Re: [Mesa-dev] [PATCH v2] util/u_atomic: provide 64bit atomics where they're missing
On Mon, Apr 3, 2017 at 1:37 PM, Mark Janes wrote: > > This commit appears to intermittently provoke gpu hangs on 32-bit Intel > systems when running > > piglit.shaders.glsl-max-varyings >max_varying_components > > Since the behavior is intermittent, I may have identified the wrong > patch. Please let me know if this patch seems unlikely to affect the > test. I think it's not this patch, but I haven't figured out which it is yet. Applying this patch doesn't change i965_dri.so (md5sum is the same). ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/mesa-dev
Re: [Mesa-dev] [PATCH v2] util/u_atomic: provide 64bit atomics where they're missing
This commit appears to intermittently provoke gpu hangs on 32-bit Intel systems when running piglit.shaders.glsl-max-varyings >max_varying_components Since the behavior is intermittent, I may have identified the wrong patch. Please let me know if this patch seems unlikely to affect the test. Matt Turner writes: > On Thu, Mar 30, 2017 at 3:47 PM, Matt Turner wrote: >> On Thu, Mar 30, 2017 at 3:26 PM, Grazvydas Ignotas > wrote: >>> There are still some distributions trying to support unfortunate people >>> with old or exotic CPUs that don't have 64bit atomic operations. When >>> compiling for such a machine, gcc conveniently inserts a library call to >>> a helper, but it's implementation is missing and we get a linker error. >>> This allows us to provide our own implementation, which is marked weak >>> to prefer a better implementation, should one exist. >>> >>> v2: changed copyright, some style adjustments >>> >>> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=93089 >>> Signed-off-by: Grazvydas Ignotas >>> Reviewed-by: Matt Turner >> >> Thanks, I'll commit this. >> >>> --- >>> no commit access, but request sent: >>> https://bugs.freedesktop.org/show_bug.cgi?id=100467 >> >> Thanks. I commented on the bug and said I approve :) > > I modified the patch slightly to print the results of the test with > AC_MSG_CHECKING/AC_MSG_RESULT and pushed it as commit a6a38a038bd62e. > > I'll do some build tests on various architectures. I think I like the idea > of this going to the stable 17.0 branch. > ___ > mesa-dev mailing list > mesa-dev@lists.freedesktop.org > https://lists.freedesktop.org/mailman/listinfo/mesa-dev ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/mesa-dev
Re: [Mesa-dev] [PATCH v2] util/u_atomic: provide 64bit atomics where they're missing
On Thu, Mar 30, 2017 at 3:47 PM, Matt Turner wrote: > On Thu, Mar 30, 2017 at 3:26 PM, Grazvydas Ignotas wrote: >> There are still some distributions trying to support unfortunate people >> with old or exotic CPUs that don't have 64bit atomic operations. When >> compiling for such a machine, gcc conveniently inserts a library call to >> a helper, but it's implementation is missing and we get a linker error. >> This allows us to provide our own implementation, which is marked weak >> to prefer a better implementation, should one exist. >> >> v2: changed copyright, some style adjustments >> >> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=93089 >> Signed-off-by: Grazvydas Ignotas >> Reviewed-by: Matt Turner > > Thanks, I'll commit this. > >> --- >> no commit access, but request sent: >> https://bugs.freedesktop.org/show_bug.cgi?id=100467 > > Thanks. I commented on the bug and said I approve :) I modified the patch slightly to print the results of the test with AC_MSG_CHECKING/AC_MSG_RESULT and pushed it as commit a6a38a038bd62e. I'll do some build tests on various architectures. I think I like the idea of this going to the stable 17.0 branch. ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/mesa-dev
Re: [Mesa-dev] [PATCH v2] util/u_atomic: provide 64bit atomics where they're missing
On Thu, Mar 30, 2017 at 3:26 PM, Grazvydas Ignotas wrote: > There are still some distributions trying to support unfortunate people > with old or exotic CPUs that don't have 64bit atomic operations. When > compiling for such a machine, gcc conveniently inserts a library call to > a helper, but it's implementation is missing and we get a linker error. > This allows us to provide our own implementation, which is marked weak > to prefer a better implementation, should one exist. > > v2: changed copyright, some style adjustments > > Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=93089 > Signed-off-by: Grazvydas Ignotas > Reviewed-by: Matt Turner Thanks, I'll commit this. > --- > no commit access, but request sent: > https://bugs.freedesktop.org/show_bug.cgi?id=100467 Thanks. I commented on the bug and said I approve :) ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/mesa-dev
[Mesa-dev] [PATCH v2] util/u_atomic: provide 64bit atomics where they're missing
There are still some distributions trying to support unfortunate people with old or exotic CPUs that don't have 64bit atomic operations. When compiling for such a machine, gcc conveniently inserts a library call to a helper, but it's implementation is missing and we get a linker error. This allows us to provide our own implementation, which is marked weak to prefer a better implementation, should one exist. v2: changed copyright, some style adjustments Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=93089 Signed-off-by: Grazvydas Ignotas Reviewed-by: Matt Turner --- no commit access, but request sent: https://bugs.freedesktop.org/show_bug.cgi?id=100467 configure.ac | 12 src/util/Makefile.sources | 1 + src/util/u_atomic.c | 75 +++ 3 files changed, 88 insertions(+) create mode 100644 src/util/u_atomic.c diff --git a/configure.ac b/configure.ac index 70885fb..74b870e 100644 --- a/configure.ac +++ b/configure.ac @@ -413,10 +413,22 @@ int main() { if test "x$GCC_ATOMIC_BUILTINS_SUPPORTED" = x1; then DEFINES="$DEFINES -DUSE_GCC_ATOMIC_BUILTINS" fi AM_CONDITIONAL([GCC_ATOMIC_BUILTINS_SUPPORTED], [test x$GCC_ATOMIC_BUILTINS_SUPPORTED = x1]) +dnl Check if host supports 64bit atomics +dnl note that lack of support usually results in link (not compile) error +AC_LINK_IFELSE([AC_LANG_SOURCE([[ +#include +uint64_t v; +int main() { +return __sync_add_and_fetch(&v, (uint64_t)1); +}]])], GCC_64BIT_ATOMICS_SUPPORTED=1) +if test "x$GCC_64BIT_ATOMICS_SUPPORTED" != x1; then +DEFINES="$DEFINES -DMISSING_64BIT_ATOMICS" +fi + dnl Check for Endianness AC_C_BIGENDIAN( little_endian=no, little_endian=yes, little_endian=no, diff --git a/src/util/Makefile.sources b/src/util/Makefile.sources index 8ee45d5..e905734 100644 --- a/src/util/Makefile.sources +++ b/src/util/Makefile.sources @@ -41,10 +41,11 @@ MESA_UTIL_FILES := \ string_to_uint_map.h \ strndup.h \ strtod.c \ strtod.h \ texcompress_rgtc_tmp.h \ + u_atomic.c \ u_atomic.h \ u_endian.h \ u_queue.c \ u_queue.h \ u_string.h \ diff --git a/src/util/u_atomic.c b/src/util/u_atomic.c new file mode 100644 index 000..44b75fb --- /dev/null +++ b/src/util/u_atomic.c @@ -0,0 +1,75 @@ +/* + * Copyright © 2017 Gražvydas Ignotas + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS + * IN THE SOFTWARE. + */ + +#if defined(MISSING_64BIT_ATOMICS) && defined(HAVE_PTHREAD) + +#include +#include + +#if defined(HAVE_FUNC_ATTRIBUTE_WEAK) && !defined(__CYGWIN__) +#define WEAK __attribute__((weak)) +#else +#define WEAK +#endif + +static pthread_mutex_t sync_mutex = PTHREAD_MUTEX_INITIALIZER; + +WEAK uint64_t +__sync_add_and_fetch_8(uint64_t *ptr, uint64_t val) +{ + uint64_t r; + + pthread_mutex_lock(&sync_mutex); + *ptr += val; + r = *ptr; + pthread_mutex_unlock(&sync_mutex); + + return r; +} + +WEAK uint64_t +__sync_sub_and_fetch_8(uint64_t *ptr, uint64_t val) +{ + uint64_t r; + + pthread_mutex_lock(&sync_mutex); + *ptr -= val; + r = *ptr; + pthread_mutex_unlock(&sync_mutex); + + return r; +} + +WEAK uint64_t +__atomic_fetch_add_8(uint64_t *ptr, uint64_t val, int memorder) +{ + return __sync_add_and_fetch(ptr, val); +} + +WEAK uint64_t +__atomic_fetch_sub_8(uint64_t *ptr, uint64_t val, int memorder) +{ + return __sync_sub_and_fetch(ptr, val); +} + +#endif -- 2.7.4 ___ mesa-dev mailing list mesa-dev@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/mesa-dev