Re: [Intel-gfx] [PATCH] intel-gpu-tools: Change type of variable from unsigned to uint64_t in gem_stress
Hi Daniel, We have tons of places which use gem_aperture_size. Have you reviewed them all? I've already looked at these places in latest trunk. I observed that type overflow also can happen in tests gem_evict_alignment and gem_evict_everything. But for our needs we only ported intel-gpu-tools 1.3 which doesn't contain these tests and I cannot check them on our configuration. At the moment we use 4GB for aperture size for our driver and observed that only gem_stress doesn't work correctly. We caught the bug after this test was fixed. Thanks for this test! There's a procedure to remove it (it's a exchange setting for all countries), the OTC wiki has a howto (no link atm). I found the link https://opensource.intel.com/linux-wiki/EmailDisclaimer. Thanks for the tip. I'll do it. Regards, Pavel -Original Message- From: Daniel Vetter [mailto:daniel.vet...@ffwll.ch] On Behalf Of Daniel Vetter Sent: Monday, August 04, 2014 4:10 PM To: Popov, Pavel E Cc: intel-gfx@lists.freedesktop.org Subject: Re: [Intel-gfx] [PATCH] intel-gpu-tools: Change type of variable from unsigned to uint64_t in gem_stress On Fri, Aug 01, 2014 at 12:10:44PM +0700, Pavel Popov wrote: To run gem_stress with the correct number of buffers even if aperture size = 4GB. Signed-off-by: Pavel Popov pavel.e.po...@intel.com --- tests/gem_stress.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tests/gem_stress.c b/tests/gem_stress.c index 3bbe487..313a82a 100644 --- a/tests/gem_stress.c +++ b/tests/gem_stress.c @@ -737,7 +737,7 @@ static int parse_options(int opt, int opt_index) static void init(void) { int i; - unsigned tmp; + uint64_t tmp; We have tons of places which use gem_aperture_size. Have you reviewed them all? -Daniel if (options.num_buffers == 0) { tmp = gem_aperture_size(drm_fd); -- 1.9.1 Closed Joint Stock Company Intel A/O Registered legal address: Krylatsky Hills Business Park, 17 Krylatskaya Str., Bldg 4, Moscow 121614, Russian Federation This e-mail and any attachments may contain confidential material for the sole use of the intended recipient(s). Any review or distribution by others is strictly prohibited. If you are not the intended recipient, please contact the sender and delete all copies. ___ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx -- Daniel Vetter Software Engineer, Intel Corporation +41 (0) 79 365 57 48 - http://blog.ffwll.ch Closed Joint Stock Company Intel A/O Registered legal address: Krylatsky Hills Business Park, 17 Krylatskaya Str., Bldg 4, Moscow 121614, Russian Federation This e-mail and any attachments may contain confidential material for the sole use of the intended recipient(s). Any review or distribution by others is strictly prohibited. If you are not the intended recipient, please contact the sender and delete all copies. ___ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Re: [Intel-gfx] [PATCH] intel-gpu-tools: Change type of variable from unsigned to uint64_t in gem_stress
On Tue, Aug 5, 2014 at 1:38 PM, Popov, Pavel E pavel.e.po...@intel.com wrote: We have tons of places which use gem_aperture_size. Have you reviewed them all? I've already looked at these places in latest trunk. I observed that type overflow also can happen in tests gem_evict_alignment and gem_evict_everything. But for our needs we only ported intel-gpu-tools 1.3 which doesn't contain these tests and I cannot check them on our configuration. At the moment we use 4GB for aperture size for our driver and observed that only gem_stress doesn't work correctly. We caught the bug after this test was fixed. Thanks for this test! Uhg, i-g-t 1.3 is rather old ... Can you try to roll forward your copy of i-g-t (_highly_ recommneded to stay at current code for validation, or at least always the latest release) and look at the other places? Unfortunately I don't know of any gcc tricks to make sure there's no integer precision truncation, so we have to rely on testing and review :( -Daniel -- Daniel Vetter Software Engineer, Intel Corporation +41 (0) 79 365 57 48 - http://blog.ffwll.ch ___ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Re: [Intel-gfx] [PATCH] intel-gpu-tools: Change type of variable from unsigned to uint64_t in gem_stress
On Fri, Aug 01, 2014 at 12:10:44PM +0700, Pavel Popov wrote: To run gem_stress with the correct number of buffers even if aperture size = 4GB. Signed-off-by: Pavel Popov pavel.e.po...@intel.com --- tests/gem_stress.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tests/gem_stress.c b/tests/gem_stress.c index 3bbe487..313a82a 100644 --- a/tests/gem_stress.c +++ b/tests/gem_stress.c @@ -737,7 +737,7 @@ static int parse_options(int opt, int opt_index) static void init(void) { int i; - unsigned tmp; + uint64_t tmp; We have tons of places which use gem_aperture_size. Have you reviewed them all? -Daniel if (options.num_buffers == 0) { tmp = gem_aperture_size(drm_fd); -- 1.9.1 Closed Joint Stock Company Intel A/O Registered legal address: Krylatsky Hills Business Park, 17 Krylatskaya Str., Bldg 4, Moscow 121614, Russian Federation This e-mail and any attachments may contain confidential material for the sole use of the intended recipient(s). Any review or distribution by others is strictly prohibited. If you are not the intended recipient, please contact the sender and delete all copies. ___ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx -- Daniel Vetter Software Engineer, Intel Corporation +41 (0) 79 365 57 48 - http://blog.ffwll.ch ___ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx
[Intel-gfx] [PATCH] intel-gpu-tools: Change type of variable from unsigned to uint64_t in gem_stress
To run gem_stress with the correct number of buffers even if aperture size = 4GB. Signed-off-by: Pavel Popov pavel.e.po...@intel.com --- tests/gem_stress.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tests/gem_stress.c b/tests/gem_stress.c index 3bbe487..313a82a 100644 --- a/tests/gem_stress.c +++ b/tests/gem_stress.c @@ -737,7 +737,7 @@ static int parse_options(int opt, int opt_index) static void init(void) { int i; - unsigned tmp; + uint64_t tmp; if (options.num_buffers == 0) { tmp = gem_aperture_size(drm_fd); -- 1.9.1 Closed Joint Stock Company Intel A/O Registered legal address: Krylatsky Hills Business Park, 17 Krylatskaya Str., Bldg 4, Moscow 121614, Russian Federation This e-mail and any attachments may contain confidential material for the sole use of the intended recipient(s). Any review or distribution by others is strictly prohibited. If you are not the intended recipient, please contact the sender and delete all copies. ___ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx