Re: [Intel-gfx] [PATCH v3 0/2] Don't use stolen memory or BAR mappings for ring buffers
On Fri, 2023-02-17 at 09:18 -0800, John Harrison wrote: > On 2/17/2023 00:39, Hogander, Jouni wrote: > > On Wed, 2023-02-15 at 17:10 -0800, john.c.harri...@intel.com wrote: > > > From: John Harrison > > > > > > Instruction from hardware arch is that stolen memory and BAR > > > mappings > > > are unsafe for use as ring buffers. There can be issues with > > > cache > > > aliasing due to the CPU access going to memory via the BAR. So, > > > don't > > > do it. > > Tested these patches for GPU Hang I was debugging. Seem to fix that > > one > > as well: > > > > Tested-by: Jouni Högander > Sweet! Out of interest, which platform was that? And how reproducible > was it? It would be interesting to know if an IGT was actually > regularly > showing the issue and we had just been ignoring it! It was Alderlake. It wasn't igt test. Opened several browser instances: https://webglsamples.org/aquarium/aquarium.html Took max. couple of minutes to reproduce. For some reason PSR2 DEEP_SLEEP entry/exit was some kind of trigger for this. Without PSR2 DEEP_SLEEP I couldn't reproduce the issue. BR, Jouni Högander > John. > > > > > > v2: Dont use BAR mappings either. > > > Make conditional on LLC so as not to change platforms that don't > > > need > > > to change (Daniele). > > > Add 'Fixes' tags (Tvrtko). > > > v3: Fix dumb typo. > > > > > > Signed-off-by: John Harrison > > > > > > > > > John Harrison (2): > > > drm/i915: Don't use stolen memory for ring buffers with LLC > > > drm/i915: Don't use BAR mappings for ring buffers with LLC > > > > > > drivers/gpu/drm/i915/gt/intel_ring.c | 6 +++--- > > > 1 file changed, 3 insertions(+), 3 deletions(-) > > > >
Re: [Intel-gfx] [PATCH] drm/i915: move a Kconfig symbol to unbreak the menu presentation
On 2023.02.16 22:32:33 -0800, Randy Dunlap wrote: > Hi, > > On 2/16/23 17:52, Zhenyu Wang wrote: > > On 2023.02.14 20:45:33 -0800, Randy Dunlap wrote: > >> Inserting a Kconfig symbol that does not have a dependency (DRM_I915_GVT) > >> into a list of other symbols that do have a dependency (on DRM_I915) > >> breaks the driver menu presentation in 'make *config'. > >> > > > > I'm not sure what's the actual failure in presentation, I'm not quite > > familiar > > with Kconfig, could you help to elaborate? > > > > thanks! > > For menuconfig and nconfig, it's a subtle difference. The following menu > items are indented more after the patch (i.e., they are not indented enough > before the patch): > > │Enable KVM host support Intel GVT-g graphics virtualization > │ > │ [*] Enable Intel PXP support > │ > │ drm/i915 Debugging ---> > │ > │ drm/i915 Profile Guided Optimisation ---> > > Same menu items for gconfig: they should all be subordinate (so indented) > to the main > Intel 8xx/9xx/G3x/G4x/HD Graphics > menu. > > For xconfig, it's worse. "drm/i915 Debugging" and "drm/i915 Profile Guided > Optimisation" > are shown on the left side window, while are of the other i915 options are > shown in the > right side window (before the patch). > After the patch, all subordinate options are listed in the right side window > under the > main "Intel 8xx/9xx/G3x/G4x/HD Graphics" menu item. > > See attached photos for comparisons. > I wasn't awared of the wrong indentation. Thanks a lot, Randy! Acked-by: Zhenyu Wang signature.asc Description: PGP signature