about it (client
> requests [1]). At least could client-requests be used to blacklist the
> memory-range from being scanned/monitored?
Does this Memcheck option do what you want?
--ignore-ranges=0xPP-0xQQ[,0xRR-0xSS] assume given addresses are OK
Nick
|
|
From: Stefan K. <en...@ho...> - 2008-11-03 19:51:08
|
hi, Nicholas Nethercote schrieb: > On Thu, 30 Oct 2008, Stefan Kost wrote: > >> application using fontconfig (and that basically whole of gnome) get >> several valgrind warnings as fontconfig uses a kind of hackish >> addressing scheme. Fontconfig uses mmapped caches and for this absolute >> pointers are useless. Thus they use the lower bit to signal that the >> pointer is relative (if bit==1). >> Now I wonder if there is a way to tell valgrind about it (client >> requests [1]). At least could client-requests be used to blacklist the >> memory-range from being scanned/monitored? > > Does this Memcheck option do what you want? > > --ignore-ranges=0xPP-0xQQ[,0xRR-0xSS] assume given addresses are OK > This is unfortunately not documented in my man-page. Although the tool has is. What the semantics ? --ignore-ranges=<start1>-<end1>,<start2>-<end2>,... I tried --ignore-ranges=0x461DDF5-0x46298F8 in the example below, but that did not help. Stefan ==18610== 6,312 (2,304 direct, 4,008 indirect) bytes in 6 blocks are definitely lost in loss record 18,383 of 18,396 ==18610== at 0x4024E7C: realloc (in /usr/lib/valgrind/x86-linux/vgpreload_memcheck.so) ==18610== by 0x462C841: (within /usr/lib/libfontconfig.so.1.2.0) ==18610== by 0x461C8F1: (within /usr/lib/libfontconfig.so.1.2.0) ==18610== by 0x461DDF5: FcConfigSubstituteWithPat (in /usr/lib/libfontconfig.so.1.2.0) ==18610== by 0x46298F8: FcFontRenderPrepare (in /usr/lib/libfontconfig.so.1.2.0) ==18610== by 0x409CBD0: pango_fc_font_map_load_fontset (pangofc-fontmap.c:1093) ==18610== by 0x451F439: pango_font_map_load_fontset (pango-fontmap.c:107) ==18610== by 0x409D1C4: pango_fc_font_map_load_font (pangofc-fontmap.c:1152) ==18610== by 0x451F4C2: pango_font_map_load_font (pango-fontmap.c:64) ==18610== by 0x451D852: pango_context_load_font (pango-context.c:292) ==18610== by 0x4523604: pango_layout_get_empty_extents_at_index (pango-layout.c:4333) ==18610== by 0x4523AD1: pango_layout_line_get_extents (pango-layout.c:4373) ==18610== by 0x4523BB6: get_line_extents_layout_coords (pango-layout.c:2358) ==18610== by 0x45272D0: pango_layout_get_extents_internal (pango-layout.c:2469) ==18610== by 0x4528D5D: pango_layout_get_pixel_extents (pango-layout.c:2615) ==18610== by 0x4528DC9: pango_layout_get_pixel_size (pango-layout.c:2664) ==18610== by 0x40718BA: (within /usr/local/lib/libgnomecanvas-2.so.0.2189.3) > Nick > > ------------------------------------------------------------------------- > This SF.Net email is sponsored by the Moblin Your Move Developer's challenge > Build the coolest Linux based applications with Moblin SDK & win great prizes > Grand prize is a trip for two to an Open Source event anywhere in the world > http://moblin-contest.org/redirect.php?banner_id=100&url=/ > _______________________________________________ > Valgrind-users mailing list > Val...@li... > https://lists.sourceforge.net/lists/listinfo/valgrind-users |
|
From: Stefan K. <en...@ho...> - 2008-12-02 21:22:19
|
hi again, no idea if this is solvable already? Stefan Stefan Kost schrieb: > hi, > > Nicholas Nethercote schrieb: >> On Thu, 30 Oct 2008, Stefan Kost wrote: >> >>> application using fontconfig (and that basically whole of gnome) get >>> several valgrind warnings as fontconfig uses a kind of hackish >>> addressing scheme. Fontconfig uses mmapped caches and for this absolute >>> pointers are useless. Thus they use the lower bit to signal that the >>> pointer is relative (if bit==1). >>> Now I wonder if there is a way to tell valgrind about it (client >>> requests [1]). At least could client-requests be used to blacklist the >>> memory-range from being scanned/monitored? >> Does this Memcheck option do what you want? >> >> --ignore-ranges=0xPP-0xQQ[,0xRR-0xSS] assume given addresses are OK >> > > This is unfortunately not documented in my man-page. Although the tool has is. > What the semantics ? --ignore-ranges=<start1>-<end1>,<start2>-<end2>,... > > I tried --ignore-ranges=0x461DDF5-0x46298F8 in the example below, but that did > not help. > > Stefan > > ==18610== 6,312 (2,304 direct, 4,008 indirect) bytes in 6 blocks are definitely > lost in loss record 18,383 of 18,396 > ==18610== at 0x4024E7C: realloc (in > /usr/lib/valgrind/x86-linux/vgpreload_memcheck.so) > ==18610== by 0x462C841: (within /usr/lib/libfontconfig.so.1.2.0) > ==18610== by 0x461C8F1: (within /usr/lib/libfontconfig.so.1.2.0) > ==18610== by 0x461DDF5: FcConfigSubstituteWithPat (in > /usr/lib/libfontconfig.so.1.2.0) > ==18610== by 0x46298F8: FcFontRenderPrepare (in /usr/lib/libfontconfig.so.1.2.0) > ==18610== by 0x409CBD0: pango_fc_font_map_load_fontset (pangofc-fontmap.c:1093) > ==18610== by 0x451F439: pango_font_map_load_fontset (pango-fontmap.c:107) > ==18610== by 0x409D1C4: pango_fc_font_map_load_font (pangofc-fontmap.c:1152) > ==18610== by 0x451F4C2: pango_font_map_load_font (pango-fontmap.c:64) > ==18610== by 0x451D852: pango_context_load_font (pango-context.c:292) > ==18610== by 0x4523604: pango_layout_get_empty_extents_at_index > (pango-layout.c:4333) > ==18610== by 0x4523AD1: pango_layout_line_get_extents (pango-layout.c:4373) > ==18610== by 0x4523BB6: get_line_extents_layout_coords (pango-layout.c:2358) > ==18610== by 0x45272D0: pango_layout_get_extents_internal (pango-layout.c:2469) > ==18610== by 0x4528D5D: pango_layout_get_pixel_extents (pango-layout.c:2615) > ==18610== by 0x4528DC9: pango_layout_get_pixel_size (pango-layout.c:2664) > ==18610== by 0x40718BA: (within /usr/local/lib/libgnomecanvas-2.so.0.2189.3) > > > >> Nick >> >> ------------------------------------------------------------------------- >> This SF.Net email is sponsored by the Moblin Your Move Developer's challenge >> Build the coolest Linux based applications with Moblin SDK & win great prizes >> Grand prize is a trip for two to an Open Source event anywhere in the world >> http://moblin-contest.org/redirect.php?banner_id=100&url=/ >> _______________________________________________ >> Valgrind-users mailing list >> Val...@li... >> https://lists.sourceforge.net/lists/listinfo/valgrind-users > > > ------------------------------------------------------------------------- > This SF.Net email is sponsored by the Moblin Your Move Developer's challenge > Build the coolest Linux based applications with Moblin SDK & win great prizes > Grand prize is a trip for two to an Open Source event anywhere in the world > http://moblin-contest.org/redirect.php?banner_id=100&url=/ > _______________________________________________ > Valgrind-users mailing list > Val...@li... > https://lists.sourceforge.net/lists/listinfo/valgrind-users |