Bug#475973: mplayer: FTBFS: cabac.h:525: error: can't find a register in class 'GENERAL_REGS' while reloading 'asm'
On Mon, Apr 14, 2008 at 04:38:47PM +0200, Lucas Nussbaum wrote: Don't blame gcc 4.3. It built that code fine last week. Blame dpkg (see my other mail) If gcc fails to allocate registers, it is a bug in gcc. Report it to them. Diego -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#475973: mplayer: FTBFS: cabac.h:525: error: can't find a register in class 'GENERAL_REGS' while reloading 'asm'
Package: mplayer Version: 1.0~rc2-10 Severity: serious User: [EMAIL PROTECTED] Usertags: qa-ftbfs-20080413 qa-ftbfs Justification: FTBFS on i386 Hi, During a rebuild of all packages in sid, your package failed to build on i386. This rebuild was done with gcc 4.3 instead of gcc 4.2, because gcc 4.3 is now the default on most architectures (even if it's not the case on i386 yet). Feel free to downgrade this bug to 'important' if your package is only built on i386, and this bug is specific to gcc 4.3 (i.e the package builds fine with gcc 4.2). Relevant part: cc -g -O2 -I./libavcodec -I./libavformat -Wdisabled-optimization -Wno-pointer-sign -Wdeclaration-after-statement -I. -I. -I./libavutil -g -O2 -D_REENTRANT -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64 -D_LARGEFILE64_SOURCE -DHAVE_CONFIG_H -I/usr/include/directfb -I/usr/include/ -I/usr/include/SDL -D_REENTRANT -I/usr/include/freetype2 -I/usr/include -I/usr/include/gtk-2.0 -I/usr/lib/gtk-2.0/include -I/usr/include/atk-1.0 -I/usr/include/cairo -I/usr/include/pango-1.0 -I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include -I/usr/include/freetype2 -I/usr/include/libpng12 -I../libswscale -I../libavcodec -DHAVE_AV_CONFIG_H -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE -D_ISOC9X_SOURCE -I.. -I.. -I../libavutil -Wdisabled-optimization -Wno-pointer-sign -Wdeclaration-after-statement -I. -I.. -I../libavutil -g -O2 -D_REENTRANT -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64 -D_LARGEFILE64_SOURCE -DHAVE_CONFIG_H -I/usr/include/directfb -I/usr/include/ -I/usr/include/SDL -D_REENTRANT -I/usr/include/freetype2 -I/usr/include -I/usr/include/gtk-2.0 -I/usr/lib/gtk-2.0/include -I/usr/include/atk-1.0 -I/usr/include/cairo -I/usr/include/pango-1.0 -I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include -I/usr/include/freetype2 -I/usr/include/libpng12-c -o h264.o h264.c h264.c: In function 'decode_cabac_residual': h264.c:5350: warning: passing argument 4 of 'decode_significance_8x8_x86' discards qualifiers from pointer target type cabac.h: In function 'get_cabac_noinline': cabac.h:525: error: can't find a register in class 'GENERAL_REGS' while reloading 'asm' cabac.h:525: error: 'asm' operand has impossible constraints make[2]: *** [h264.o] Error 1 The full build log is available from: http://people.debian.org/~lucas/logs/2008/04/13 A list of current common problems and possible solutions is available at http://wiki.debian.org/qa.debian.org/FTBFS . You're welcome to contribute! About the archive rebuild: The rebuild was done on about 50 AMD64 nodes of the Grid'5000 platform, using a clean chroot containing a sid i386 environment. Internet was not accessible from the build systems. -- | Lucas Nussbaum | [EMAIL PROTECTED] http://www.lucas-nussbaum.net/ | | jabber: [EMAIL PROTECTED] GPG: 1024D/023B3F4F |
Bug#475973: mplayer: FTBFS: cabac.h:525: error: can't find a register in class 'GENERAL_REGS' while reloading 'asm'
On Mon, Apr 14, 2008 at 12:48:56PM +0200, Lucas Nussbaum wrote: During a rebuild of all packages in sid, your package failed to build on i386. This rebuild was done with gcc 4.3 instead of gcc 4.2, because gcc 4.3 is now the default on most architectures (even if it's not the case on i386 yet). Feel free to downgrade this bug to 'important' if your package is only built on i386, and this bug is specific to gcc 4.3 (i.e the package builds fine with gcc 4.2). Relevant part: cc -g -O2 -I./libavcodec -I./libavformat -Wdisabled-optimization -Wno-pointer-sign -Wdeclaration-after-statement -I. -I. -I./libavutil -g -O2 -D_REENTRANT -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64 -D_LARGEFILE64_SOURCE -DHAVE_CONFIG_H -I/usr/include/directfb -I/usr/include/ -I/usr/include/SDL -D_REENTRANT -I/usr/include/freetype2 -I/usr/include -I/usr/include/gtk-2.0 -I/usr/lib/gtk-2.0/include -I/usr/include/atk-1.0 -I/usr/include/cairo -I/usr/include/pango-1.0 -I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include -I/usr/include/freetype2 -I/usr/include/libpng12 -I../libswscale -I../libavcodec -DHAVE_AV_CONFIG_H -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE -D_ISOC9X_SOURCE -I.. -I.. -I../libavutil -Wdisabled-optimization -Wno-pointer-sign -Wdeclaration-after-statement -I. -I.. -I../libavutil -g -O2 -D_REENTRANT -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64 -D_LARGEFILE64_SOURCE -DHAVE_CONFIG_H -I/usr/include/directfb -I/usr/include/ -I/usr/include/SDL -D_REENTRANT -I/usr/include/freetype2 -I/usr/include -I/usr/include/gtk-2.0 -I/usr/lib/gtk-2.0/include -I/usr/include/atk-1.0 -I/usr/include/cairo -I/usr/include/pango-1.0 -I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include -I/usr/include/freetype2 -I/usr/include/libpng12-c -o h264.o h264.c h264.c: In function 'decode_cabac_residual': h264.c:5350: warning: passing argument 4 of 'decode_significance_8x8_x86' discards qualifiers from pointer target type cabac.h: In function 'get_cabac_noinline': cabac.h:525: error: can't find a register in class 'GENERAL_REGS' while reloading 'asm' cabac.h:525: error: 'asm' operand has impossible constraints make[2]: *** [h264.o] Error 1 That is a bug in gcc, not MPlayer. gcc should not run out of registers. Diego
Bug#475973: mplayer: FTBFS: cabac.h:525: error: can't find a register in class 'GENERAL_REGS' while reloading 'asm'
On Mon, Apr 14, 2008 at 12:48:56PM +0200, Lucas Nussbaum wrote: Package: mplayer Version: 1.0~rc2-10 Severity: serious User: [EMAIL PROTECTED] Usertags: qa-ftbfs-20080413 qa-ftbfs Justification: FTBFS on i386 Hi, During a rebuild of all packages in sid, your package failed to build on i386. This rebuild was done with gcc 4.3 instead of gcc 4.2, because gcc 4.3 is now the default on most architectures (even if it's not the case on i386 yet). Feel free to downgrade this bug to 'important' if your package is only built on i386, and this bug is specific to gcc 4.3 (i.e the package builds fine with gcc 4.2). hi I am not sure on what you mean by the above conditional ''your package is only built on i386'' AND ''this bug is specific to gcc 4.3'' What I am sure of is that mplayer builds fine with gcc 4.2, so I see this as a bug in gcc 4.3 , not a bug in mplayer. And this is the second bug with gcc 4.3 not being able to build mplayer, the other bug is 475153 Maybe gcc 4.3 is not yet mature enough? Maybe the decision to switch to gcc 4.3 was a bit too fast? a. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#475973: mplayer: FTBFS: cabac.h:525: error: can't find a register in class 'GENERAL_REGS' while reloading 'asm'
hi Read from http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=402950#44 on. As you see, that piece of code worked in the past, if enough optimization was allowed. The fact that it does not work anymore is quite surprising and I would consider this a regression in gcc 4.3 anyhow I will do some cross checks and see a. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#475973: mplayer: FTBFS: cabac.h:525: error: can't find a register in class 'GENERAL_REGS' while reloading 'asm'
On 14/04/08 at 13:55 +0200, A Mennucc wrote: On Mon, Apr 14, 2008 at 12:48:56PM +0200, Lucas Nussbaum wrote: Package: mplayer Version: 1.0~rc2-10 Severity: serious User: [EMAIL PROTECTED] Usertags: qa-ftbfs-20080413 qa-ftbfs Justification: FTBFS on i386 Hi, During a rebuild of all packages in sid, your package failed to build on i386. This rebuild was done with gcc 4.3 instead of gcc 4.2, because gcc 4.3 is now the default on most architectures (even if it's not the case on i386 yet). Feel free to downgrade this bug to 'important' if your package is only built on i386, and this bug is specific to gcc 4.3 (i.e the package builds fine with gcc 4.2). hi I am not sure on what you mean by the above conditional ''your package is only built on i386'' AND ''this bug is specific to gcc 4.3'' What I am sure of is that mplayer builds fine with gcc 4.2, so I see this as a bug in gcc 4.3 , not a bug in mplayer. And this is the second bug with gcc 4.3 not being able to build mplayer, the other bug is 475153 Maybe gcc 4.3 is not yet mature enough? Maybe the decision to switch to gcc 4.3 was a bit too fast? Don't blame gcc 4.3. It built that code fine last week. Blame dpkg (see my other mail) -- | Lucas Nussbaum | [EMAIL PROTECTED] http://www.lucas-nussbaum.net/ | | jabber: [EMAIL PROTECTED] GPG: 1024D/023B3F4F | -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#475973: mplayer: FTBFS: cabac.h:525: error: can't find a register in class 'GENERAL_REGS' while reloading 'asm'
On 14/04/08 at 11:58 +0200, Diego Biurrun wrote: On Mon, Apr 14, 2008 at 12:48:56PM +0200, Lucas Nussbaum wrote: During a rebuild of all packages in sid, your package failed to build on i386. This rebuild was done with gcc 4.3 instead of gcc 4.2, because gcc 4.3 is now the default on most architectures (even if it's not the case on i386 yet). Feel free to downgrade this bug to 'important' if your package is only built on i386, and this bug is specific to gcc 4.3 (i.e the package builds fine with gcc 4.2). Relevant part: cc -g -O2 -I./libavcodec -I./libavformat -Wdisabled-optimization -Wno-pointer-sign -Wdeclaration-after-statement -I. -I. -I./libavutil -g -O2 -D_REENTRANT -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64 -D_LARGEFILE64_SOURCE -DHAVE_CONFIG_H -I/usr/include/directfb -I/usr/include/ -I/usr/include/SDL -D_REENTRANT -I/usr/include/freetype2 -I/usr/include -I/usr/include/gtk-2.0 -I/usr/lib/gtk-2.0/include -I/usr/include/atk-1.0 -I/usr/include/cairo -I/usr/include/pango-1.0 -I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include -I/usr/include/freetype2 -I/usr/include/libpng12 -I../libswscale -I../libavcodec -DHAVE_AV_CONFIG_H -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE -D_ISOC9X_SOURCE -I.. -I.. -I../libavutil -Wdisabled-optimization -Wno-pointer-sign -Wdeclaration-after-statement -I. -I.. -I../libavutil -g -O2 -D_REENTRANT -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64 -D_LARGEFILE64_SOURCE -DHAVE_CONFIG_H -I/usr/include/directfb -I/usr/include/ -I/usr/include/SDL -D_REENTRANT -I/usr/include/freetype2 -I/usr/include -I/usr/include/gtk-2.0 -I/usr/lib/gtk-2.0/include -I/usr/include/atk-1.0 -I/usr/include/cairo -I/usr/include/pango-1.0 -I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include -I/usr/include/freetype2 -I/usr/include/libpng12-c -o h264.o h264.c h264.c: In function 'decode_cabac_residual': h264.c:5350: warning: passing argument 4 of 'decode_significance_8x8_x86' discards qualifiers from pointer target type cabac.h: In function 'get_cabac_noinline': cabac.h:525: error: can't find a register in class 'GENERAL_REGS' while reloading 'asm' cabac.h:525: error: 'asm' operand has impossible constraints make[2]: *** [h264.o] Error 1 That is a bug in gcc, not MPlayer. gcc should not run out of registers. Hi, This might be caused by the default value of CFLAGS now set by dpkg-buildpackage. Maybe this code doesn't build with -O2 ? -- | Lucas Nussbaum | [EMAIL PROTECTED] http://www.lucas-nussbaum.net/ | | jabber: [EMAIL PROTECTED] GPG: 1024D/023B3F4F |
Bug#475973: mplayer: FTBFS: cabac.h:525: error: can't find a register in class 'GENERAL_REGS' while reloading 'asm'
On Mon, Apr 14, 2008 at 04:30:24PM +0200, Lucas Nussbaum wrote: On 14/04/08 at 11:58 +0200, Diego Biurrun wrote: On Mon, Apr 14, 2008 at 12:48:56PM +0200, Lucas Nussbaum wrote: During a rebuild of all packages in sid, your package failed to build on i386. This rebuild was done with gcc 4.3 instead of gcc 4.2, because gcc 4.3 is now the default on most architectures (even if it's not the case on i386 yet). Feel free to downgrade this bug to 'important' if your package is only built on i386, and this bug is specific to gcc 4.3 (i.e the package builds fine with gcc 4.2). Relevant part: cc -g -O2 -I./libavcodec -I./libavformat -Wdisabled-optimization -Wno-pointer-sign -Wdeclaration-after-statement -I. -I. -I./libavutil -g -O2 -D_REENTRANT -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64 -D_LARGEFILE64_SOURCE -DHAVE_CONFIG_H -I/usr/include/directfb -I/usr/include/ -I/usr/include/SDL -D_REENTRANT -I/usr/include/freetype2 -I/usr/include -I/usr/include/gtk-2.0 -I/usr/lib/gtk-2.0/include -I/usr/include/atk-1.0 -I/usr/include/cairo -I/usr/include/pango-1.0 -I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include -I/usr/include/freetype2 -I/usr/include/libpng12 -I../libswscale -I../libavcodec -DHAVE_AV_CONFIG_H -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE -D_ISOC9X_SOURCE -I.. -I.. -I../libavutil -Wdisabled-optimization -Wno-pointer-sign -Wdeclaration-after-statement -I. -I.. -I../libavutil -g -O2 -D_REENTRANT -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64 -D_LARGEFILE64_SOURCE -DHAVE_CONFIG_H -I/usr/include/directfb -I/usr/include/ -I/usr/include/SDL -D_REENTRANT -I/usr/include/freetype2 -I/usr/include -I/usr/include/gtk-2.0 -I/usr/lib/gtk-2.0/include -I/usr/include/atk-1.0 -I/usr/include/cairo -I/usr/include/pango-1.0 -I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include -I/usr/include/freetype2 -I/usr/include/libpng12-c -o h264.o h264.c h264.c: In function 'decode_cabac_residual': h264.c:5350: warning: passing argument 4 of 'decode_significance_8x8_x86' discards qualifiers from pointer target type cabac.h: In function 'get_cabac_noinline': cabac.h:525: error: can't find a register in class 'GENERAL_REGS' while reloading 'asm' cabac.h:525: error: 'asm' operand has impossible constraints make[2]: *** [h264.o] Error 1 That is a bug in gcc, not MPlayer. gcc should not run out of registers. This might be caused by the default value of CFLAGS now set by dpkg-buildpackage. Maybe this code doesn't build with -O2 ? No, certainly not. From the MPlayer developers side, only the build flags chosen by MPlayer's configure are acceptable and anything else is unsupported and very likely to fail. In particular, -fomit-frame-pointer is mandatory. While there is alternative code that works without, it is (for H.264 about 10%) slower and thus not really acceptable. Also, -O2 compared to the default of -O4 (equivalent to -O3) has very different inlining behaviour and might cause significant performance differences as well. Greetings, Reimar Döffinger
Bug#475973: mplayer: FTBFS: cabac.h:525: error: can't find a register in class 'GENERAL_REGS' while reloading 'asm'
On 14/04/08 at 15:06 +0200, Reimar Döffinger wrote: On Mon, Apr 14, 2008 at 04:30:24PM +0200, Lucas Nussbaum wrote: On 14/04/08 at 11:58 +0200, Diego Biurrun wrote: On Mon, Apr 14, 2008 at 12:48:56PM +0200, Lucas Nussbaum wrote: During a rebuild of all packages in sid, your package failed to build on i386. This rebuild was done with gcc 4.3 instead of gcc 4.2, because gcc 4.3 is now the default on most architectures (even if it's not the case on i386 yet). Feel free to downgrade this bug to 'important' if your package is only built on i386, and this bug is specific to gcc 4.3 (i.e the package builds fine with gcc 4.2). Relevant part: cc -g -O2 -I./libavcodec -I./libavformat -Wdisabled-optimization -Wno-pointer-sign -Wdeclaration-after-statement -I. -I. -I./libavutil -g -O2 -D_REENTRANT -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64 -D_LARGEFILE64_SOURCE -DHAVE_CONFIG_H -I/usr/include/directfb -I/usr/include/ -I/usr/include/SDL -D_REENTRANT -I/usr/include/freetype2 -I/usr/include -I/usr/include/gtk-2.0 -I/usr/lib/gtk-2.0/include -I/usr/include/atk-1.0 -I/usr/include/cairo -I/usr/include/pango-1.0 -I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include -I/usr/include/freetype2 -I/usr/include/libpng12 -I../libswscale -I../libavcodec -DHAVE_AV_CONFIG_H -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE -D_ISOC9X_SOURCE -I.. -I.. -I../libavutil -Wdisabled-optimization -Wno-pointer-sign -Wdeclaration-after-statement -I. -I.. -I../libavutil -g -O2 -D_REENTRANT -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64 -D_LARGEFILE64_SOURCE -DHAVE_CONFIG_H -I/usr/include/directfb -I/usr/include/ -I/usr/include/SDL -D_REENTRANT -I/usr/include/freetype2 -I/usr/include -I/usr/include/gtk-2.0 -I/usr/lib/gtk-2.0/include -I/usr/include/atk-1.0 -I/usr/include/cairo -I/usr/include/pango-1.0 -I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include -I/usr/include/freetype2 -I/usr/include/libpng12-c -o h264.o h264.c h264.c: In function 'decode_cabac_residual': h264.c:5350: warning: passing argument 4 of 'decode_significance_8x8_x86' discards qualifiers from pointer target type cabac.h: In function 'get_cabac_noinline': cabac.h:525: error: can't find a register in class 'GENERAL_REGS' while reloading 'asm' cabac.h:525: error: 'asm' operand has impossible constraints make[2]: *** [h264.o] Error 1 That is a bug in gcc, not MPlayer. gcc should not run out of registers. This might be caused by the default value of CFLAGS now set by dpkg-buildpackage. Maybe this code doesn't build with -O2 ? No, certainly not. From the MPlayer developers side, only the build flags chosen by MPlayer's configure are acceptable and anything else is unsupported and very likely to fail. In particular, -fomit-frame-pointer is mandatory. While there is alternative code that works without, it is (for H.264 about 10%) slower and thus not really acceptable. Also, -O2 compared to the default of -O4 (equivalent to -O3) has very different inlining behaviour and might cause significant performance differences as well. OK, so you just need to make sure to override dpkg-buildpackage's CFLAGS by your value. -- | Lucas Nussbaum | [EMAIL PROTECTED] http://www.lucas-nussbaum.net/ | | jabber: [EMAIL PROTECTED] GPG: 1024D/023B3F4F |
Bug#475973: mplayer: FTBFS: cabac.h:525: error: can't find a register in class 'GENERAL_REGS' while reloading 'asm'
On Mon, Apr 14, 2008 at 03:06:50PM +0200, Reimar Döffinger wrote: This might be caused by the default value of CFLAGS now set by dpkg-buildpackage. No, certainly not. well, looking at the original bug report, it really says cc -g -O2 more details are in http://people.debian.org/~lucas/logs/2008/04/13/mplayer_1.0~rc2-10_sid32-gcc43.buildlog Maybe this code doesn't build with -O2 ? This is exactly that. OK, thanks everyone, I think we found what is causing the build failure. In particular, thanks to lucas for telling me about this new feature of dpkg-buildpackage . I will study how to stop the dpkg-buildpackage CFLAGS from creeping down to the building of mplayer. a.
Bug#475973: mplayer: FTBFS: cabac.h:525: error: can't find a register in class 'GENERAL_REGS' while reloading 'asm'
On Mon, Apr 14, 2008 at 04:38:47PM +0200, Lucas Nussbaum wrote: Don't blame gcc 4.3. It built that code fine last week. Blame dpkg (see my other mail) actually, I am happy to know that. Having one gcc bug that stops mplayer from transitioning from sid to lenny was already one gcc bug too many. a. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
