mirror of
git://nv-tegra.nvidia.com/linux-nvgpu.git
synced 2025-12-22 09:12:24 +03:00
dfd9feace6
When a pbdma fault needs a channel teardown, do the recovery/teardown process before acking the pbdma interrupt status back. Acking it causes the hardware to proceed which could release fences too early before the involved channel(s) have been found to be broken. With these host copyengine interrupts, the teardown sequence is light and proceeds even with the pbdma intr flag still set; there are no engines to reset when these pbdma launch check interrupts happen. The bad tsg is just disabled and the channels in it aborted. A few unit tests are so heavily affected by this refactor that they would need to be rewritten. They're not strictly needed at the moment, so do only half of the rewrite: just delete them. Bug 200611198 Change-Id: Id126fb158b6d05e46ba124cd426389046eedc053 Signed-off-by: Konsta Hölttä <kholtta@nvidia.com> Reviewed-on: https://git-master.nvidia.com/r/c/linux-nvgpu/+/2392669 Reviewed-by: automaticguardword <automaticguardword@nvidia.com> Reviewed-by: svc-mobile-coverity <svc-mobile-coverity@nvidia.com> Reviewed-by: svc-mobile-cert <svc-mobile-cert@nvidia.com> Reviewed-by: svc-mobile-misra <svc-mobile-misra@nvidia.com> Reviewed-by: Alex Waterman <alexw@nvidia.com> Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com> GVS: Gerrit_Virtual_Submit Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>