* [PATCH v4 0/3] mmu_notifier vs fs_reclaim lockdep annotations @ 2020-11-25 16:25 Daniel Vetter 2020-11-25 16:25 ` [PATCH v4 1/3] mm: Track mmu notifiers in fs_reclaim_acquire/release Daniel Vetter ` (3 more replies) 0 siblings, 4 replies; 27+ messages in thread From: Daniel Vetter @ 2020-11-25 16:25 UTC (permalink / raw) To: DRI Development Cc: Daniel Vetter, Intel Graphics Development, LKML, linux-xfs, linux-mm, linux-fsdevel Hi all, Just resending with the polish applied, no functional changes at all. Previous versions. v3: https://lore.kernel.org/dri-devel/20201120095445.1195585-1-daniel.vetter@ffwll.ch/ v2: https://lore.kernel.org/dri-devel/20200610194101.1668038-1-daniel.vetter@ffwll.ch/ Changes since v3: - more acks/r-b - typos in the kerneldoc fixed Changes since v2: - Now hopefully the bug that bombed xfs fixed. - With unit-tests (that's the part I really wanted and never got to) - might_alloc() helper thrown in for good. I think if we have an ack/review from fs-devel this should be good to land. Last version that landed in -mm (v2) broke xfs pretty badly. Unfortuantely I don't have a working email from Qian anymore, who reported the xfs issue. Maybe Dave Chinner instead? Cheers, Daniel Daniel Vetter (3): mm: Track mmu notifiers in fs_reclaim_acquire/release mm: Extract might_alloc() debug check locking/selftests: Add testcases for fs_reclaim include/linux/sched/mm.h | 16 ++++++++++++++ lib/locking-selftest.c | 47 ++++++++++++++++++++++++++++++++++++++++ mm/mmu_notifier.c | 7 ------ mm/page_alloc.c | 31 ++++++++++++++++---------- mm/slab.h | 5 +---- mm/slob.c | 6 ++--- 6 files changed, 86 insertions(+), 26 deletions(-) -- 2.29.2 _______________________________________________ dri-devel mailing list dri-devel@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/dri-devel ^ permalink raw reply [flat|nested] 27+ messages in thread
* [PATCH v4 1/3] mm: Track mmu notifiers in fs_reclaim_acquire/release 2020-11-25 16:25 [PATCH v4 0/3] mmu_notifier vs fs_reclaim lockdep annotations Daniel Vetter @ 2020-11-25 16:25 ` Daniel Vetter 2020-11-25 16:25 ` [PATCH v4 2/3] mm: Extract might_alloc() debug check Daniel Vetter ` (2 subsequent siblings) 3 siblings, 0 replies; 27+ messages in thread From: Daniel Vetter @ 2020-11-25 16:25 UTC (permalink / raw) To: DRI Development Cc: linux-rdma, Daniel Vetter, Intel Graphics Development, Dave Chinner, LKML, Matthew Wilcox (Oracle), Christian König, linux-xfs, linux-mm, Jason Gunthorpe, Qian Cai, Jason Gunthorpe, linux-fsdevel, Daniel Vetter, Andrew Morton, Thomas Hellström fs_reclaim_acquire/release nicely catch recursion issues when allocating GFP_KERNEL memory against shrinkers (which gpu drivers tend to use to keep the excessive caches in check). For mmu notifier recursions we do have lockdep annotations since 23b68395c7c7 ("mm/mmu_notifiers: add a lockdep map for invalidate_range_start/end"). But these only fire if a path actually results in some pte invalidation - for most small allocations that's very rarely the case. The other trouble is that pte invalidation can happen any time when __GFP_RECLAIM is set. Which means only really GFP_ATOMIC is a safe choice, GFP_NOIO isn't good enough to avoid potential mmu notifier recursion. I was pondering whether we should just do the general annotation, but there's always the risk for false positives. Plus I'm assuming that the core fs and io code is a lot better reviewed and tested than random mmu notifier code in drivers. Hence why I decide to only annotate for that specific case. Furthermore even if we'd create a lockdep map for direct reclaim, we'd still need to explicit pull in the mmu notifier map - there's a lot more places that do pte invalidation than just direct reclaim, these two contexts arent the same. Note that the mmu notifiers needing their own independent lockdep map is also the reason we can't hold them from fs_reclaim_acquire to fs_reclaim_release - it would nest with the acquistion in the pte invalidation code, causing a lockdep splat. And we can't remove the annotations from pte invalidation and all the other places since they're called from many other places than page reclaim. Hence we can only do the equivalent of might_lock, but on the raw lockdep map. With this we can also remove the lockdep priming added in 66204f1d2d1b ("mm/mmu_notifiers: prime lockdep") since the new annotations are strictly more powerful. v2: Review from Thomas Hellstrom: - unbotch the fs_reclaim context check, I accidentally inverted it, but it didn't blow up because I inverted it immediately - fix compiling for !CONFIG_MMU_NOTIFIER v3: Unbreak the PF_MEMALLOC_ context flags. Thanks to Qian for the report and Dave for explaining what I failed to see. Reviewed-by: Jason Gunthorpe <jgg@nvidia.com> Cc: linux-fsdevel@vger.kernel.org Cc: Dave Chinner <david@fromorbit.com> Cc: Qian Cai <cai@lca.pw> Cc: linux-xfs@vger.kernel.org Cc: Thomas Hellström (Intel) <thomas_os@shipmail.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Jason Gunthorpe <jgg@mellanox.com> Cc: linux-mm@kvack.org Cc: linux-rdma@vger.kernel.org Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Cc: Christian König <christian.koenig@amd.com> Cc: "Matthew Wilcox (Oracle)" <willy@infradead.org> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> --- mm/mmu_notifier.c | 7 ------- mm/page_alloc.c | 31 ++++++++++++++++++++----------- 2 files changed, 20 insertions(+), 18 deletions(-) diff --git a/mm/mmu_notifier.c b/mm/mmu_notifier.c index 5654dd19addc..61ee40ed804e 100644 --- a/mm/mmu_notifier.c +++ b/mm/mmu_notifier.c @@ -612,13 +612,6 @@ int __mmu_notifier_register(struct mmu_notifier *subscription, mmap_assert_write_locked(mm); BUG_ON(atomic_read(&mm->mm_users) <= 0); - if (IS_ENABLED(CONFIG_LOCKDEP)) { - fs_reclaim_acquire(GFP_KERNEL); - lock_map_acquire(&__mmu_notifier_invalidate_range_start_map); - lock_map_release(&__mmu_notifier_invalidate_range_start_map); - fs_reclaim_release(GFP_KERNEL); - } - if (!mm->notifier_subscriptions) { /* * kmalloc cannot be called under mm_take_all_locks(), but we diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 23f5066bd4a5..ff0f9a84b8de 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -57,6 +57,7 @@ #include <trace/events/oom.h> #include <linux/prefetch.h> #include <linux/mm_inline.h> +#include <linux/mmu_notifier.h> #include <linux/migrate.h> #include <linux/hugetlb.h> #include <linux/sched/rt.h> @@ -4264,10 +4265,8 @@ should_compact_retry(struct alloc_context *ac, unsigned int order, int alloc_fla static struct lockdep_map __fs_reclaim_map = STATIC_LOCKDEP_MAP_INIT("fs_reclaim", &__fs_reclaim_map); -static bool __need_fs_reclaim(gfp_t gfp_mask) +static bool __need_reclaim(gfp_t gfp_mask) { - gfp_mask = current_gfp_context(gfp_mask); - /* no reclaim without waiting on it */ if (!(gfp_mask & __GFP_DIRECT_RECLAIM)) return false; @@ -4276,10 +4275,6 @@ static bool __need_fs_reclaim(gfp_t gfp_mask) if (current->flags & PF_MEMALLOC) return false; - /* We're only interested __GFP_FS allocations for now */ - if (!(gfp_mask & __GFP_FS)) - return false; - if (gfp_mask & __GFP_NOLOCKDEP) return false; @@ -4298,15 +4293,29 @@ void __fs_reclaim_release(void) void fs_reclaim_acquire(gfp_t gfp_mask) { - if (__need_fs_reclaim(gfp_mask)) - __fs_reclaim_acquire(); + gfp_mask = current_gfp_context(gfp_mask); + + if (__need_reclaim(gfp_mask)) { + if (gfp_mask & __GFP_FS) + __fs_reclaim_acquire(); + +#ifdef CONFIG_MMU_NOTIFIER + lock_map_acquire(&__mmu_notifier_invalidate_range_start_map); + lock_map_release(&__mmu_notifier_invalidate_range_start_map); +#endif + + } } EXPORT_SYMBOL_GPL(fs_reclaim_acquire); void fs_reclaim_release(gfp_t gfp_mask) { - if (__need_fs_reclaim(gfp_mask)) - __fs_reclaim_release(); + gfp_mask = current_gfp_context(gfp_mask); + + if (__need_reclaim(gfp_mask)) { + if (gfp_mask & __GFP_FS) + __fs_reclaim_release(); + } } EXPORT_SYMBOL_GPL(fs_reclaim_release); #endif -- 2.29.2 _______________________________________________ dri-devel mailing list dri-devel@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/dri-devel ^ permalink raw reply related [flat|nested] 27+ messages in thread
* [PATCH v4 2/3] mm: Extract might_alloc() debug check 2020-11-25 16:25 [PATCH v4 0/3] mmu_notifier vs fs_reclaim lockdep annotations Daniel Vetter 2020-11-25 16:25 ` [PATCH v4 1/3] mm: Track mmu notifiers in fs_reclaim_acquire/release Daniel Vetter @ 2020-11-25 16:25 ` Daniel Vetter 2020-11-25 16:25 ` [PATCH v4 3/3] locking/selftests: Add testcases for fs_reclaim Daniel Vetter 2020-11-25 16:25 ` [PATCH] drm/ttm: don't set page->mapping Daniel Vetter 3 siblings, 0 replies; 27+ messages in thread From: Daniel Vetter @ 2020-11-25 16:25 UTC (permalink / raw) To: DRI Development Cc: Peter Zijlstra, Daniel Vetter, Sebastian Andrzej Siewior, Dave Chinner, linux-mm, Daniel Vetter, Christoph Lameter, Michel Lespinasse, Ingo Molnar, linux-xfs, Matthew Wilcox (Oracle), Jason Gunthorpe, David Rientjes, Waiman Long, Paul E . McKenney, Intel Graphics Development, Mathieu Desnoyers, Thomas Gleixner, Joonsoo Kim, Vlastimil Babka, Randy Dunlap, LKML, Pekka Enberg, linux-fsdevel, Qian Cai, Andrew Morton Extracted from slab.h, which seems to have the most complete version including the correct might_sleep() check. Roll it out to slob.c. Motivated by a discussion with Paul about possibly changing call_rcu behaviour to allocate memory, but only roughly every 500th call. There are a lot fewer places in the kernel that care about whether allocating memory is allowed or not (due to deadlocks with reclaim code) than places that care whether sleeping is allowed. But debugging these also tends to be a lot harder, so nice descriptive checks could come in handy. I might have some use eventually for annotations in drivers/gpu. Note that unlike fs_reclaim_acquire/release gfpflags_allow_blocking does not consult the PF_MEMALLOC flags. But there is no flag equivalent for GFP_NOWAIT, hence this check can't go wrong due to memalloc_no*_save/restore contexts. Willy is working on a patch series which might change this: https://lore.kernel.org/linux-mm/20200625113122.7540-7-willy@infradead.org/ I think best would be if that updates gfpflags_allow_blocking(), since there's a ton of callers all over the place for that already. v2: Fix typos in kerneldoc (Randy) Acked-by: Vlastimil Babka <vbabka@suse.cz> Acked-by: Paul E. McKenney <paulmck@kernel.org> Reviewed-by: Jason Gunthorpe <jgg@nvidia.com> Cc. Randy Dunlap <rdunlap@infradead.org> Cc: Paul E. McKenney <paulmck@kernel.org> Cc: Christoph Lameter <cl@linux.com> Cc: Pekka Enberg <penberg@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Ingo Molnar <mingo@kernel.org> Cc: Vlastimil Babka <vbabka@suse.cz> Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Cc: Sebastian Andrzej Siewior <bigeasy@linutronix.de> Cc: Michel Lespinasse <walken@google.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: Waiman Long <longman@redhat.com> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Randy Dunlap <rdunlap@infradead.org> Cc: linux-mm@kvack.org Cc: linux-fsdevel@vger.kernel.org Cc: Dave Chinner <david@fromorbit.com> Cc: Qian Cai <cai@lca.pw> Cc: linux-xfs@vger.kernel.org Cc: "Matthew Wilcox (Oracle)" <willy@infradead.org> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> --- include/linux/sched/mm.h | 16 ++++++++++++++++ mm/slab.h | 5 +---- mm/slob.c | 6 ++---- 3 files changed, 19 insertions(+), 8 deletions(-) diff --git a/include/linux/sched/mm.h b/include/linux/sched/mm.h index d5ece7a9a403..a11a61b5226f 100644 --- a/include/linux/sched/mm.h +++ b/include/linux/sched/mm.h @@ -180,6 +180,22 @@ static inline void fs_reclaim_acquire(gfp_t gfp_mask) { } static inline void fs_reclaim_release(gfp_t gfp_mask) { } #endif +/** + * might_alloc - Mark possible allocation sites + * @gfp_mask: gfp_t flags that would be used to allocate + * + * Similar to might_sleep() and other annotations, this can be used in functions + * that might allocate, but often don't. Compiles to nothing without + * CONFIG_LOCKDEP. Includes a conditional might_sleep() if @gfp allows blocking. + */ +static inline void might_alloc(gfp_t gfp_mask) +{ + fs_reclaim_acquire(gfp_mask); + fs_reclaim_release(gfp_mask); + + might_sleep_if(gfpflags_allow_blocking(gfp_mask)); +} + /** * memalloc_noio_save - Marks implicit GFP_NOIO allocation scope. * diff --git a/mm/slab.h b/mm/slab.h index 6d7c6a5056ba..37b981247e5d 100644 --- a/mm/slab.h +++ b/mm/slab.h @@ -500,10 +500,7 @@ static inline struct kmem_cache *slab_pre_alloc_hook(struct kmem_cache *s, { flags &= gfp_allowed_mask; - fs_reclaim_acquire(flags); - fs_reclaim_release(flags); - - might_sleep_if(gfpflags_allow_blocking(flags)); + might_alloc(flags); if (should_failslab(s, flags)) return NULL; diff --git a/mm/slob.c b/mm/slob.c index 7cc9805c8091..8d4bfa46247f 100644 --- a/mm/slob.c +++ b/mm/slob.c @@ -474,8 +474,7 @@ __do_kmalloc_node(size_t size, gfp_t gfp, int node, unsigned long caller) gfp &= gfp_allowed_mask; - fs_reclaim_acquire(gfp); - fs_reclaim_release(gfp); + might_alloc(gfp); if (size < PAGE_SIZE - minalign) { int align = minalign; @@ -597,8 +596,7 @@ static void *slob_alloc_node(struct kmem_cache *c, gfp_t flags, int node) flags &= gfp_allowed_mask; - fs_reclaim_acquire(flags); - fs_reclaim_release(flags); + might_alloc(flags); if (c->size < PAGE_SIZE) { b = slob_alloc(c->size, flags, c->align, node, 0); -- 2.29.2 _______________________________________________ dri-devel mailing list dri-devel@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/dri-devel ^ permalink raw reply related [flat|nested] 27+ messages in thread
* [PATCH v4 3/3] locking/selftests: Add testcases for fs_reclaim 2020-11-25 16:25 [PATCH v4 0/3] mmu_notifier vs fs_reclaim lockdep annotations Daniel Vetter 2020-11-25 16:25 ` [PATCH v4 1/3] mm: Track mmu notifiers in fs_reclaim_acquire/release Daniel Vetter 2020-11-25 16:25 ` [PATCH v4 2/3] mm: Extract might_alloc() debug check Daniel Vetter @ 2020-11-25 16:25 ` Daniel Vetter 2020-11-25 16:25 ` [PATCH] drm/ttm: don't set page->mapping Daniel Vetter 3 siblings, 0 replies; 27+ messages in thread From: Daniel Vetter @ 2020-11-25 16:25 UTC (permalink / raw) To: DRI Development Cc: Will Deacon, Peter Zijlstra, Daniel Vetter, Intel Graphics Development, Dave Chinner, LKML, Matthew Wilcox (Oracle), Christian König, linux-xfs, linux-mm, Jason Gunthorpe, Qian Cai, linux-fsdevel, Daniel Vetter, Andrew Morton, Ingo Molnar, Thomas Hellström, linux-rdma Since I butchered this I figured better to make sure we have testcases for this now. Since we only have a locking context for __GFP_FS that's the only thing we're testing right now. Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org> Cc: linux-fsdevel@vger.kernel.org Cc: Dave Chinner <david@fromorbit.com> Cc: Qian Cai <cai@lca.pw> Cc: linux-xfs@vger.kernel.org Cc: Thomas Hellström (Intel) <thomas_os@shipmail.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Jason Gunthorpe <jgg@mellanox.com> Cc: linux-mm@kvack.org Cc: linux-rdma@vger.kernel.org Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Cc: Christian König <christian.koenig@amd.com> Cc: "Matthew Wilcox (Oracle)" <willy@infradead.org> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Will Deacon <will@kernel.org> Cc: linux-kernel@vger.kernel.org --- lib/locking-selftest.c | 47 ++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 47 insertions(+) diff --git a/lib/locking-selftest.c b/lib/locking-selftest.c index a899b3f0e2e5..ad47c3358e30 100644 --- a/lib/locking-selftest.c +++ b/lib/locking-selftest.c @@ -15,6 +15,7 @@ #include <linux/mutex.h> #include <linux/ww_mutex.h> #include <linux/sched.h> +#include <linux/sched/mm.h> #include <linux/delay.h> #include <linux/lockdep.h> #include <linux/spinlock.h> @@ -2357,6 +2358,50 @@ static void queued_read_lock_tests(void) pr_cont("\n"); } +static void fs_reclaim_correct_nesting(void) +{ + fs_reclaim_acquire(GFP_KERNEL); + might_alloc(GFP_NOFS); + fs_reclaim_release(GFP_KERNEL); +} + +static void fs_reclaim_wrong_nesting(void) +{ + fs_reclaim_acquire(GFP_KERNEL); + might_alloc(GFP_KERNEL); + fs_reclaim_release(GFP_KERNEL); +} + +static void fs_reclaim_protected_nesting(void) +{ + unsigned int flags; + + fs_reclaim_acquire(GFP_KERNEL); + flags = memalloc_nofs_save(); + might_alloc(GFP_KERNEL); + memalloc_nofs_restore(flags); + fs_reclaim_release(GFP_KERNEL); +} + +static void fs_reclaim_tests(void) +{ + printk(" --------------------\n"); + printk(" | fs_reclaim tests |\n"); + printk(" --------------------\n"); + + print_testname("correct nesting"); + dotest(fs_reclaim_correct_nesting, SUCCESS, 0); + pr_cont("\n"); + + print_testname("wrong nesting"); + dotest(fs_reclaim_wrong_nesting, FAILURE, 0); + pr_cont("\n"); + + print_testname("protected nesting"); + dotest(fs_reclaim_protected_nesting, SUCCESS, 0); + pr_cont("\n"); +} + void locking_selftest(void) { /* @@ -2478,6 +2523,8 @@ void locking_selftest(void) if (IS_ENABLED(CONFIG_QUEUED_RWLOCKS)) queued_read_lock_tests(); + fs_reclaim_tests(); + if (unexpected_testcase_failures) { printk("-----------------------------------------------------------------\n"); debug_locks = 0; -- 2.29.2 _______________________________________________ dri-devel mailing list dri-devel@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/dri-devel ^ permalink raw reply related [flat|nested] 27+ messages in thread
* [PATCH] drm/ttm: don't set page->mapping 2020-11-25 16:25 [PATCH v4 0/3] mmu_notifier vs fs_reclaim lockdep annotations Daniel Vetter ` (2 preceding siblings ...) 2020-11-25 16:25 ` [PATCH v4 3/3] locking/selftests: Add testcases for fs_reclaim Daniel Vetter @ 2020-11-25 16:25 ` Daniel Vetter 2020-11-25 16:28 ` Daniel Vetter 3 siblings, 1 reply; 27+ messages in thread From: Daniel Vetter @ 2020-11-25 16:25 UTC (permalink / raw) To: DRI Development Cc: Thomas Hellstrom, Daniel Vetter, Intel Graphics Development, LKML, linux-xfs, linux-mm, Huang Rui, Brian Paul, linux-fsdevel, Daniel Vetter, Christian Koenig Random observation while trying to review Christian's patch series to stop looking at struct page for dma-buf imports. This was originally added in commit 58aa6622d32af7d2c08d45085f44c54554a16ed7 Author: Thomas Hellstrom <thellstrom@vmware.com> Date: Fri Jan 3 11:47:23 2014 +0100 drm/ttm: Correctly set page mapping and -index members Needed for some vm operations; most notably unmap_mapping_range() with even_cows = 0. Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com> Reviewed-by: Brian Paul <brianp@vmware.com> but we do not have a single caller of unmap_mapping_range with even_cows == 0. And all the gem drivers don't do this, so another small thing we could standardize between drm and ttm drivers. Plus I don't really see a need for unamp_mapping_range where we don't want to indiscriminately shoot down all ptes. Cc: Thomas Hellstrom <thellstrom@vmware.com> Cc: Brian Paul <brianp@vmware.com> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> Cc: Christian Koenig <christian.koenig@amd.com> Cc: Huang Rui <ray.huang@amd.com> --- drivers/gpu/drm/ttm/ttm_tt.c | 12 ------------ 1 file changed, 12 deletions(-) diff --git a/drivers/gpu/drm/ttm/ttm_tt.c b/drivers/gpu/drm/ttm/ttm_tt.c index da9eeffe0c6d..5b2eb6d58bb7 100644 --- a/drivers/gpu/drm/ttm/ttm_tt.c +++ b/drivers/gpu/drm/ttm/ttm_tt.c @@ -284,17 +284,6 @@ int ttm_tt_swapout(struct ttm_bo_device *bdev, struct ttm_tt *ttm) return ret; } -static void ttm_tt_add_mapping(struct ttm_bo_device *bdev, struct ttm_tt *ttm) -{ - pgoff_t i; - - if (ttm->page_flags & TTM_PAGE_FLAG_SG) - return; - - for (i = 0; i < ttm->num_pages; ++i) - ttm->pages[i]->mapping = bdev->dev_mapping; -} - int ttm_tt_populate(struct ttm_bo_device *bdev, struct ttm_tt *ttm, struct ttm_operation_ctx *ctx) { @@ -313,7 +302,6 @@ int ttm_tt_populate(struct ttm_bo_device *bdev, if (ret) return ret; - ttm_tt_add_mapping(bdev, ttm); ttm->page_flags |= TTM_PAGE_FLAG_PRIV_POPULATED; if (unlikely(ttm->page_flags & TTM_PAGE_FLAG_SWAPPED)) { ret = ttm_tt_swapin(ttm); -- 2.29.2 _______________________________________________ dri-devel mailing list dri-devel@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/dri-devel ^ permalink raw reply related [flat|nested] 27+ messages in thread
* Re: [PATCH] drm/ttm: don't set page->mapping 2020-11-25 16:25 ` [PATCH] drm/ttm: don't set page->mapping Daniel Vetter @ 2020-11-25 16:28 ` Daniel Vetter 2020-11-25 18:06 ` Jason Gunthorpe 0 siblings, 1 reply; 27+ messages in thread From: Daniel Vetter @ 2020-11-25 16:28 UTC (permalink / raw) To: DRI Development Cc: Thomas Hellstrom, Intel Graphics Development, LKML, linux-xfs, Linux MM, Huang Rui, Brian Paul, linux-fsdevel, Daniel Vetter, Christian Koenig On Wed, Nov 25, 2020 at 5:25 PM Daniel Vetter <daniel.vetter@ffwll.ch> wrote: > > Random observation while trying to review Christian's patch series to > stop looking at struct page for dma-buf imports. > > This was originally added in > > commit 58aa6622d32af7d2c08d45085f44c54554a16ed7 > Author: Thomas Hellstrom <thellstrom@vmware.com> > Date: Fri Jan 3 11:47:23 2014 +0100 > > drm/ttm: Correctly set page mapping and -index members > > Needed for some vm operations; most notably unmap_mapping_range() with > even_cows = 0. > > Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com> > Reviewed-by: Brian Paul <brianp@vmware.com> > > but we do not have a single caller of unmap_mapping_range with > even_cows == 0. And all the gem drivers don't do this, so another > small thing we could standardize between drm and ttm drivers. > > Plus I don't really see a need for unamp_mapping_range where we don't > want to indiscriminately shoot down all ptes. > > Cc: Thomas Hellstrom <thellstrom@vmware.com> > Cc: Brian Paul <brianp@vmware.com> > Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> > Cc: Christian Koenig <christian.koenig@amd.com> > Cc: Huang Rui <ray.huang@amd.com> Apologies again, this shouldn't have been included. But at least I have an idea now why this patch somehow was included in the git send-email. Lovely interface :-/ -Daniel > --- > drivers/gpu/drm/ttm/ttm_tt.c | 12 ------------ > 1 file changed, 12 deletions(-) > > diff --git a/drivers/gpu/drm/ttm/ttm_tt.c b/drivers/gpu/drm/ttm/ttm_tt.c > index da9eeffe0c6d..5b2eb6d58bb7 100644 > --- a/drivers/gpu/drm/ttm/ttm_tt.c > +++ b/drivers/gpu/drm/ttm/ttm_tt.c > @@ -284,17 +284,6 @@ int ttm_tt_swapout(struct ttm_bo_device *bdev, struct ttm_tt *ttm) > return ret; > } > > -static void ttm_tt_add_mapping(struct ttm_bo_device *bdev, struct ttm_tt *ttm) > -{ > - pgoff_t i; > - > - if (ttm->page_flags & TTM_PAGE_FLAG_SG) > - return; > - > - for (i = 0; i < ttm->num_pages; ++i) > - ttm->pages[i]->mapping = bdev->dev_mapping; > -} > - > int ttm_tt_populate(struct ttm_bo_device *bdev, > struct ttm_tt *ttm, struct ttm_operation_ctx *ctx) > { > @@ -313,7 +302,6 @@ int ttm_tt_populate(struct ttm_bo_device *bdev, > if (ret) > return ret; > > - ttm_tt_add_mapping(bdev, ttm); > ttm->page_flags |= TTM_PAGE_FLAG_PRIV_POPULATED; > if (unlikely(ttm->page_flags & TTM_PAGE_FLAG_SWAPPED)) { > ret = ttm_tt_swapin(ttm); > -- > 2.29.2 > -- Daniel Vetter Software Engineer, Intel Corporation http://blog.ffwll.ch _______________________________________________ dri-devel mailing list dri-devel@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/dri-devel ^ permalink raw reply [flat|nested] 27+ messages in thread
* Re: [PATCH] drm/ttm: don't set page->mapping 2020-11-25 16:28 ` Daniel Vetter @ 2020-11-25 18:06 ` Jason Gunthorpe 2020-11-25 18:16 ` Daniel Stone [not found] ` <20201125181129.GA1858@infradead.org> 0 siblings, 2 replies; 27+ messages in thread From: Jason Gunthorpe @ 2020-11-25 18:06 UTC (permalink / raw) To: Daniel Vetter Cc: Thomas Hellstrom, Intel Graphics Development, LKML, DRI Development, linux-xfs, Linux MM, Huang Rui, Brian Paul, linux-fsdevel, Daniel Vetter, Christian Koenig On Wed, Nov 25, 2020 at 05:28:32PM +0100, Daniel Vetter wrote: > On Wed, Nov 25, 2020 at 5:25 PM Daniel Vetter <daniel.vetter@ffwll.ch> wrote: > > > > Random observation while trying to review Christian's patch series to > > stop looking at struct page for dma-buf imports. > > > > This was originally added in > > > > commit 58aa6622d32af7d2c08d45085f44c54554a16ed7 > > Author: Thomas Hellstrom <thellstrom@vmware.com> > > Date: Fri Jan 3 11:47:23 2014 +0100 > > > > drm/ttm: Correctly set page mapping and -index members > > > > Needed for some vm operations; most notably unmap_mapping_range() with > > even_cows = 0. > > > > Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com> > > Reviewed-by: Brian Paul <brianp@vmware.com> > > > > but we do not have a single caller of unmap_mapping_range with > > even_cows == 0. And all the gem drivers don't do this, so another > > small thing we could standardize between drm and ttm drivers. > > > > Plus I don't really see a need for unamp_mapping_range where we don't > > want to indiscriminately shoot down all ptes. > > > > Cc: Thomas Hellstrom <thellstrom@vmware.com> > > Cc: Brian Paul <brianp@vmware.com> > > Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> > > Cc: Christian Koenig <christian.koenig@amd.com> > > Cc: Huang Rui <ray.huang@amd.com> > > Apologies again, this shouldn't have been included. But at least I > have an idea now why this patch somehow was included in the git > send-email. Lovely interface :-/ I wrote a bit of a script around this because git send-email just too hard to use The key workflow change I made was to have it prepare all the emails to send and open them in an editor for review - exactly as they would be sent to the lists. It uses a empty 'cover-letter' commit and automatically transforms it into exactly the right stuff. Keeps track of everything you send in git, and there is a little tool to auto-run git range-diff to help build change logs.. https://github.com/jgunthorpe/Kernel-Maintainer-Tools/blob/master/gj_tools/cmd_send_patches.py I've been occasionaly wondering if I should suggest Konstantin add a sending side to b4, maybe using some of those ideas.. (careful if you run it, it does autosend without prompting) Jason _______________________________________________ dri-devel mailing list dri-devel@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/dri-devel ^ permalink raw reply [flat|nested] 27+ messages in thread
* Re: [PATCH] drm/ttm: don't set page->mapping 2020-11-25 18:06 ` Jason Gunthorpe @ 2020-11-25 18:16 ` Daniel Stone [not found] ` <20201125181129.GA1858@infradead.org> 1 sibling, 0 replies; 27+ messages in thread From: Daniel Stone @ 2020-11-25 18:16 UTC (permalink / raw) To: Jason Gunthorpe Cc: Thomas Hellstrom, Daniel Vetter, Intel Graphics Development, LKML, DRI Development, linux-xfs, Linux MM, Huang Rui, Brian Paul, linux-fsdevel, Daniel Vetter, Christian Koenig Hi, On Wed, 25 Nov 2020 at 18:06, Jason Gunthorpe <jgg@ziepe.ca> wrote: > On Wed, Nov 25, 2020 at 05:28:32PM +0100, Daniel Vetter wrote: > > Apologies again, this shouldn't have been included. But at least I > > have an idea now why this patch somehow was included in the git > > send-email. Lovely interface :-/ > > I wrote a bit of a script around this because git send-email just too > hard to use > > The key workflow change I made was to have it prepare all the emails > to send and open them in an editor for review - exactly as they would > be sent to the lists. > > It uses a empty 'cover-letter' commit and automatically transforms it > into exactly the right stuff. Keeps track of everything you send in > git, and there is a little tool to auto-run git range-diff to help > build change logs.. This sounds a fair bit like patman, which does something similar and also lets you annotate commit messages with changelogs. But of course, suggesting different methods of carving patches into stone tablets to someone who's written their own, is even more of a windmill tilt than rDMA. ;) Cheers, Daniel _______________________________________________ dri-devel mailing list dri-devel@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/dri-devel ^ permalink raw reply [flat|nested] 27+ messages in thread
[parent not found: <20201125181129.GA1858@infradead.org>]
* Re: [PATCH] drm/ttm: don't set page->mapping [not found] ` <20201125181129.GA1858@infradead.org> @ 2020-11-25 23:57 ` Jason Gunthorpe 0 siblings, 0 replies; 27+ messages in thread From: Jason Gunthorpe @ 2020-11-25 23:57 UTC (permalink / raw) To: Christoph Hellwig Cc: Thomas Hellstrom, Daniel Vetter, Intel Graphics Development, LKML, DRI Development, linux-xfs, Linux MM, Huang Rui, Brian Paul, linux-fsdevel, Daniel Vetter, Christian Koenig On Wed, Nov 25, 2020 at 06:11:29PM +0000, Christoph Hellwig wrote: > On Wed, Nov 25, 2020 at 02:06:06PM -0400, Jason Gunthorpe wrote: > > It uses a empty 'cover-letter' commit and automatically transforms it > > into exactly the right stuff. Keeps track of everything you send in > > git, and there is a little tool to auto-run git range-diff to help > > build change logs.. > > > > https://github.com/jgunthorpe/Kernel-Maintainer-Tools/blob/master/gj_tools/cmd_send_patches.py > > > > I've been occasionaly wondering if I should suggest Konstantin add a > > sending side to b4, maybe using some of those ideas.. > > > > (careful if you run it, it does autosend without prompting) > > The looks pretty fancy. Here is my trivial patchbomb.sh script > > #!/bin/sh > > COVERLETTER=$1 > PATCHES=$2 > > git send-email --annotate --to-cover --cc-cover $1 $2 > > still needs the git basecommit..endcommit notation, but it fires > up the series for review. annotate is OK, I used that for a long time.. My main gripe was it didn't setup the to/cc until after the annotate editor closes. Jason _______________________________________________ dri-devel mailing list dri-devel@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/dri-devel ^ permalink raw reply [flat|nested] 27+ messages in thread
* [PATCH 0/3] mmu_notifier fs fs_reclaim lockdep annotations @ 2020-11-20 9:54 Daniel Vetter 2020-11-20 9:54 ` [PATCH] drm/ttm: don't set page->mapping Daniel Vetter 0 siblings, 1 reply; 27+ messages in thread From: Daniel Vetter @ 2020-11-20 9:54 UTC (permalink / raw) To: DRI Development Cc: Daniel Vetter, Intel Graphics Development, LKML, linux-xfs, linux-mm, linux-fsdevel Hi all, I've finally gotten around to polish of my lockdep anntotation patches from a while ago: https://lore.kernel.org/dri-devel/20200610194101.1668038-1-daniel.vetter@ffwll.ch/ That patch has been in -mm for a few days already, but it immediately hit some issues with xfs. Changes since v2: - Now hopefully the bug that bombed xfs fixed. - With unit-tests (that's the part I really wanted and never got to) - might_alloc() helper thrown in for good. The unit test stuff was the major drag until I figured out how to make this very easy with the locking selftests. Comments, review, testing all very much welcome. Cheers, Daniel Daniel Vetter (3): mm: Track mmu notifiers in fs_reclaim_acquire/release mm: Extract might_alloc() debug check locking/selftests: Add testcases for fs_reclaim include/linux/sched/mm.h | 16 ++++++++++++++ lib/locking-selftest.c | 47 ++++++++++++++++++++++++++++++++++++++++ mm/mmu_notifier.c | 7 ------ mm/page_alloc.c | 31 ++++++++++++++++---------- mm/slab.h | 5 +---- mm/slob.c | 6 ++--- 6 files changed, 86 insertions(+), 26 deletions(-) -- 2.29.2 _______________________________________________ dri-devel mailing list dri-devel@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/dri-devel ^ permalink raw reply [flat|nested] 27+ messages in thread
* [PATCH] drm/ttm: don't set page->mapping 2020-11-20 9:54 [PATCH 0/3] mmu_notifier fs fs_reclaim lockdep annotations Daniel Vetter @ 2020-11-20 9:54 ` Daniel Vetter 2020-11-20 10:04 ` Christian König 0 siblings, 1 reply; 27+ messages in thread From: Daniel Vetter @ 2020-11-20 9:54 UTC (permalink / raw) To: DRI Development Cc: Thomas Hellstrom, Daniel Vetter, Intel Graphics Development, LKML, linux-xfs, linux-mm, Huang Rui, Brian Paul, linux-fsdevel, Daniel Vetter, Christian Koenig Random observation while trying to review Christian's patch series to stop looking at struct page for dma-buf imports. This was originally added in commit 58aa6622d32af7d2c08d45085f44c54554a16ed7 Author: Thomas Hellstrom <thellstrom@vmware.com> Date: Fri Jan 3 11:47:23 2014 +0100 drm/ttm: Correctly set page mapping and -index members Needed for some vm operations; most notably unmap_mapping_range() with even_cows = 0. Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com> Reviewed-by: Brian Paul <brianp@vmware.com> but we do not have a single caller of unmap_mapping_range with even_cows == 0. And all the gem drivers don't do this, so another small thing we could standardize between drm and ttm drivers. Plus I don't really see a need for unamp_mapping_range where we don't want to indiscriminately shoot down all ptes. Cc: Thomas Hellstrom <thellstrom@vmware.com> Cc: Brian Paul <brianp@vmware.com> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> Cc: Christian Koenig <christian.koenig@amd.com> Cc: Huang Rui <ray.huang@amd.com> --- drivers/gpu/drm/ttm/ttm_tt.c | 12 ------------ 1 file changed, 12 deletions(-) diff --git a/drivers/gpu/drm/ttm/ttm_tt.c b/drivers/gpu/drm/ttm/ttm_tt.c index da9eeffe0c6d..5b2eb6d58bb7 100644 --- a/drivers/gpu/drm/ttm/ttm_tt.c +++ b/drivers/gpu/drm/ttm/ttm_tt.c @@ -284,17 +284,6 @@ int ttm_tt_swapout(struct ttm_bo_device *bdev, struct ttm_tt *ttm) return ret; } -static void ttm_tt_add_mapping(struct ttm_bo_device *bdev, struct ttm_tt *ttm) -{ - pgoff_t i; - - if (ttm->page_flags & TTM_PAGE_FLAG_SG) - return; - - for (i = 0; i < ttm->num_pages; ++i) - ttm->pages[i]->mapping = bdev->dev_mapping; -} - int ttm_tt_populate(struct ttm_bo_device *bdev, struct ttm_tt *ttm, struct ttm_operation_ctx *ctx) { @@ -313,7 +302,6 @@ int ttm_tt_populate(struct ttm_bo_device *bdev, if (ret) return ret; - ttm_tt_add_mapping(bdev, ttm); ttm->page_flags |= TTM_PAGE_FLAG_PRIV_POPULATED; if (unlikely(ttm->page_flags & TTM_PAGE_FLAG_SWAPPED)) { ret = ttm_tt_swapin(ttm); -- 2.29.2 _______________________________________________ dri-devel mailing list dri-devel@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/dri-devel ^ permalink raw reply related [flat|nested] 27+ messages in thread
* Re: [PATCH] drm/ttm: don't set page->mapping 2020-11-20 9:54 ` [PATCH] drm/ttm: don't set page->mapping Daniel Vetter @ 2020-11-20 10:04 ` Christian König 2020-11-20 10:05 ` Daniel Vetter 0 siblings, 1 reply; 27+ messages in thread From: Christian König @ 2020-11-20 10:04 UTC (permalink / raw) To: Daniel Vetter, DRI Development Cc: Thomas Hellstrom, Intel Graphics Development, LKML, linux-xfs, linux-mm, Huang Rui, Brian Paul, linux-fsdevel, Daniel Vetter Am 20.11.20 um 10:54 schrieb Daniel Vetter: > Random observation while trying to review Christian's patch series to > stop looking at struct page for dma-buf imports. > > This was originally added in > > commit 58aa6622d32af7d2c08d45085f44c54554a16ed7 > Author: Thomas Hellstrom <thellstrom@vmware.com> > Date: Fri Jan 3 11:47:23 2014 +0100 > > drm/ttm: Correctly set page mapping and -index members > > Needed for some vm operations; most notably unmap_mapping_range() with > even_cows = 0. > > Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com> > Reviewed-by: Brian Paul <brianp@vmware.com> > > but we do not have a single caller of unmap_mapping_range with > even_cows == 0. And all the gem drivers don't do this, so another > small thing we could standardize between drm and ttm drivers. > > Plus I don't really see a need for unamp_mapping_range where we don't > want to indiscriminately shoot down all ptes. > > Cc: Thomas Hellstrom <thellstrom@vmware.com> > Cc: Brian Paul <brianp@vmware.com> > Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> > Cc: Christian Koenig <christian.koenig@amd.com> > Cc: Huang Rui <ray.huang@amd.com> This is still a NAK as long as we can't come up with a better way to track TTMs page allocations. Additional to that page_mapping() is used quite extensively in the mm code and I'm not sure if that isn't needed for other stuff as well. Regards, Christian. > --- > drivers/gpu/drm/ttm/ttm_tt.c | 12 ------------ > 1 file changed, 12 deletions(-) > > diff --git a/drivers/gpu/drm/ttm/ttm_tt.c b/drivers/gpu/drm/ttm/ttm_tt.c > index da9eeffe0c6d..5b2eb6d58bb7 100644 > --- a/drivers/gpu/drm/ttm/ttm_tt.c > +++ b/drivers/gpu/drm/ttm/ttm_tt.c > @@ -284,17 +284,6 @@ int ttm_tt_swapout(struct ttm_bo_device *bdev, struct ttm_tt *ttm) > return ret; > } > > -static void ttm_tt_add_mapping(struct ttm_bo_device *bdev, struct ttm_tt *ttm) > -{ > - pgoff_t i; > - > - if (ttm->page_flags & TTM_PAGE_FLAG_SG) > - return; > - > - for (i = 0; i < ttm->num_pages; ++i) > - ttm->pages[i]->mapping = bdev->dev_mapping; > -} > - > int ttm_tt_populate(struct ttm_bo_device *bdev, > struct ttm_tt *ttm, struct ttm_operation_ctx *ctx) > { > @@ -313,7 +302,6 @@ int ttm_tt_populate(struct ttm_bo_device *bdev, > if (ret) > return ret; > > - ttm_tt_add_mapping(bdev, ttm); > ttm->page_flags |= TTM_PAGE_FLAG_PRIV_POPULATED; > if (unlikely(ttm->page_flags & TTM_PAGE_FLAG_SWAPPED)) { > ret = ttm_tt_swapin(ttm); _______________________________________________ dri-devel mailing list dri-devel@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/dri-devel ^ permalink raw reply [flat|nested] 27+ messages in thread
* Re: [PATCH] drm/ttm: don't set page->mapping 2020-11-20 10:04 ` Christian König @ 2020-11-20 10:05 ` Daniel Vetter 2020-11-20 10:08 ` Christian König 0 siblings, 1 reply; 27+ messages in thread From: Daniel Vetter @ 2020-11-20 10:05 UTC (permalink / raw) To: Christian König Cc: Thomas Hellstrom, Intel Graphics Development, LKML, DRI Development, linux-xfs, Linux MM, Huang Rui, Brian Paul, linux-fsdevel, Daniel Vetter On Fri, Nov 20, 2020 at 11:04 AM Christian König <christian.koenig@amd.com> wrote: > > Am 20.11.20 um 10:54 schrieb Daniel Vetter: > > Random observation while trying to review Christian's patch series to > > stop looking at struct page for dma-buf imports. > > > > This was originally added in > > > > commit 58aa6622d32af7d2c08d45085f44c54554a16ed7 > > Author: Thomas Hellstrom <thellstrom@vmware.com> > > Date: Fri Jan 3 11:47:23 2014 +0100 > > > > drm/ttm: Correctly set page mapping and -index members > > > > Needed for some vm operations; most notably unmap_mapping_range() with > > even_cows = 0. > > > > Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com> > > Reviewed-by: Brian Paul <brianp@vmware.com> > > > > but we do not have a single caller of unmap_mapping_range with > > even_cows == 0. And all the gem drivers don't do this, so another > > small thing we could standardize between drm and ttm drivers. > > > > Plus I don't really see a need for unamp_mapping_range where we don't > > want to indiscriminately shoot down all ptes. > > > > Cc: Thomas Hellstrom <thellstrom@vmware.com> > > Cc: Brian Paul <brianp@vmware.com> > > Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> > > Cc: Christian Koenig <christian.koenig@amd.com> > > Cc: Huang Rui <ray.huang@amd.com> > > This is still a NAK as long as we can't come up with a better way to > track TTMs page allocations. > > Additional to that page_mapping() is used quite extensively in the mm > code and I'm not sure if that isn't needed for other stuff as well. Apologies, I'm honestly not quite sure how this lone patch here ended up in this submission. I didn't want to send it out. -Daniel > > Regards, > Christian. > > > --- > > drivers/gpu/drm/ttm/ttm_tt.c | 12 ------------ > > 1 file changed, 12 deletions(-) > > > > diff --git a/drivers/gpu/drm/ttm/ttm_tt.c b/drivers/gpu/drm/ttm/ttm_tt.c > > index da9eeffe0c6d..5b2eb6d58bb7 100644 > > --- a/drivers/gpu/drm/ttm/ttm_tt.c > > +++ b/drivers/gpu/drm/ttm/ttm_tt.c > > @@ -284,17 +284,6 @@ int ttm_tt_swapout(struct ttm_bo_device *bdev, struct ttm_tt *ttm) > > return ret; > > } > > > > -static void ttm_tt_add_mapping(struct ttm_bo_device *bdev, struct ttm_tt *ttm) > > -{ > > - pgoff_t i; > > - > > - if (ttm->page_flags & TTM_PAGE_FLAG_SG) > > - return; > > - > > - for (i = 0; i < ttm->num_pages; ++i) > > - ttm->pages[i]->mapping = bdev->dev_mapping; > > -} > > - > > int ttm_tt_populate(struct ttm_bo_device *bdev, > > struct ttm_tt *ttm, struct ttm_operation_ctx *ctx) > > { > > @@ -313,7 +302,6 @@ int ttm_tt_populate(struct ttm_bo_device *bdev, > > if (ret) > > return ret; > > > > - ttm_tt_add_mapping(bdev, ttm); > > ttm->page_flags |= TTM_PAGE_FLAG_PRIV_POPULATED; > > if (unlikely(ttm->page_flags & TTM_PAGE_FLAG_SWAPPED)) { > > ret = ttm_tt_swapin(ttm); > -- Daniel Vetter Software Engineer, Intel Corporation http://blog.ffwll.ch _______________________________________________ dri-devel mailing list dri-devel@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/dri-devel ^ permalink raw reply [flat|nested] 27+ messages in thread
* Re: [PATCH] drm/ttm: don't set page->mapping 2020-11-20 10:05 ` Daniel Vetter @ 2020-11-20 10:08 ` Christian König 2020-11-20 15:01 ` Daniel Vetter 0 siblings, 1 reply; 27+ messages in thread From: Christian König @ 2020-11-20 10:08 UTC (permalink / raw) To: Daniel Vetter Cc: Thomas Hellstrom, Intel Graphics Development, LKML, DRI Development, linux-xfs, Linux MM, Huang Rui, Brian Paul, linux-fsdevel, Daniel Vetter Am 20.11.20 um 11:05 schrieb Daniel Vetter: > On Fri, Nov 20, 2020 at 11:04 AM Christian König > <christian.koenig@amd.com> wrote: >> Am 20.11.20 um 10:54 schrieb Daniel Vetter: >>> Random observation while trying to review Christian's patch series to >>> stop looking at struct page for dma-buf imports. >>> >>> This was originally added in >>> >>> commit 58aa6622d32af7d2c08d45085f44c54554a16ed7 >>> Author: Thomas Hellstrom <thellstrom@vmware.com> >>> Date: Fri Jan 3 11:47:23 2014 +0100 >>> >>> drm/ttm: Correctly set page mapping and -index members >>> >>> Needed for some vm operations; most notably unmap_mapping_range() with >>> even_cows = 0. >>> >>> Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com> >>> Reviewed-by: Brian Paul <brianp@vmware.com> >>> >>> but we do not have a single caller of unmap_mapping_range with >>> even_cows == 0. And all the gem drivers don't do this, so another >>> small thing we could standardize between drm and ttm drivers. >>> >>> Plus I don't really see a need for unamp_mapping_range where we don't >>> want to indiscriminately shoot down all ptes. >>> >>> Cc: Thomas Hellstrom <thellstrom@vmware.com> >>> Cc: Brian Paul <brianp@vmware.com> >>> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> >>> Cc: Christian Koenig <christian.koenig@amd.com> >>> Cc: Huang Rui <ray.huang@amd.com> >> This is still a NAK as long as we can't come up with a better way to >> track TTMs page allocations. >> >> Additional to that page_mapping() is used quite extensively in the mm >> code and I'm not sure if that isn't needed for other stuff as well. > Apologies, I'm honestly not quite sure how this lone patch here ended > up in this submission. I didn't want to send it out. No problem. But looking a bit deeper into the mm code that other drm drivers don't set this correctly and still use unmap_mapping_range() sounds like quite a bug to me. Going to track down what exactly that is used for. Christian. > -Daniel > >> Regards, >> Christian. >> >>> --- >>> drivers/gpu/drm/ttm/ttm_tt.c | 12 ------------ >>> 1 file changed, 12 deletions(-) >>> >>> diff --git a/drivers/gpu/drm/ttm/ttm_tt.c b/drivers/gpu/drm/ttm/ttm_tt.c >>> index da9eeffe0c6d..5b2eb6d58bb7 100644 >>> --- a/drivers/gpu/drm/ttm/ttm_tt.c >>> +++ b/drivers/gpu/drm/ttm/ttm_tt.c >>> @@ -284,17 +284,6 @@ int ttm_tt_swapout(struct ttm_bo_device *bdev, struct ttm_tt *ttm) >>> return ret; >>> } >>> >>> -static void ttm_tt_add_mapping(struct ttm_bo_device *bdev, struct ttm_tt *ttm) >>> -{ >>> - pgoff_t i; >>> - >>> - if (ttm->page_flags & TTM_PAGE_FLAG_SG) >>> - return; >>> - >>> - for (i = 0; i < ttm->num_pages; ++i) >>> - ttm->pages[i]->mapping = bdev->dev_mapping; >>> -} >>> - >>> int ttm_tt_populate(struct ttm_bo_device *bdev, >>> struct ttm_tt *ttm, struct ttm_operation_ctx *ctx) >>> { >>> @@ -313,7 +302,6 @@ int ttm_tt_populate(struct ttm_bo_device *bdev, >>> if (ret) >>> return ret; >>> >>> - ttm_tt_add_mapping(bdev, ttm); >>> ttm->page_flags |= TTM_PAGE_FLAG_PRIV_POPULATED; >>> if (unlikely(ttm->page_flags & TTM_PAGE_FLAG_SWAPPED)) { >>> ret = ttm_tt_swapin(ttm); > _______________________________________________ dri-devel mailing list dri-devel@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/dri-devel ^ permalink raw reply [flat|nested] 27+ messages in thread
* Re: [PATCH] drm/ttm: don't set page->mapping 2020-11-20 10:08 ` Christian König @ 2020-11-20 15:01 ` Daniel Vetter 0 siblings, 0 replies; 27+ messages in thread From: Daniel Vetter @ 2020-11-20 15:01 UTC (permalink / raw) To: Christian König Cc: Thomas Hellstrom, Daniel Vetter, Intel Graphics Development, LKML, DRI Development, linux-xfs, Linux MM, Huang Rui, Brian Paul, linux-fsdevel, Daniel Vetter On Fri, Nov 20, 2020 at 11:08:31AM +0100, Christian König wrote: > Am 20.11.20 um 11:05 schrieb Daniel Vetter: > > On Fri, Nov 20, 2020 at 11:04 AM Christian König > > <christian.koenig@amd.com> wrote: > > > Am 20.11.20 um 10:54 schrieb Daniel Vetter: > > > > Random observation while trying to review Christian's patch series to > > > > stop looking at struct page for dma-buf imports. > > > > > > > > This was originally added in > > > > > > > > commit 58aa6622d32af7d2c08d45085f44c54554a16ed7 > > > > Author: Thomas Hellstrom <thellstrom@vmware.com> > > > > Date: Fri Jan 3 11:47:23 2014 +0100 > > > > > > > > drm/ttm: Correctly set page mapping and -index members > > > > > > > > Needed for some vm operations; most notably unmap_mapping_range() with > > > > even_cows = 0. > > > > > > > > Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com> > > > > Reviewed-by: Brian Paul <brianp@vmware.com> > > > > > > > > but we do not have a single caller of unmap_mapping_range with > > > > even_cows == 0. And all the gem drivers don't do this, so another > > > > small thing we could standardize between drm and ttm drivers. > > > > > > > > Plus I don't really see a need for unamp_mapping_range where we don't > > > > want to indiscriminately shoot down all ptes. > > > > > > > > Cc: Thomas Hellstrom <thellstrom@vmware.com> > > > > Cc: Brian Paul <brianp@vmware.com> > > > > Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> > > > > Cc: Christian Koenig <christian.koenig@amd.com> > > > > Cc: Huang Rui <ray.huang@amd.com> > > > This is still a NAK as long as we can't come up with a better way to > > > track TTMs page allocations. > > > > > > Additional to that page_mapping() is used quite extensively in the mm > > > code and I'm not sure if that isn't needed for other stuff as well. > > Apologies, I'm honestly not quite sure how this lone patch here ended > > up in this submission. I didn't want to send it out. > > No problem. > > But looking a bit deeper into the mm code that other drm drivers don't set > this correctly and still use unmap_mapping_range() sounds like quite a bug > to me. > > Going to track down what exactly that is used for. Pagecache shootdown. unmap_mapping_range only shoots down from the virtual side. Since that's all we care about, we don't need to set up the address_space in the page. -Daniel > > Christian. > > > -Daniel > > > > > Regards, > > > Christian. > > > > > > > --- > > > > drivers/gpu/drm/ttm/ttm_tt.c | 12 ------------ > > > > 1 file changed, 12 deletions(-) > > > > > > > > diff --git a/drivers/gpu/drm/ttm/ttm_tt.c b/drivers/gpu/drm/ttm/ttm_tt.c > > > > index da9eeffe0c6d..5b2eb6d58bb7 100644 > > > > --- a/drivers/gpu/drm/ttm/ttm_tt.c > > > > +++ b/drivers/gpu/drm/ttm/ttm_tt.c > > > > @@ -284,17 +284,6 @@ int ttm_tt_swapout(struct ttm_bo_device *bdev, struct ttm_tt *ttm) > > > > return ret; > > > > } > > > > > > > > -static void ttm_tt_add_mapping(struct ttm_bo_device *bdev, struct ttm_tt *ttm) > > > > -{ > > > > - pgoff_t i; > > > > - > > > > - if (ttm->page_flags & TTM_PAGE_FLAG_SG) > > > > - return; > > > > - > > > > - for (i = 0; i < ttm->num_pages; ++i) > > > > - ttm->pages[i]->mapping = bdev->dev_mapping; > > > > -} > > > > - > > > > int ttm_tt_populate(struct ttm_bo_device *bdev, > > > > struct ttm_tt *ttm, struct ttm_operation_ctx *ctx) > > > > { > > > > @@ -313,7 +302,6 @@ int ttm_tt_populate(struct ttm_bo_device *bdev, > > > > if (ret) > > > > return ret; > > > > > > > > - ttm_tt_add_mapping(bdev, ttm); > > > > ttm->page_flags |= TTM_PAGE_FLAG_PRIV_POPULATED; > > > > if (unlikely(ttm->page_flags & TTM_PAGE_FLAG_SWAPPED)) { > > > > ret = ttm_tt_swapin(ttm); > > > -- Daniel Vetter Software Engineer, Intel Corporation http://blog.ffwll.ch _______________________________________________ dri-devel mailing list dri-devel@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/dri-devel ^ permalink raw reply [flat|nested] 27+ messages in thread
* [PATCH] drm/ttm: don't set page->mapping @ 2020-11-04 16:50 Daniel Vetter 2020-11-05 7:59 ` Christian König 0 siblings, 1 reply; 27+ messages in thread From: Daniel Vetter @ 2020-11-04 16:50 UTC (permalink / raw) To: DRI Development Cc: Thomas Hellstrom, Daniel Vetter, Huang Rui, Brian Paul, Daniel Vetter, Christian Koenig Random observation while trying to review Christian's patch series to stop looking at struct page for dma-buf imports. This was originally added in commit 58aa6622d32af7d2c08d45085f44c54554a16ed7 Author: Thomas Hellstrom <thellstrom@vmware.com> Date: Fri Jan 3 11:47:23 2014 +0100 drm/ttm: Correctly set page mapping and -index members Needed for some vm operations; most notably unmap_mapping_range() with even_cows = 0. Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com> Reviewed-by: Brian Paul <brianp@vmware.com> but we do not have a single caller of unmap_mapping_range with even_cows == 0. And all the gem drivers don't do this, so another small thing we could standardize between drm and ttm drivers. Plus I don't really see a need for unamp_mapping_range where we don't want to indiscriminately shoot down all ptes. Cc: Thomas Hellstrom <thellstrom@vmware.com> Cc: Brian Paul <brianp@vmware.com> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> Cc: Christian Koenig <christian.koenig@amd.com> Cc: Huang Rui <ray.huang@amd.com> --- drivers/gpu/drm/ttm/ttm_tt.c | 12 ------------ 1 file changed, 12 deletions(-) diff --git a/drivers/gpu/drm/ttm/ttm_tt.c b/drivers/gpu/drm/ttm/ttm_tt.c index 8861a74ac335..438ea43fd8c1 100644 --- a/drivers/gpu/drm/ttm/ttm_tt.c +++ b/drivers/gpu/drm/ttm/ttm_tt.c @@ -291,17 +291,6 @@ int ttm_tt_swapout(struct ttm_bo_device *bdev, struct ttm_tt *ttm) return ret; } -static void ttm_tt_add_mapping(struct ttm_bo_device *bdev, struct ttm_tt *ttm) -{ - pgoff_t i; - - if (ttm->page_flags & TTM_PAGE_FLAG_SG) - return; - - for (i = 0; i < ttm->num_pages; ++i) - ttm->pages[i]->mapping = bdev->dev_mapping; -} - int ttm_tt_populate(struct ttm_bo_device *bdev, struct ttm_tt *ttm, struct ttm_operation_ctx *ctx) { @@ -320,7 +309,6 @@ int ttm_tt_populate(struct ttm_bo_device *bdev, if (ret) return ret; - ttm_tt_add_mapping(bdev, ttm); ttm->page_flags |= TTM_PAGE_FLAG_PRIV_POPULATED; if (unlikely(ttm->page_flags & TTM_PAGE_FLAG_SWAPPED)) { ret = ttm_tt_swapin(ttm); -- 2.28.0 _______________________________________________ dri-devel mailing list dri-devel@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/dri-devel ^ permalink raw reply related [flat|nested] 27+ messages in thread
* Re: [PATCH] drm/ttm: don't set page->mapping 2020-11-04 16:50 Daniel Vetter @ 2020-11-05 7:59 ` Christian König 2020-11-05 9:11 ` Daniel Vetter 0 siblings, 1 reply; 27+ messages in thread From: Christian König @ 2020-11-05 7:59 UTC (permalink / raw) To: Daniel Vetter, DRI Development Cc: Daniel Vetter, Thomas Hellstrom, Huang Rui, Brian Paul Am 04.11.20 um 17:50 schrieb Daniel Vetter: > Random observation while trying to review Christian's patch series to > stop looking at struct page for dma-buf imports. > > This was originally added in > > commit 58aa6622d32af7d2c08d45085f44c54554a16ed7 > Author: Thomas Hellstrom <thellstrom@vmware.com> > Date: Fri Jan 3 11:47:23 2014 +0100 > > drm/ttm: Correctly set page mapping and -index members > > Needed for some vm operations; most notably unmap_mapping_range() with > even_cows = 0. > > Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com> > Reviewed-by: Brian Paul <brianp@vmware.com> > > but we do not have a single caller of unmap_mapping_range with > even_cows == 0. And all the gem drivers don't do this, so another > small thing we could standardize between drm and ttm drivers. > > Plus I don't really see a need for unamp_mapping_range where we don't > want to indiscriminately shoot down all ptes. NAK, we use this to determine if a pages belongs to the driver or not in amdgpu for example. Mostly used for debugging, but I would really like to keep that. Christian. > > Cc: Thomas Hellstrom <thellstrom@vmware.com> > Cc: Brian Paul <brianp@vmware.com> > Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> > Cc: Christian Koenig <christian.koenig@amd.com> > Cc: Huang Rui <ray.huang@amd.com> > --- > drivers/gpu/drm/ttm/ttm_tt.c | 12 ------------ > 1 file changed, 12 deletions(-) > > diff --git a/drivers/gpu/drm/ttm/ttm_tt.c b/drivers/gpu/drm/ttm/ttm_tt.c > index 8861a74ac335..438ea43fd8c1 100644 > --- a/drivers/gpu/drm/ttm/ttm_tt.c > +++ b/drivers/gpu/drm/ttm/ttm_tt.c > @@ -291,17 +291,6 @@ int ttm_tt_swapout(struct ttm_bo_device *bdev, struct ttm_tt *ttm) > return ret; > } > > -static void ttm_tt_add_mapping(struct ttm_bo_device *bdev, struct ttm_tt *ttm) > -{ > - pgoff_t i; > - > - if (ttm->page_flags & TTM_PAGE_FLAG_SG) > - return; > - > - for (i = 0; i < ttm->num_pages; ++i) > - ttm->pages[i]->mapping = bdev->dev_mapping; > -} > - > int ttm_tt_populate(struct ttm_bo_device *bdev, > struct ttm_tt *ttm, struct ttm_operation_ctx *ctx) > { > @@ -320,7 +309,6 @@ int ttm_tt_populate(struct ttm_bo_device *bdev, > if (ret) > return ret; > > - ttm_tt_add_mapping(bdev, ttm); > ttm->page_flags |= TTM_PAGE_FLAG_PRIV_POPULATED; > if (unlikely(ttm->page_flags & TTM_PAGE_FLAG_SWAPPED)) { > ret = ttm_tt_swapin(ttm); _______________________________________________ dri-devel mailing list dri-devel@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/dri-devel ^ permalink raw reply [flat|nested] 27+ messages in thread
* Re: [PATCH] drm/ttm: don't set page->mapping 2020-11-05 7:59 ` Christian König @ 2020-11-05 9:11 ` Daniel Vetter 2020-11-05 12:29 ` Christian König 0 siblings, 1 reply; 27+ messages in thread From: Daniel Vetter @ 2020-11-05 9:11 UTC (permalink / raw) To: Christian König Cc: Daniel Vetter, Thomas Hellstrom, Huang Rui, Brian Paul, DRI Development On Thu, Nov 5, 2020 at 9:00 AM Christian König <christian.koenig@amd.com> wrote: > > Am 04.11.20 um 17:50 schrieb Daniel Vetter: > > Random observation while trying to review Christian's patch series to > > stop looking at struct page for dma-buf imports. > > > > This was originally added in > > > > commit 58aa6622d32af7d2c08d45085f44c54554a16ed7 > > Author: Thomas Hellstrom <thellstrom@vmware.com> > > Date: Fri Jan 3 11:47:23 2014 +0100 > > > > drm/ttm: Correctly set page mapping and -index members > > > > Needed for some vm operations; most notably unmap_mapping_range() with > > even_cows = 0. > > > > Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com> > > Reviewed-by: Brian Paul <brianp@vmware.com> > > > > but we do not have a single caller of unmap_mapping_range with > > even_cows == 0. And all the gem drivers don't do this, so another > > small thing we could standardize between drm and ttm drivers. > > > > Plus I don't really see a need for unamp_mapping_range where we don't > > want to indiscriminately shoot down all ptes. > > NAK, we use this to determine if a pages belongs to the driver or not in > amdgpu for example. > > Mostly used for debugging, but I would really like to keep that. Can you pls point me at that code? A quick grep hasn't really found much at all. -Daniel > > Christian. > > > > > Cc: Thomas Hellstrom <thellstrom@vmware.com> > > Cc: Brian Paul <brianp@vmware.com> > > Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> > > Cc: Christian Koenig <christian.koenig@amd.com> > > Cc: Huang Rui <ray.huang@amd.com> > > --- > > drivers/gpu/drm/ttm/ttm_tt.c | 12 ------------ > > 1 file changed, 12 deletions(-) > > > > diff --git a/drivers/gpu/drm/ttm/ttm_tt.c b/drivers/gpu/drm/ttm/ttm_tt.c > > index 8861a74ac335..438ea43fd8c1 100644 > > --- a/drivers/gpu/drm/ttm/ttm_tt.c > > +++ b/drivers/gpu/drm/ttm/ttm_tt.c > > @@ -291,17 +291,6 @@ int ttm_tt_swapout(struct ttm_bo_device *bdev, struct ttm_tt *ttm) > > return ret; > > } > > > > -static void ttm_tt_add_mapping(struct ttm_bo_device *bdev, struct ttm_tt *ttm) > > -{ > > - pgoff_t i; > > - > > - if (ttm->page_flags & TTM_PAGE_FLAG_SG) > > - return; > > - > > - for (i = 0; i < ttm->num_pages; ++i) > > - ttm->pages[i]->mapping = bdev->dev_mapping; > > -} > > - > > int ttm_tt_populate(struct ttm_bo_device *bdev, > > struct ttm_tt *ttm, struct ttm_operation_ctx *ctx) > > { > > @@ -320,7 +309,6 @@ int ttm_tt_populate(struct ttm_bo_device *bdev, > > if (ret) > > return ret; > > > > - ttm_tt_add_mapping(bdev, ttm); > > ttm->page_flags |= TTM_PAGE_FLAG_PRIV_POPULATED; > > if (unlikely(ttm->page_flags & TTM_PAGE_FLAG_SWAPPED)) { > > ret = ttm_tt_swapin(ttm); > -- Daniel Vetter Software Engineer, Intel Corporation http://blog.ffwll.ch _______________________________________________ dri-devel mailing list dri-devel@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/dri-devel ^ permalink raw reply [flat|nested] 27+ messages in thread
* Re: [PATCH] drm/ttm: don't set page->mapping 2020-11-05 9:11 ` Daniel Vetter @ 2020-11-05 12:29 ` Christian König 2020-11-05 12:50 ` Daniel Vetter 0 siblings, 1 reply; 27+ messages in thread From: Christian König @ 2020-11-05 12:29 UTC (permalink / raw) To: Daniel Vetter Cc: Daniel Vetter, Thomas Hellstrom, Huang Rui, Brian Paul, DRI Development Am 05.11.20 um 10:11 schrieb Daniel Vetter: > On Thu, Nov 5, 2020 at 9:00 AM Christian König <christian.koenig@amd.com> wrote: >> Am 04.11.20 um 17:50 schrieb Daniel Vetter: >>> Random observation while trying to review Christian's patch series to >>> stop looking at struct page for dma-buf imports. >>> >>> This was originally added in >>> >>> commit 58aa6622d32af7d2c08d45085f44c54554a16ed7 >>> Author: Thomas Hellstrom <thellstrom@vmware.com> >>> Date: Fri Jan 3 11:47:23 2014 +0100 >>> >>> drm/ttm: Correctly set page mapping and -index members >>> >>> Needed for some vm operations; most notably unmap_mapping_range() with >>> even_cows = 0. >>> >>> Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com> >>> Reviewed-by: Brian Paul <brianp@vmware.com> >>> >>> but we do not have a single caller of unmap_mapping_range with >>> even_cows == 0. And all the gem drivers don't do this, so another >>> small thing we could standardize between drm and ttm drivers. >>> >>> Plus I don't really see a need for unamp_mapping_range where we don't >>> want to indiscriminately shoot down all ptes. >> NAK, we use this to determine if a pages belongs to the driver or not in >> amdgpu for example. >> >> Mostly used for debugging, but I would really like to keep that. > Can you pls point me at that code? A quick grep hasn't really found much at all. See amdgpu_iomem_read() for an example: > if (p->mapping != adev->mman.bdev.dev_mapping) > return -EPERM; Christian. > -Daniel > >> Christian. >> >>> Cc: Thomas Hellstrom <thellstrom@vmware.com> >>> Cc: Brian Paul <brianp@vmware.com> >>> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> >>> Cc: Christian Koenig <christian.koenig@amd.com> >>> Cc: Huang Rui <ray.huang@amd.com> >>> --- >>> drivers/gpu/drm/ttm/ttm_tt.c | 12 ------------ >>> 1 file changed, 12 deletions(-) >>> >>> diff --git a/drivers/gpu/drm/ttm/ttm_tt.c b/drivers/gpu/drm/ttm/ttm_tt.c >>> index 8861a74ac335..438ea43fd8c1 100644 >>> --- a/drivers/gpu/drm/ttm/ttm_tt.c >>> +++ b/drivers/gpu/drm/ttm/ttm_tt.c >>> @@ -291,17 +291,6 @@ int ttm_tt_swapout(struct ttm_bo_device *bdev, struct ttm_tt *ttm) >>> return ret; >>> } >>> >>> -static void ttm_tt_add_mapping(struct ttm_bo_device *bdev, struct ttm_tt *ttm) >>> -{ >>> - pgoff_t i; >>> - >>> - if (ttm->page_flags & TTM_PAGE_FLAG_SG) >>> - return; >>> - >>> - for (i = 0; i < ttm->num_pages; ++i) >>> - ttm->pages[i]->mapping = bdev->dev_mapping; >>> -} >>> - >>> int ttm_tt_populate(struct ttm_bo_device *bdev, >>> struct ttm_tt *ttm, struct ttm_operation_ctx *ctx) >>> { >>> @@ -320,7 +309,6 @@ int ttm_tt_populate(struct ttm_bo_device *bdev, >>> if (ret) >>> return ret; >>> >>> - ttm_tt_add_mapping(bdev, ttm); >>> ttm->page_flags |= TTM_PAGE_FLAG_PRIV_POPULATED; >>> if (unlikely(ttm->page_flags & TTM_PAGE_FLAG_SWAPPED)) { >>> ret = ttm_tt_swapin(ttm); > _______________________________________________ dri-devel mailing list dri-devel@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/dri-devel ^ permalink raw reply [flat|nested] 27+ messages in thread
* Re: [PATCH] drm/ttm: don't set page->mapping 2020-11-05 12:29 ` Christian König @ 2020-11-05 12:50 ` Daniel Vetter 2020-11-05 12:56 ` Christian König 0 siblings, 1 reply; 27+ messages in thread From: Daniel Vetter @ 2020-11-05 12:50 UTC (permalink / raw) To: Christian König Cc: Thomas Hellstrom, Daniel Vetter, DRI Development, Huang Rui, Brian Paul, Daniel Vetter On Thu, Nov 05, 2020 at 01:29:50PM +0100, Christian König wrote: > Am 05.11.20 um 10:11 schrieb Daniel Vetter: > > On Thu, Nov 5, 2020 at 9:00 AM Christian König <christian.koenig@amd.com> wrote: > > > Am 04.11.20 um 17:50 schrieb Daniel Vetter: > > > > Random observation while trying to review Christian's patch series to > > > > stop looking at struct page for dma-buf imports. > > > > > > > > This was originally added in > > > > > > > > commit 58aa6622d32af7d2c08d45085f44c54554a16ed7 > > > > Author: Thomas Hellstrom <thellstrom@vmware.com> > > > > Date: Fri Jan 3 11:47:23 2014 +0100 > > > > > > > > drm/ttm: Correctly set page mapping and -index members > > > > > > > > Needed for some vm operations; most notably unmap_mapping_range() with > > > > even_cows = 0. > > > > > > > > Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com> > > > > Reviewed-by: Brian Paul <brianp@vmware.com> > > > > > > > > but we do not have a single caller of unmap_mapping_range with > > > > even_cows == 0. And all the gem drivers don't do this, so another > > > > small thing we could standardize between drm and ttm drivers. > > > > > > > > Plus I don't really see a need for unamp_mapping_range where we don't > > > > want to indiscriminately shoot down all ptes. > > > NAK, we use this to determine if a pages belongs to the driver or not in > > > amdgpu for example. > > > > > > Mostly used for debugging, but I would really like to keep that. > > Can you pls point me at that code? A quick grep hasn't really found much at all. > > See amdgpu_iomem_read() for an example: Why do you reject this? If this is to avoid issues with userptr, then I think there's a simple trick: - grab page reference - recheck that the iova still points at the same address - do read/write, safe in the knowledge that this page cannot be reused for anything else - drop page reference Of course this can still race against iova updates, but that seems to be a fundamental part of your debug interface here. Or am I missing something? Just pondering this more since setting the page->mapping pointer for just this seems somewhat wild abuse of ->mapping semantics :-) -Daniel > > > if (p->mapping != adev->mman.bdev.dev_mapping) > > return -EPERM; > > Christian. > > > -Daniel > > > > > Christian. > > > > > > > Cc: Thomas Hellstrom <thellstrom@vmware.com> > > > > Cc: Brian Paul <brianp@vmware.com> > > > > Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> > > > > Cc: Christian Koenig <christian.koenig@amd.com> > > > > Cc: Huang Rui <ray.huang@amd.com> > > > > --- > > > > drivers/gpu/drm/ttm/ttm_tt.c | 12 ------------ > > > > 1 file changed, 12 deletions(-) > > > > > > > > diff --git a/drivers/gpu/drm/ttm/ttm_tt.c b/drivers/gpu/drm/ttm/ttm_tt.c > > > > index 8861a74ac335..438ea43fd8c1 100644 > > > > --- a/drivers/gpu/drm/ttm/ttm_tt.c > > > > +++ b/drivers/gpu/drm/ttm/ttm_tt.c > > > > @@ -291,17 +291,6 @@ int ttm_tt_swapout(struct ttm_bo_device *bdev, struct ttm_tt *ttm) > > > > return ret; > > > > } > > > > > > > > -static void ttm_tt_add_mapping(struct ttm_bo_device *bdev, struct ttm_tt *ttm) > > > > -{ > > > > - pgoff_t i; > > > > - > > > > - if (ttm->page_flags & TTM_PAGE_FLAG_SG) > > > > - return; > > > > - > > > > - for (i = 0; i < ttm->num_pages; ++i) > > > > - ttm->pages[i]->mapping = bdev->dev_mapping; > > > > -} > > > > - > > > > int ttm_tt_populate(struct ttm_bo_device *bdev, > > > > struct ttm_tt *ttm, struct ttm_operation_ctx *ctx) > > > > { > > > > @@ -320,7 +309,6 @@ int ttm_tt_populate(struct ttm_bo_device *bdev, > > > > if (ret) > > > > return ret; > > > > > > > > - ttm_tt_add_mapping(bdev, ttm); > > > > ttm->page_flags |= TTM_PAGE_FLAG_PRIV_POPULATED; > > > > if (unlikely(ttm->page_flags & TTM_PAGE_FLAG_SWAPPED)) { > > > > ret = ttm_tt_swapin(ttm); > > > -- Daniel Vetter Software Engineer, Intel Corporation http://blog.ffwll.ch _______________________________________________ dri-devel mailing list dri-devel@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/dri-devel ^ permalink raw reply [flat|nested] 27+ messages in thread
* Re: [PATCH] drm/ttm: don't set page->mapping 2020-11-05 12:50 ` Daniel Vetter @ 2020-11-05 12:56 ` Christian König 2020-11-05 13:20 ` Daniel Vetter 0 siblings, 1 reply; 27+ messages in thread From: Christian König @ 2020-11-05 12:56 UTC (permalink / raw) To: Daniel Vetter Cc: Thomas Hellstrom, Daniel Vetter, DRI Development, Huang Rui, Brian Paul, Daniel Vetter Am 05.11.20 um 13:50 schrieb Daniel Vetter: > On Thu, Nov 05, 2020 at 01:29:50PM +0100, Christian König wrote: >> Am 05.11.20 um 10:11 schrieb Daniel Vetter: >>> On Thu, Nov 5, 2020 at 9:00 AM Christian König <christian.koenig@amd.com> wrote: >>>> Am 04.11.20 um 17:50 schrieb Daniel Vetter: >>>>> Random observation while trying to review Christian's patch series to >>>>> stop looking at struct page for dma-buf imports. >>>>> >>>>> This was originally added in >>>>> >>>>> commit 58aa6622d32af7d2c08d45085f44c54554a16ed7 >>>>> Author: Thomas Hellstrom <thellstrom@vmware.com> >>>>> Date: Fri Jan 3 11:47:23 2014 +0100 >>>>> >>>>> drm/ttm: Correctly set page mapping and -index members >>>>> >>>>> Needed for some vm operations; most notably unmap_mapping_range() with >>>>> even_cows = 0. >>>>> >>>>> Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com> >>>>> Reviewed-by: Brian Paul <brianp@vmware.com> >>>>> >>>>> but we do not have a single caller of unmap_mapping_range with >>>>> even_cows == 0. And all the gem drivers don't do this, so another >>>>> small thing we could standardize between drm and ttm drivers. >>>>> >>>>> Plus I don't really see a need for unamp_mapping_range where we don't >>>>> want to indiscriminately shoot down all ptes. >>>> NAK, we use this to determine if a pages belongs to the driver or not in >>>> amdgpu for example. >>>> >>>> Mostly used for debugging, but I would really like to keep that. >>> Can you pls point me at that code? A quick grep hasn't really found much at all. >> See amdgpu_iomem_read() for an example: > Why do you reject this? When IOMMU is disabled or uses an 1 to 1 mapping we would otherwise give the same access as /dev/mem to system memory and that is forbidden. But as I noted this is just for the debugfs file. When I tried a few years ago to not set the page->mapping I immediately ran into issues with our eviction test. So I think that this is used elsewhere as well. Regards, Christian. > > If this is to avoid issues with userptr, then I think there's a simple > trick: > - grab page reference > - recheck that the iova still points at the same address > - do read/write, safe in the knowledge that this page cannot be reused for > anything else > - drop page reference > > Of course this can still race against iova updates, but that seems to be a > fundamental part of your debug interface here. > > Or am I missing something? > > Just pondering this more since setting the page->mapping pointer for just > this seems somewhat wild abuse of ->mapping semantics :-) > -Daniel > >>> if (p->mapping != adev->mman.bdev.dev_mapping) >>> return -EPERM; >> Christian. >> >>> -Daniel >>> >>>> Christian. >>>> >>>>> Cc: Thomas Hellstrom <thellstrom@vmware.com> >>>>> Cc: Brian Paul <brianp@vmware.com> >>>>> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> >>>>> Cc: Christian Koenig <christian.koenig@amd.com> >>>>> Cc: Huang Rui <ray.huang@amd.com> >>>>> --- >>>>> drivers/gpu/drm/ttm/ttm_tt.c | 12 ------------ >>>>> 1 file changed, 12 deletions(-) >>>>> >>>>> diff --git a/drivers/gpu/drm/ttm/ttm_tt.c b/drivers/gpu/drm/ttm/ttm_tt.c >>>>> index 8861a74ac335..438ea43fd8c1 100644 >>>>> --- a/drivers/gpu/drm/ttm/ttm_tt.c >>>>> +++ b/drivers/gpu/drm/ttm/ttm_tt.c >>>>> @@ -291,17 +291,6 @@ int ttm_tt_swapout(struct ttm_bo_device *bdev, struct ttm_tt *ttm) >>>>> return ret; >>>>> } >>>>> >>>>> -static void ttm_tt_add_mapping(struct ttm_bo_device *bdev, struct ttm_tt *ttm) >>>>> -{ >>>>> - pgoff_t i; >>>>> - >>>>> - if (ttm->page_flags & TTM_PAGE_FLAG_SG) >>>>> - return; >>>>> - >>>>> - for (i = 0; i < ttm->num_pages; ++i) >>>>> - ttm->pages[i]->mapping = bdev->dev_mapping; >>>>> -} >>>>> - >>>>> int ttm_tt_populate(struct ttm_bo_device *bdev, >>>>> struct ttm_tt *ttm, struct ttm_operation_ctx *ctx) >>>>> { >>>>> @@ -320,7 +309,6 @@ int ttm_tt_populate(struct ttm_bo_device *bdev, >>>>> if (ret) >>>>> return ret; >>>>> >>>>> - ttm_tt_add_mapping(bdev, ttm); >>>>> ttm->page_flags |= TTM_PAGE_FLAG_PRIV_POPULATED; >>>>> if (unlikely(ttm->page_flags & TTM_PAGE_FLAG_SWAPPED)) { >>>>> ret = ttm_tt_swapin(ttm); _______________________________________________ dri-devel mailing list dri-devel@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/dri-devel ^ permalink raw reply [flat|nested] 27+ messages in thread
* Re: [PATCH] drm/ttm: don't set page->mapping 2020-11-05 12:56 ` Christian König @ 2020-11-05 13:20 ` Daniel Vetter 2020-11-05 13:22 ` Christian König 0 siblings, 1 reply; 27+ messages in thread From: Daniel Vetter @ 2020-11-05 13:20 UTC (permalink / raw) To: Christian König Cc: Thomas Hellstrom, Daniel Vetter, DRI Development, Huang Rui, Brian Paul, Daniel Vetter On Thu, Nov 05, 2020 at 01:56:22PM +0100, Christian König wrote: > Am 05.11.20 um 13:50 schrieb Daniel Vetter: > > On Thu, Nov 05, 2020 at 01:29:50PM +0100, Christian König wrote: > > > Am 05.11.20 um 10:11 schrieb Daniel Vetter: > > > > On Thu, Nov 5, 2020 at 9:00 AM Christian König <christian.koenig@amd.com> wrote: > > > > > Am 04.11.20 um 17:50 schrieb Daniel Vetter: > > > > > > Random observation while trying to review Christian's patch series to > > > > > > stop looking at struct page for dma-buf imports. > > > > > > > > > > > > This was originally added in > > > > > > > > > > > > commit 58aa6622d32af7d2c08d45085f44c54554a16ed7 > > > > > > Author: Thomas Hellstrom <thellstrom@vmware.com> > > > > > > Date: Fri Jan 3 11:47:23 2014 +0100 > > > > > > > > > > > > drm/ttm: Correctly set page mapping and -index members > > > > > > > > > > > > Needed for some vm operations; most notably unmap_mapping_range() with > > > > > > even_cows = 0. > > > > > > > > > > > > Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com> > > > > > > Reviewed-by: Brian Paul <brianp@vmware.com> > > > > > > > > > > > > but we do not have a single caller of unmap_mapping_range with > > > > > > even_cows == 0. And all the gem drivers don't do this, so another > > > > > > small thing we could standardize between drm and ttm drivers. > > > > > > > > > > > > Plus I don't really see a need for unamp_mapping_range where we don't > > > > > > want to indiscriminately shoot down all ptes. > > > > > NAK, we use this to determine if a pages belongs to the driver or not in > > > > > amdgpu for example. > > > > > > > > > > Mostly used for debugging, but I would really like to keep that. > > > > Can you pls point me at that code? A quick grep hasn't really found much at all. > > > See amdgpu_iomem_read() for an example: > > Why do you reject this? > > When IOMMU is disabled or uses an 1 to 1 mapping we would otherwise give the > same access as /dev/mem to system memory and that is forbidden. But as I > noted this is just for the debugfs file. Ah, there's a config option for that. Plus it's debugfs, anything goes in debugfs, but if you're worried about that hole we should just disable the entire debugfs file for CONFIG_STRICT_DEVMEM. I can perhaps throw that on top, that follow_pfn patch series I'm baking is all about this kind of fun. > When I tried a few years ago to not set the page->mapping I immediately ran > into issues with our eviction test. So I think that this is used elsewhere > as well. That's the kind of interaction I'm worried about here tbh. If this does some kind of shrinking of some sorts, I think a real shrinker should take over. An improved grep shows nothing else, so the only the above is the only thing I can think of. What kind of eviction test goes boom if you clear ->mapping here? I'd be happy to type up the clever trick for the debugfs files. -Daniel > > Regards, > Christian. > > > > > If this is to avoid issues with userptr, then I think there's a simple > > trick: > > - grab page reference > > - recheck that the iova still points at the same address > > - do read/write, safe in the knowledge that this page cannot be reused for > > anything else > > - drop page reference > > > > Of course this can still race against iova updates, but that seems to be a > > fundamental part of your debug interface here. > > > > Or am I missing something? > > > > Just pondering this more since setting the page->mapping pointer for just > > this seems somewhat wild abuse of ->mapping semantics :-) > > -Daniel > > > > > > if (p->mapping != adev->mman.bdev.dev_mapping) > > > > return -EPERM; > > > Christian. > > > > > > > -Daniel > > > > > > > > > Christian. > > > > > > > > > > > Cc: Thomas Hellstrom <thellstrom@vmware.com> > > > > > > Cc: Brian Paul <brianp@vmware.com> > > > > > > Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> > > > > > > Cc: Christian Koenig <christian.koenig@amd.com> > > > > > > Cc: Huang Rui <ray.huang@amd.com> > > > > > > --- > > > > > > drivers/gpu/drm/ttm/ttm_tt.c | 12 ------------ > > > > > > 1 file changed, 12 deletions(-) > > > > > > > > > > > > diff --git a/drivers/gpu/drm/ttm/ttm_tt.c b/drivers/gpu/drm/ttm/ttm_tt.c > > > > > > index 8861a74ac335..438ea43fd8c1 100644 > > > > > > --- a/drivers/gpu/drm/ttm/ttm_tt.c > > > > > > +++ b/drivers/gpu/drm/ttm/ttm_tt.c > > > > > > @@ -291,17 +291,6 @@ int ttm_tt_swapout(struct ttm_bo_device *bdev, struct ttm_tt *ttm) > > > > > > return ret; > > > > > > } > > > > > > > > > > > > -static void ttm_tt_add_mapping(struct ttm_bo_device *bdev, struct ttm_tt *ttm) > > > > > > -{ > > > > > > - pgoff_t i; > > > > > > - > > > > > > - if (ttm->page_flags & TTM_PAGE_FLAG_SG) > > > > > > - return; > > > > > > - > > > > > > - for (i = 0; i < ttm->num_pages; ++i) > > > > > > - ttm->pages[i]->mapping = bdev->dev_mapping; > > > > > > -} > > > > > > - > > > > > > int ttm_tt_populate(struct ttm_bo_device *bdev, > > > > > > struct ttm_tt *ttm, struct ttm_operation_ctx *ctx) > > > > > > { > > > > > > @@ -320,7 +309,6 @@ int ttm_tt_populate(struct ttm_bo_device *bdev, > > > > > > if (ret) > > > > > > return ret; > > > > > > > > > > > > - ttm_tt_add_mapping(bdev, ttm); > > > > > > ttm->page_flags |= TTM_PAGE_FLAG_PRIV_POPULATED; > > > > > > if (unlikely(ttm->page_flags & TTM_PAGE_FLAG_SWAPPED)) { > > > > > > ret = ttm_tt_swapin(ttm); > -- Daniel Vetter Software Engineer, Intel Corporation http://blog.ffwll.ch _______________________________________________ dri-devel mailing list dri-devel@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/dri-devel ^ permalink raw reply [flat|nested] 27+ messages in thread
* Re: [PATCH] drm/ttm: don't set page->mapping 2020-11-05 13:20 ` Daniel Vetter @ 2020-11-05 13:22 ` Christian König 2020-11-05 14:31 ` Daniel Vetter 0 siblings, 1 reply; 27+ messages in thread From: Christian König @ 2020-11-05 13:22 UTC (permalink / raw) To: Daniel Vetter Cc: Thomas Hellstrom, Daniel Vetter, DRI Development, Huang Rui, Brian Paul, Daniel Vetter Am 05.11.20 um 14:20 schrieb Daniel Vetter: > On Thu, Nov 05, 2020 at 01:56:22PM +0100, Christian König wrote: >> Am 05.11.20 um 13:50 schrieb Daniel Vetter: >>> On Thu, Nov 05, 2020 at 01:29:50PM +0100, Christian König wrote: >>>> Am 05.11.20 um 10:11 schrieb Daniel Vetter: >>>>> On Thu, Nov 5, 2020 at 9:00 AM Christian König <christian.koenig@amd.com> wrote: >>>>>> Am 04.11.20 um 17:50 schrieb Daniel Vetter: >>>>>>> Random observation while trying to review Christian's patch series to >>>>>>> stop looking at struct page for dma-buf imports. >>>>>>> >>>>>>> This was originally added in >>>>>>> >>>>>>> commit 58aa6622d32af7d2c08d45085f44c54554a16ed7 >>>>>>> Author: Thomas Hellstrom <thellstrom@vmware.com> >>>>>>> Date: Fri Jan 3 11:47:23 2014 +0100 >>>>>>> >>>>>>> drm/ttm: Correctly set page mapping and -index members >>>>>>> >>>>>>> Needed for some vm operations; most notably unmap_mapping_range() with >>>>>>> even_cows = 0. >>>>>>> >>>>>>> Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com> >>>>>>> Reviewed-by: Brian Paul <brianp@vmware.com> >>>>>>> >>>>>>> but we do not have a single caller of unmap_mapping_range with >>>>>>> even_cows == 0. And all the gem drivers don't do this, so another >>>>>>> small thing we could standardize between drm and ttm drivers. >>>>>>> >>>>>>> Plus I don't really see a need for unamp_mapping_range where we don't >>>>>>> want to indiscriminately shoot down all ptes. >>>>>> NAK, we use this to determine if a pages belongs to the driver or not in >>>>>> amdgpu for example. >>>>>> >>>>>> Mostly used for debugging, but I would really like to keep that. >>>>> Can you pls point me at that code? A quick grep hasn't really found much at all. >>>> See amdgpu_iomem_read() for an example: >>> Why do you reject this? >> When IOMMU is disabled or uses an 1 to 1 mapping we would otherwise give the >> same access as /dev/mem to system memory and that is forbidden. But as I >> noted this is just for the debugfs file. > Ah, there's a config option for that. Plus it's debugfs, anything goes in > debugfs, but if you're worried about that hole we should just disable the > entire debugfs file for CONFIG_STRICT_DEVMEM. I can perhaps throw that on > top, that follow_pfn patch series I'm baking is all about this kind of > fun. And exactly that would get a NAK from us. We have specially created that debugfs file as an alternative when CONFIG_STRICT_DEVMEM is set. Regards, Christian. > >> When I tried a few years ago to not set the page->mapping I immediately ran >> into issues with our eviction test. So I think that this is used elsewhere >> as well. > That's the kind of interaction I'm worried about here tbh. If this does > some kind of shrinking of some sorts, I think a real shrinker should take > over. > > An improved grep shows nothing else, so the only the above is the only > thing I can think of. What kind of eviction test goes boom if you clear > ->mapping here? I'd be happy to type up the clever trick for the debugfs > files. > -Daniel > >> Regards, >> Christian. >> >>> If this is to avoid issues with userptr, then I think there's a simple >>> trick: >>> - grab page reference >>> - recheck that the iova still points at the same address >>> - do read/write, safe in the knowledge that this page cannot be reused for >>> anything else >>> - drop page reference >>> >>> Of course this can still race against iova updates, but that seems to be a >>> fundamental part of your debug interface here. >>> >>> Or am I missing something? >>> >>> Just pondering this more since setting the page->mapping pointer for just >>> this seems somewhat wild abuse of ->mapping semantics :-) >>> -Daniel >>> >>>>> if (p->mapping != adev->mman.bdev.dev_mapping) >>>>> return -EPERM; >>>> Christian. >>>> >>>>> -Daniel >>>>> >>>>>> Christian. >>>>>> >>>>>>> Cc: Thomas Hellstrom <thellstrom@vmware.com> >>>>>>> Cc: Brian Paul <brianp@vmware.com> >>>>>>> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> >>>>>>> Cc: Christian Koenig <christian.koenig@amd.com> >>>>>>> Cc: Huang Rui <ray.huang@amd.com> >>>>>>> --- >>>>>>> drivers/gpu/drm/ttm/ttm_tt.c | 12 ------------ >>>>>>> 1 file changed, 12 deletions(-) >>>>>>> >>>>>>> diff --git a/drivers/gpu/drm/ttm/ttm_tt.c b/drivers/gpu/drm/ttm/ttm_tt.c >>>>>>> index 8861a74ac335..438ea43fd8c1 100644 >>>>>>> --- a/drivers/gpu/drm/ttm/ttm_tt.c >>>>>>> +++ b/drivers/gpu/drm/ttm/ttm_tt.c >>>>>>> @@ -291,17 +291,6 @@ int ttm_tt_swapout(struct ttm_bo_device *bdev, struct ttm_tt *ttm) >>>>>>> return ret; >>>>>>> } >>>>>>> >>>>>>> -static void ttm_tt_add_mapping(struct ttm_bo_device *bdev, struct ttm_tt *ttm) >>>>>>> -{ >>>>>>> - pgoff_t i; >>>>>>> - >>>>>>> - if (ttm->page_flags & TTM_PAGE_FLAG_SG) >>>>>>> - return; >>>>>>> - >>>>>>> - for (i = 0; i < ttm->num_pages; ++i) >>>>>>> - ttm->pages[i]->mapping = bdev->dev_mapping; >>>>>>> -} >>>>>>> - >>>>>>> int ttm_tt_populate(struct ttm_bo_device *bdev, >>>>>>> struct ttm_tt *ttm, struct ttm_operation_ctx *ctx) >>>>>>> { >>>>>>> @@ -320,7 +309,6 @@ int ttm_tt_populate(struct ttm_bo_device *bdev, >>>>>>> if (ret) >>>>>>> return ret; >>>>>>> >>>>>>> - ttm_tt_add_mapping(bdev, ttm); >>>>>>> ttm->page_flags |= TTM_PAGE_FLAG_PRIV_POPULATED; >>>>>>> if (unlikely(ttm->page_flags & TTM_PAGE_FLAG_SWAPPED)) { >>>>>>> ret = ttm_tt_swapin(ttm); _______________________________________________ dri-devel mailing list dri-devel@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/dri-devel ^ permalink raw reply [flat|nested] 27+ messages in thread
* Re: [PATCH] drm/ttm: don't set page->mapping 2020-11-05 13:22 ` Christian König @ 2020-11-05 14:31 ` Daniel Vetter 2020-11-05 14:38 ` Daniel Vetter 0 siblings, 1 reply; 27+ messages in thread From: Daniel Vetter @ 2020-11-05 14:31 UTC (permalink / raw) To: Christian König Cc: Daniel Vetter, Thomas Hellstrom, Huang Rui, Brian Paul, DRI Development On Thu, Nov 5, 2020 at 2:22 PM Christian König <christian.koenig@amd.com> wrote: > > Am 05.11.20 um 14:20 schrieb Daniel Vetter: > > On Thu, Nov 05, 2020 at 01:56:22PM +0100, Christian König wrote: > >> Am 05.11.20 um 13:50 schrieb Daniel Vetter: > >>> On Thu, Nov 05, 2020 at 01:29:50PM +0100, Christian König wrote: > >>>> Am 05.11.20 um 10:11 schrieb Daniel Vetter: > >>>>> On Thu, Nov 5, 2020 at 9:00 AM Christian König <christian.koenig@amd.com> wrote: > >>>>>> Am 04.11.20 um 17:50 schrieb Daniel Vetter: > >>>>>>> Random observation while trying to review Christian's patch series to > >>>>>>> stop looking at struct page for dma-buf imports. > >>>>>>> > >>>>>>> This was originally added in > >>>>>>> > >>>>>>> commit 58aa6622d32af7d2c08d45085f44c54554a16ed7 > >>>>>>> Author: Thomas Hellstrom <thellstrom@vmware.com> > >>>>>>> Date: Fri Jan 3 11:47:23 2014 +0100 > >>>>>>> > >>>>>>> drm/ttm: Correctly set page mapping and -index members > >>>>>>> > >>>>>>> Needed for some vm operations; most notably unmap_mapping_range() with > >>>>>>> even_cows = 0. > >>>>>>> > >>>>>>> Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com> > >>>>>>> Reviewed-by: Brian Paul <brianp@vmware.com> > >>>>>>> > >>>>>>> but we do not have a single caller of unmap_mapping_range with > >>>>>>> even_cows == 0. And all the gem drivers don't do this, so another > >>>>>>> small thing we could standardize between drm and ttm drivers. > >>>>>>> > >>>>>>> Plus I don't really see a need for unamp_mapping_range where we don't > >>>>>>> want to indiscriminately shoot down all ptes. > >>>>>> NAK, we use this to determine if a pages belongs to the driver or not in > >>>>>> amdgpu for example. > >>>>>> > >>>>>> Mostly used for debugging, but I would really like to keep that. > >>>>> Can you pls point me at that code? A quick grep hasn't really found much at all. > >>>> See amdgpu_iomem_read() for an example: > >>> Why do you reject this? > >> When IOMMU is disabled or uses an 1 to 1 mapping we would otherwise give the > >> same access as /dev/mem to system memory and that is forbidden. But as I > >> noted this is just for the debugfs file. > > Ah, there's a config option for that. Plus it's debugfs, anything goes in > > debugfs, but if you're worried about that hole we should just disable the > > entire debugfs file for CONFIG_STRICT_DEVMEM. I can perhaps throw that on > > top, that follow_pfn patch series I'm baking is all about this kind of > > fun. > > And exactly that would get a NAK from us. > > We have specially created that debugfs file as an alternative when > CONFIG_STRICT_DEVMEM is set. Uh that doesn't work if you work around core restrictions with your own debugfs paths. Maybe you can do fun like this in your dkms, but not in upstream. Like if this was specifically created to work around CONFIG_STRICT_DEVMEM (and it sounds like that) then I think this should never have landed in upstream to begin with. -Daniel > > Regards, > Christian. > > > > >> When I tried a few years ago to not set the page->mapping I immediately ran > >> into issues with our eviction test. So I think that this is used elsewhere > >> as well. > > That's the kind of interaction I'm worried about here tbh. If this does > > some kind of shrinking of some sorts, I think a real shrinker should take > > over. > > > > An improved grep shows nothing else, so the only the above is the only > > thing I can think of. What kind of eviction test goes boom if you clear > > ->mapping here? I'd be happy to type up the clever trick for the debugfs > > files. > > -Daniel > > > >> Regards, > >> Christian. > >> > >>> If this is to avoid issues with userptr, then I think there's a simple > >>> trick: > >>> - grab page reference > >>> - recheck that the iova still points at the same address > >>> - do read/write, safe in the knowledge that this page cannot be reused for > >>> anything else > >>> - drop page reference > >>> > >>> Of course this can still race against iova updates, but that seems to be a > >>> fundamental part of your debug interface here. > >>> > >>> Or am I missing something? > >>> > >>> Just pondering this more since setting the page->mapping pointer for just > >>> this seems somewhat wild abuse of ->mapping semantics :-) > >>> -Daniel > >>> > >>>>> if (p->mapping != adev->mman.bdev.dev_mapping) > >>>>> return -EPERM; > >>>> Christian. > >>>> > >>>>> -Daniel > >>>>> > >>>>>> Christian. > >>>>>> > >>>>>>> Cc: Thomas Hellstrom <thellstrom@vmware.com> > >>>>>>> Cc: Brian Paul <brianp@vmware.com> > >>>>>>> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> > >>>>>>> Cc: Christian Koenig <christian.koenig@amd.com> > >>>>>>> Cc: Huang Rui <ray.huang@amd.com> > >>>>>>> --- > >>>>>>> drivers/gpu/drm/ttm/ttm_tt.c | 12 ------------ > >>>>>>> 1 file changed, 12 deletions(-) > >>>>>>> > >>>>>>> diff --git a/drivers/gpu/drm/ttm/ttm_tt.c b/drivers/gpu/drm/ttm/ttm_tt.c > >>>>>>> index 8861a74ac335..438ea43fd8c1 100644 > >>>>>>> --- a/drivers/gpu/drm/ttm/ttm_tt.c > >>>>>>> +++ b/drivers/gpu/drm/ttm/ttm_tt.c > >>>>>>> @@ -291,17 +291,6 @@ int ttm_tt_swapout(struct ttm_bo_device *bdev, struct ttm_tt *ttm) > >>>>>>> return ret; > >>>>>>> } > >>>>>>> > >>>>>>> -static void ttm_tt_add_mapping(struct ttm_bo_device *bdev, struct ttm_tt *ttm) > >>>>>>> -{ > >>>>>>> - pgoff_t i; > >>>>>>> - > >>>>>>> - if (ttm->page_flags & TTM_PAGE_FLAG_SG) > >>>>>>> - return; > >>>>>>> - > >>>>>>> - for (i = 0; i < ttm->num_pages; ++i) > >>>>>>> - ttm->pages[i]->mapping = bdev->dev_mapping; > >>>>>>> -} > >>>>>>> - > >>>>>>> int ttm_tt_populate(struct ttm_bo_device *bdev, > >>>>>>> struct ttm_tt *ttm, struct ttm_operation_ctx *ctx) > >>>>>>> { > >>>>>>> @@ -320,7 +309,6 @@ int ttm_tt_populate(struct ttm_bo_device *bdev, > >>>>>>> if (ret) > >>>>>>> return ret; > >>>>>>> > >>>>>>> - ttm_tt_add_mapping(bdev, ttm); > >>>>>>> ttm->page_flags |= TTM_PAGE_FLAG_PRIV_POPULATED; > >>>>>>> if (unlikely(ttm->page_flags & TTM_PAGE_FLAG_SWAPPED)) { > >>>>>>> ret = ttm_tt_swapin(ttm); > -- Daniel Vetter Software Engineer, Intel Corporation http://blog.ffwll.ch _______________________________________________ dri-devel mailing list dri-devel@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/dri-devel ^ permalink raw reply [flat|nested] 27+ messages in thread
* Re: [PATCH] drm/ttm: don't set page->mapping 2020-11-05 14:31 ` Daniel Vetter @ 2020-11-05 14:38 ` Daniel Vetter 2020-11-05 15:15 ` Christian König 0 siblings, 1 reply; 27+ messages in thread From: Daniel Vetter @ 2020-11-05 14:38 UTC (permalink / raw) To: Christian König Cc: Daniel Vetter, Thomas Hellstrom, Huang Rui, Brian Paul, DRI Development On Thu, Nov 5, 2020 at 3:31 PM Daniel Vetter <daniel@ffwll.ch> wrote: > > On Thu, Nov 5, 2020 at 2:22 PM Christian König <christian.koenig@amd.com> wrote: > > > > Am 05.11.20 um 14:20 schrieb Daniel Vetter: > > > On Thu, Nov 05, 2020 at 01:56:22PM +0100, Christian König wrote: > > >> Am 05.11.20 um 13:50 schrieb Daniel Vetter: > > >>> On Thu, Nov 05, 2020 at 01:29:50PM +0100, Christian König wrote: > > >>>> Am 05.11.20 um 10:11 schrieb Daniel Vetter: > > >>>>> On Thu, Nov 5, 2020 at 9:00 AM Christian König <christian.koenig@amd.com> wrote: > > >>>>>> Am 04.11.20 um 17:50 schrieb Daniel Vetter: > > >>>>>>> Random observation while trying to review Christian's patch series to > > >>>>>>> stop looking at struct page for dma-buf imports. > > >>>>>>> > > >>>>>>> This was originally added in > > >>>>>>> > > >>>>>>> commit 58aa6622d32af7d2c08d45085f44c54554a16ed7 > > >>>>>>> Author: Thomas Hellstrom <thellstrom@vmware.com> > > >>>>>>> Date: Fri Jan 3 11:47:23 2014 +0100 > > >>>>>>> > > >>>>>>> drm/ttm: Correctly set page mapping and -index members > > >>>>>>> > > >>>>>>> Needed for some vm operations; most notably unmap_mapping_range() with > > >>>>>>> even_cows = 0. > > >>>>>>> > > >>>>>>> Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com> > > >>>>>>> Reviewed-by: Brian Paul <brianp@vmware.com> > > >>>>>>> > > >>>>>>> but we do not have a single caller of unmap_mapping_range with > > >>>>>>> even_cows == 0. And all the gem drivers don't do this, so another > > >>>>>>> small thing we could standardize between drm and ttm drivers. > > >>>>>>> > > >>>>>>> Plus I don't really see a need for unamp_mapping_range where we don't > > >>>>>>> want to indiscriminately shoot down all ptes. > > >>>>>> NAK, we use this to determine if a pages belongs to the driver or not in > > >>>>>> amdgpu for example. > > >>>>>> > > >>>>>> Mostly used for debugging, but I would really like to keep that. > > >>>>> Can you pls point me at that code? A quick grep hasn't really found much at all. > > >>>> See amdgpu_iomem_read() for an example: > > >>> Why do you reject this? > > >> When IOMMU is disabled or uses an 1 to 1 mapping we would otherwise give the > > >> same access as /dev/mem to system memory and that is forbidden. But as I > > >> noted this is just for the debugfs file. > > > Ah, there's a config option for that. Plus it's debugfs, anything goes in > > > debugfs, but if you're worried about that hole we should just disable the > > > entire debugfs file for CONFIG_STRICT_DEVMEM. I can perhaps throw that on > > > top, that follow_pfn patch series I'm baking is all about this kind of > > > fun. > > > > And exactly that would get a NAK from us. > > > > We have specially created that debugfs file as an alternative when > > CONFIG_STRICT_DEVMEM is set. > > Uh that doesn't work if you work around core restrictions with your > own debugfs paths. Maybe you can do fun like this in your dkms, but > not in upstream. Like if this was specifically created to work around > CONFIG_STRICT_DEVMEM (and it sounds like that) then I think this > should never have landed in upstream to begin with. I'm also kinda confused that there's distros with CONFIG_STRICT_DEVMEM which allow debugfs. debugfs is a pretty bad root hole all around, or at least that's been the assumption all the time. -Daniel > > >> When I tried a few years ago to not set the page->mapping I immediately ran > > >> into issues with our eviction test. So I think that this is used elsewhere > > >> as well. > > > That's the kind of interaction I'm worried about here tbh. If this does > > > some kind of shrinking of some sorts, I think a real shrinker should take > > > over. > > > > > > An improved grep shows nothing else, so the only the above is the only > > > thing I can think of. What kind of eviction test goes boom if you clear > > > ->mapping here? I'd be happy to type up the clever trick for the debugfs > > > files. > > > -Daniel > > > > > >> Regards, > > >> Christian. > > >> > > >>> If this is to avoid issues with userptr, then I think there's a simple > > >>> trick: > > >>> - grab page reference > > >>> - recheck that the iova still points at the same address > > >>> - do read/write, safe in the knowledge that this page cannot be reused for > > >>> anything else > > >>> - drop page reference > > >>> > > >>> Of course this can still race against iova updates, but that seems to be a > > >>> fundamental part of your debug interface here. > > >>> > > >>> Or am I missing something? > > >>> > > >>> Just pondering this more since setting the page->mapping pointer for just > > >>> this seems somewhat wild abuse of ->mapping semantics :-) > > >>> -Daniel > > >>> > > >>>>> if (p->mapping != adev->mman.bdev.dev_mapping) > > >>>>> return -EPERM; > > >>>> Christian. > > >>>> > > >>>>> -Daniel > > >>>>> > > >>>>>> Christian. > > >>>>>> > > >>>>>>> Cc: Thomas Hellstrom <thellstrom@vmware.com> > > >>>>>>> Cc: Brian Paul <brianp@vmware.com> > > >>>>>>> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> > > >>>>>>> Cc: Christian Koenig <christian.koenig@amd.com> > > >>>>>>> Cc: Huang Rui <ray.huang@amd.com> > > >>>>>>> --- > > >>>>>>> drivers/gpu/drm/ttm/ttm_tt.c | 12 ------------ > > >>>>>>> 1 file changed, 12 deletions(-) > > >>>>>>> > > >>>>>>> diff --git a/drivers/gpu/drm/ttm/ttm_tt.c b/drivers/gpu/drm/ttm/ttm_tt.c > > >>>>>>> index 8861a74ac335..438ea43fd8c1 100644 > > >>>>>>> --- a/drivers/gpu/drm/ttm/ttm_tt.c > > >>>>>>> +++ b/drivers/gpu/drm/ttm/ttm_tt.c > > >>>>>>> @@ -291,17 +291,6 @@ int ttm_tt_swapout(struct ttm_bo_device *bdev, struct ttm_tt *ttm) > > >>>>>>> return ret; > > >>>>>>> } > > >>>>>>> > > >>>>>>> -static void ttm_tt_add_mapping(struct ttm_bo_device *bdev, struct ttm_tt *ttm) > > >>>>>>> -{ > > >>>>>>> - pgoff_t i; > > >>>>>>> - > > >>>>>>> - if (ttm->page_flags & TTM_PAGE_FLAG_SG) > > >>>>>>> - return; > > >>>>>>> - > > >>>>>>> - for (i = 0; i < ttm->num_pages; ++i) > > >>>>>>> - ttm->pages[i]->mapping = bdev->dev_mapping; > > >>>>>>> -} > > >>>>>>> - > > >>>>>>> int ttm_tt_populate(struct ttm_bo_device *bdev, > > >>>>>>> struct ttm_tt *ttm, struct ttm_operation_ctx *ctx) > > >>>>>>> { > > >>>>>>> @@ -320,7 +309,6 @@ int ttm_tt_populate(struct ttm_bo_device *bdev, > > >>>>>>> if (ret) > > >>>>>>> return ret; > > >>>>>>> > > >>>>>>> - ttm_tt_add_mapping(bdev, ttm); > > >>>>>>> ttm->page_flags |= TTM_PAGE_FLAG_PRIV_POPULATED; > > >>>>>>> if (unlikely(ttm->page_flags & TTM_PAGE_FLAG_SWAPPED)) { > > >>>>>>> ret = ttm_tt_swapin(ttm); > > > > > -- > Daniel Vetter > Software Engineer, Intel Corporation > http://blog.ffwll.ch -- Daniel Vetter Software Engineer, Intel Corporation http://blog.ffwll.ch _______________________________________________ dri-devel mailing list dri-devel@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/dri-devel ^ permalink raw reply [flat|nested] 27+ messages in thread
* Re: [PATCH] drm/ttm: don't set page->mapping 2020-11-05 14:38 ` Daniel Vetter @ 2020-11-05 15:15 ` Christian König 2020-11-05 16:37 ` Daniel Vetter 0 siblings, 1 reply; 27+ messages in thread From: Christian König @ 2020-11-05 15:15 UTC (permalink / raw) To: Daniel Vetter; +Cc: Daniel Vetter, Huang Rui, Brian Paul, DRI Development Am 05.11.20 um 15:38 schrieb Daniel Vetter: > On Thu, Nov 5, 2020 at 3:31 PM Daniel Vetter <daniel@ffwll.ch> wrote: >> On Thu, Nov 5, 2020 at 2:22 PM Christian König <christian.koenig@amd.com> wrote: >>> Am 05.11.20 um 14:20 schrieb Daniel Vetter: >>>> On Thu, Nov 05, 2020 at 01:56:22PM +0100, Christian König wrote: >>>>> Am 05.11.20 um 13:50 schrieb Daniel Vetter: >>>>>> On Thu, Nov 05, 2020 at 01:29:50PM +0100, Christian König wrote: >>>>>>> Am 05.11.20 um 10:11 schrieb Daniel Vetter: >>>>>>>> On Thu, Nov 5, 2020 at 9:00 AM Christian König <christian.koenig@amd.com> wrote: >>>>>>>>> Am 04.11.20 um 17:50 schrieb Daniel Vetter: >>>>>>>>>> Random observation while trying to review Christian's patch series to >>>>>>>>>> stop looking at struct page for dma-buf imports. >>>>>>>>>> >>>>>>>>>> This was originally added in >>>>>>>>>> >>>>>>>>>> commit 58aa6622d32af7d2c08d45085f44c54554a16ed7 >>>>>>>>>> Author: Thomas Hellstrom <thellstrom@vmware.com> >>>>>>>>>> Date: Fri Jan 3 11:47:23 2014 +0100 >>>>>>>>>> >>>>>>>>>> drm/ttm: Correctly set page mapping and -index members >>>>>>>>>> >>>>>>>>>> Needed for some vm operations; most notably unmap_mapping_range() with >>>>>>>>>> even_cows = 0. >>>>>>>>>> >>>>>>>>>> Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com> >>>>>>>>>> Reviewed-by: Brian Paul <brianp@vmware.com> >>>>>>>>>> >>>>>>>>>> but we do not have a single caller of unmap_mapping_range with >>>>>>>>>> even_cows == 0. And all the gem drivers don't do this, so another >>>>>>>>>> small thing we could standardize between drm and ttm drivers. >>>>>>>>>> >>>>>>>>>> Plus I don't really see a need for unamp_mapping_range where we don't >>>>>>>>>> want to indiscriminately shoot down all ptes. >>>>>>>>> NAK, we use this to determine if a pages belongs to the driver or not in >>>>>>>>> amdgpu for example. >>>>>>>>> >>>>>>>>> Mostly used for debugging, but I would really like to keep that. >>>>>>>> Can you pls point me at that code? A quick grep hasn't really found much at all. >>>>>>> See amdgpu_iomem_read() for an example: >>>>>> Why do you reject this? >>>>> When IOMMU is disabled or uses an 1 to 1 mapping we would otherwise give the >>>>> same access as /dev/mem to system memory and that is forbidden. But as I >>>>> noted this is just for the debugfs file. >>>> Ah, there's a config option for that. Plus it's debugfs, anything goes in >>>> debugfs, but if you're worried about that hole we should just disable the >>>> entire debugfs file for CONFIG_STRICT_DEVMEM. I can perhaps throw that on >>>> top, that follow_pfn patch series I'm baking is all about this kind of >>>> fun. >>> And exactly that would get a NAK from us. >>> >>> We have specially created that debugfs file as an alternative when >>> CONFIG_STRICT_DEVMEM is set. >> Uh that doesn't work if you work around core restrictions with your >> own debugfs paths. That's why we have the restriction to check the mapping of the pages. This way we only expose the memory which was allocated by our driver and don't allow any uncontrolled access to the whole system memory. We have something similar for radeon as well, but there we have a global GART table which we can use for validating stuff. >> Maybe you can do fun like this in your dkms, but >> not in upstream. Like if this was specifically created to work around >> CONFIG_STRICT_DEVMEM (and it sounds like that) then I think this >> should never have landed in upstream to begin with. > I'm also kinda confused that there's distros with CONFIG_STRICT_DEVMEM > which allow debugfs. debugfs is a pretty bad root hole all around, or > at least that's been the assumption all the time. Yeah, completely agree :) But that's not my problem. Christian. > -Daniel > >>>>> When I tried a few years ago to not set the page->mapping I immediately ran >>>>> into issues with our eviction test. So I think that this is used elsewhere >>>>> as well. >>>> That's the kind of interaction I'm worried about here tbh. If this does >>>> some kind of shrinking of some sorts, I think a real shrinker should take >>>> over. >>>> >>>> An improved grep shows nothing else, so the only the above is the only >>>> thing I can think of. What kind of eviction test goes boom if you clear >>>> ->mapping here? I'd be happy to type up the clever trick for the debugfs >>>> files. >>>> -Daniel >>>> >>>>> Regards, >>>>> Christian. >>>>> >>>>>> If this is to avoid issues with userptr, then I think there's a simple >>>>>> trick: >>>>>> - grab page reference >>>>>> - recheck that the iova still points at the same address >>>>>> - do read/write, safe in the knowledge that this page cannot be reused for >>>>>> anything else >>>>>> - drop page reference >>>>>> >>>>>> Of course this can still race against iova updates, but that seems to be a >>>>>> fundamental part of your debug interface here. >>>>>> >>>>>> Or am I missing something? >>>>>> >>>>>> Just pondering this more since setting the page->mapping pointer for just >>>>>> this seems somewhat wild abuse of ->mapping semantics :-) >>>>>> -Daniel >>>>>> >>>>>>>> if (p->mapping != adev->mman.bdev.dev_mapping) >>>>>>>> return -EPERM; >>>>>>> Christian. >>>>>>> >>>>>>>> -Daniel >>>>>>>> >>>>>>>>> Christian. >>>>>>>>> >>>>>>>>>> Cc: Thomas Hellstrom <thellstrom@vmware.com> >>>>>>>>>> Cc: Brian Paul <brianp@vmware.com> >>>>>>>>>> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> >>>>>>>>>> Cc: Christian Koenig <christian.koenig@amd.com> >>>>>>>>>> Cc: Huang Rui <ray.huang@amd.com> >>>>>>>>>> --- >>>>>>>>>> drivers/gpu/drm/ttm/ttm_tt.c | 12 ------------ >>>>>>>>>> 1 file changed, 12 deletions(-) >>>>>>>>>> >>>>>>>>>> diff --git a/drivers/gpu/drm/ttm/ttm_tt.c b/drivers/gpu/drm/ttm/ttm_tt.c >>>>>>>>>> index 8861a74ac335..438ea43fd8c1 100644 >>>>>>>>>> --- a/drivers/gpu/drm/ttm/ttm_tt.c >>>>>>>>>> +++ b/drivers/gpu/drm/ttm/ttm_tt.c >>>>>>>>>> @@ -291,17 +291,6 @@ int ttm_tt_swapout(struct ttm_bo_device *bdev, struct ttm_tt *ttm) >>>>>>>>>> return ret; >>>>>>>>>> } >>>>>>>>>> >>>>>>>>>> -static void ttm_tt_add_mapping(struct ttm_bo_device *bdev, struct ttm_tt *ttm) >>>>>>>>>> -{ >>>>>>>>>> - pgoff_t i; >>>>>>>>>> - >>>>>>>>>> - if (ttm->page_flags & TTM_PAGE_FLAG_SG) >>>>>>>>>> - return; >>>>>>>>>> - >>>>>>>>>> - for (i = 0; i < ttm->num_pages; ++i) >>>>>>>>>> - ttm->pages[i]->mapping = bdev->dev_mapping; >>>>>>>>>> -} >>>>>>>>>> - >>>>>>>>>> int ttm_tt_populate(struct ttm_bo_device *bdev, >>>>>>>>>> struct ttm_tt *ttm, struct ttm_operation_ctx *ctx) >>>>>>>>>> { >>>>>>>>>> @@ -320,7 +309,6 @@ int ttm_tt_populate(struct ttm_bo_device *bdev, >>>>>>>>>> if (ret) >>>>>>>>>> return ret; >>>>>>>>>> >>>>>>>>>> - ttm_tt_add_mapping(bdev, ttm); >>>>>>>>>> ttm->page_flags |= TTM_PAGE_FLAG_PRIV_POPULATED; >>>>>>>>>> if (unlikely(ttm->page_flags & TTM_PAGE_FLAG_SWAPPED)) { >>>>>>>>>> ret = ttm_tt_swapin(ttm); >> >> -- >> Daniel Vetter >> Software Engineer, Intel Corporation >> https://nam11.safelinks.protection.outlook.com/?url=http%3A%2F%2Fblog.ffwll.ch%2F&data=04%7C01%7Cchristian.koenig%40amd.com%7C619e6a6113674691eb9708d8819874f4%7C3dd8961fe4884e608e11a82d994e183d%7C0%7C0%7C637401839082694450%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=Uo7UXS7y%2BU%2FHfnBenx2vQXuyyB%2FCuOULLOp1uL0eg4I%3D&reserved=0 > > _______________________________________________ dri-devel mailing list dri-devel@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/dri-devel ^ permalink raw reply [flat|nested] 27+ messages in thread
* Re: [PATCH] drm/ttm: don't set page->mapping 2020-11-05 15:15 ` Christian König @ 2020-11-05 16:37 ` Daniel Vetter 2020-11-06 8:30 ` Christian König 0 siblings, 1 reply; 27+ messages in thread From: Daniel Vetter @ 2020-11-05 16:37 UTC (permalink / raw) To: Christian König Cc: Daniel Vetter, Huang Rui, Brian Paul, DRI Development On Thu, Nov 5, 2020 at 4:15 PM Christian König <christian.koenig@amd.com> wrote: > > Am 05.11.20 um 15:38 schrieb Daniel Vetter: > > On Thu, Nov 5, 2020 at 3:31 PM Daniel Vetter <daniel@ffwll.ch> wrote: > >> On Thu, Nov 5, 2020 at 2:22 PM Christian König <christian.koenig@amd.com> wrote: > >>> Am 05.11.20 um 14:20 schrieb Daniel Vetter: > >>>> On Thu, Nov 05, 2020 at 01:56:22PM +0100, Christian König wrote: > >>>>> Am 05.11.20 um 13:50 schrieb Daniel Vetter: > >>>>>> On Thu, Nov 05, 2020 at 01:29:50PM +0100, Christian König wrote: > >>>>>>> Am 05.11.20 um 10:11 schrieb Daniel Vetter: > >>>>>>>> On Thu, Nov 5, 2020 at 9:00 AM Christian König <christian.koenig@amd.com> wrote: > >>>>>>>>> Am 04.11.20 um 17:50 schrieb Daniel Vetter: > >>>>>>>>>> Random observation while trying to review Christian's patch series to > >>>>>>>>>> stop looking at struct page for dma-buf imports. > >>>>>>>>>> > >>>>>>>>>> This was originally added in > >>>>>>>>>> > >>>>>>>>>> commit 58aa6622d32af7d2c08d45085f44c54554a16ed7 > >>>>>>>>>> Author: Thomas Hellstrom <thellstrom@vmware.com> > >>>>>>>>>> Date: Fri Jan 3 11:47:23 2014 +0100 > >>>>>>>>>> > >>>>>>>>>> drm/ttm: Correctly set page mapping and -index members > >>>>>>>>>> > >>>>>>>>>> Needed for some vm operations; most notably unmap_mapping_range() with > >>>>>>>>>> even_cows = 0. > >>>>>>>>>> > >>>>>>>>>> Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com> > >>>>>>>>>> Reviewed-by: Brian Paul <brianp@vmware.com> > >>>>>>>>>> > >>>>>>>>>> but we do not have a single caller of unmap_mapping_range with > >>>>>>>>>> even_cows == 0. And all the gem drivers don't do this, so another > >>>>>>>>>> small thing we could standardize between drm and ttm drivers. > >>>>>>>>>> > >>>>>>>>>> Plus I don't really see a need for unamp_mapping_range where we don't > >>>>>>>>>> want to indiscriminately shoot down all ptes. > >>>>>>>>> NAK, we use this to determine if a pages belongs to the driver or not in > >>>>>>>>> amdgpu for example. > >>>>>>>>> > >>>>>>>>> Mostly used for debugging, but I would really like to keep that. > >>>>>>>> Can you pls point me at that code? A quick grep hasn't really found much at all. > >>>>>>> See amdgpu_iomem_read() for an example: > >>>>>> Why do you reject this? > >>>>> When IOMMU is disabled or uses an 1 to 1 mapping we would otherwise give the > >>>>> same access as /dev/mem to system memory and that is forbidden. But as I > >>>>> noted this is just for the debugfs file. > >>>> Ah, there's a config option for that. Plus it's debugfs, anything goes in > >>>> debugfs, but if you're worried about that hole we should just disable the > >>>> entire debugfs file for CONFIG_STRICT_DEVMEM. I can perhaps throw that on > >>>> top, that follow_pfn patch series I'm baking is all about this kind of > >>>> fun. > >>> And exactly that would get a NAK from us. > >>> > >>> We have specially created that debugfs file as an alternative when > >>> CONFIG_STRICT_DEVMEM is set. > >> Uh that doesn't work if you work around core restrictions with your > >> own debugfs paths. > > That's why we have the restriction to check the mapping of the pages. > > This way we only expose the memory which was allocated by our driver and > don't allow any uncontrolled access to the whole system memory. > > We have something similar for radeon as well, but there we have a global > GART table which we can use for validating stuff. The check doesn't take any locks over the check and copy*user, I don't think it's protecting anything really against somewhat adverse userspace. I mean fundamentally locking down stuff like STRICT_DEVMEM or all the others makes debugging harder, that's kinda the expected tradeoff. > >> Maybe you can do fun like this in your dkms, but > >> not in upstream. Like if this was specifically created to work around > >> CONFIG_STRICT_DEVMEM (and it sounds like that) then I think this > >> should never have landed in upstream to begin with. > > I'm also kinda confused that there's distros with CONFIG_STRICT_DEVMEM > > which allow debugfs. debugfs is a pretty bad root hole all around, or > > at least that's been the assumption all the time. > > Yeah, completely agree :) But that's not my problem. I guess I'll do another rfc series and poke a pile of people ... seems to be a habit I'm developing :-) -Daniel > > Christian. > > > -Daniel > > > >>>>> When I tried a few years ago to not set the page->mapping I immediately ran > >>>>> into issues with our eviction test. So I think that this is used elsewhere > >>>>> as well. > >>>> That's the kind of interaction I'm worried about here tbh. If this does > >>>> some kind of shrinking of some sorts, I think a real shrinker should take > >>>> over. > >>>> > >>>> An improved grep shows nothing else, so the only the above is the only > >>>> thing I can think of. What kind of eviction test goes boom if you clear > >>>> ->mapping here? I'd be happy to type up the clever trick for the debugfs > >>>> files. > >>>> -Daniel > >>>> > >>>>> Regards, > >>>>> Christian. > >>>>> > >>>>>> If this is to avoid issues with userptr, then I think there's a simple > >>>>>> trick: > >>>>>> - grab page reference > >>>>>> - recheck that the iova still points at the same address > >>>>>> - do read/write, safe in the knowledge that this page cannot be reused for > >>>>>> anything else > >>>>>> - drop page reference > >>>>>> > >>>>>> Of course this can still race against iova updates, but that seems to be a > >>>>>> fundamental part of your debug interface here. > >>>>>> > >>>>>> Or am I missing something? > >>>>>> > >>>>>> Just pondering this more since setting the page->mapping pointer for just > >>>>>> this seems somewhat wild abuse of ->mapping semantics :-) > >>>>>> -Daniel > >>>>>> > >>>>>>>> if (p->mapping != adev->mman.bdev.dev_mapping) > >>>>>>>> return -EPERM; > >>>>>>> Christian. > >>>>>>> > >>>>>>>> -Daniel > >>>>>>>> > >>>>>>>>> Christian. > >>>>>>>>> > >>>>>>>>>> Cc: Thomas Hellstrom <thellstrom@vmware.com> > >>>>>>>>>> Cc: Brian Paul <brianp@vmware.com> > >>>>>>>>>> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> > >>>>>>>>>> Cc: Christian Koenig <christian.koenig@amd.com> > >>>>>>>>>> Cc: Huang Rui <ray.huang@amd.com> > >>>>>>>>>> --- > >>>>>>>>>> drivers/gpu/drm/ttm/ttm_tt.c | 12 ------------ > >>>>>>>>>> 1 file changed, 12 deletions(-) > >>>>>>>>>> > >>>>>>>>>> diff --git a/drivers/gpu/drm/ttm/ttm_tt.c b/drivers/gpu/drm/ttm/ttm_tt.c > >>>>>>>>>> index 8861a74ac335..438ea43fd8c1 100644 > >>>>>>>>>> --- a/drivers/gpu/drm/ttm/ttm_tt.c > >>>>>>>>>> +++ b/drivers/gpu/drm/ttm/ttm_tt.c > >>>>>>>>>> @@ -291,17 +291,6 @@ int ttm_tt_swapout(struct ttm_bo_device *bdev, struct ttm_tt *ttm) > >>>>>>>>>> return ret; > >>>>>>>>>> } > >>>>>>>>>> > >>>>>>>>>> -static void ttm_tt_add_mapping(struct ttm_bo_device *bdev, struct ttm_tt *ttm) > >>>>>>>>>> -{ > >>>>>>>>>> - pgoff_t i; > >>>>>>>>>> - > >>>>>>>>>> - if (ttm->page_flags & TTM_PAGE_FLAG_SG) > >>>>>>>>>> - return; > >>>>>>>>>> - > >>>>>>>>>> - for (i = 0; i < ttm->num_pages; ++i) > >>>>>>>>>> - ttm->pages[i]->mapping = bdev->dev_mapping; > >>>>>>>>>> -} > >>>>>>>>>> - > >>>>>>>>>> int ttm_tt_populate(struct ttm_bo_device *bdev, > >>>>>>>>>> struct ttm_tt *ttm, struct ttm_operation_ctx *ctx) > >>>>>>>>>> { > >>>>>>>>>> @@ -320,7 +309,6 @@ int ttm_tt_populate(struct ttm_bo_device *bdev, > >>>>>>>>>> if (ret) > >>>>>>>>>> return ret; > >>>>>>>>>> > >>>>>>>>>> - ttm_tt_add_mapping(bdev, ttm); > >>>>>>>>>> ttm->page_flags |= TTM_PAGE_FLAG_PRIV_POPULATED; > >>>>>>>>>> if (unlikely(ttm->page_flags & TTM_PAGE_FLAG_SWAPPED)) { > >>>>>>>>>> ret = ttm_tt_swapin(ttm); > >> > >> -- > >> Daniel Vetter > >> Software Engineer, Intel Corporation > >> https://nam11.safelinks.protection.outlook.com/?url=http%3A%2F%2Fblog.ffwll.ch%2F&data=04%7C01%7Cchristian.koenig%40amd.com%7C619e6a6113674691eb9708d8819874f4%7C3dd8961fe4884e608e11a82d994e183d%7C0%7C0%7C637401839082694450%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=Uo7UXS7y%2BU%2FHfnBenx2vQXuyyB%2FCuOULLOp1uL0eg4I%3D&reserved=0 > > > > > -- Daniel Vetter Software Engineer, Intel Corporation http://blog.ffwll.ch _______________________________________________ dri-devel mailing list dri-devel@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/dri-devel ^ permalink raw reply [flat|nested] 27+ messages in thread
* Re: [PATCH] drm/ttm: don't set page->mapping 2020-11-05 16:37 ` Daniel Vetter @ 2020-11-06 8:30 ` Christian König 0 siblings, 0 replies; 27+ messages in thread From: Christian König @ 2020-11-06 8:30 UTC (permalink / raw) To: Daniel Vetter; +Cc: Daniel Vetter, Huang Rui, Brian Paul, DRI Development Am 05.11.20 um 17:37 schrieb Daniel Vetter: > [SNIP] >>>>>>>>>>> NAK, we use this to determine if a pages belongs to the driver or not in >>>>>>>>>>> amdgpu for example. >>>>>>>>>>> >>>>>>>>>>> Mostly used for debugging, but I would really like to keep that. >>>>>>>>>> Can you pls point me at that code? A quick grep hasn't really found much at all. >>>>>>>>> See amdgpu_iomem_read() for an example: >>>>>>>> Why do you reject this? >>>>>>> When IOMMU is disabled or uses an 1 to 1 mapping we would otherwise give the >>>>>>> same access as /dev/mem to system memory and that is forbidden. But as I >>>>>>> noted this is just for the debugfs file. >>>>>> Ah, there's a config option for that. Plus it's debugfs, anything goes in >>>>>> debugfs, but if you're worried about that hole we should just disable the >>>>>> entire debugfs file for CONFIG_STRICT_DEVMEM. I can perhaps throw that on >>>>>> top, that follow_pfn patch series I'm baking is all about this kind of >>>>>> fun. >>>>> And exactly that would get a NAK from us. >>>>> >>>>> We have specially created that debugfs file as an alternative when >>>>> CONFIG_STRICT_DEVMEM is set. >>>> Uh that doesn't work if you work around core restrictions with your >>>> own debugfs paths. >> That's why we have the restriction to check the mapping of the pages. >> >> This way we only expose the memory which was allocated by our driver and >> don't allow any uncontrolled access to the whole system memory. >> >> We have something similar for radeon as well, but there we have a global >> GART table which we can use for validating stuff. > The check doesn't take any locks over the check and copy*user, I don't > think it's protecting anything really against somewhat adverse > userspace. You mean that it's racing with freeing the pages? Yes that is an obvious problem, but we already had the same issue for radeon as well. On the other hand we could easily prevent that with a lock. > I mean fundamentally locking down stuff like STRICT_DEVMEM or all the > others makes debugging harder, that's kinda the expected tradeoff. Well we just wanted to have the same debugging functionality we had for radeon. As you said debugfs is a rabbit hole regarding attack vectors anyway. If you are root you can just go to /sys and start reprogramming the hardware to do what you want to do. Regards, Christian. > >>>> Maybe you can do fun like this in your dkms, but >>>> not in upstream. Like if this was specifically created to work around >>>> CONFIG_STRICT_DEVMEM (and it sounds like that) then I think this >>>> should never have landed in upstream to begin with. >>> I'm also kinda confused that there's distros with CONFIG_STRICT_DEVMEM >>> which allow debugfs. debugfs is a pretty bad root hole all around, or >>> at least that's been the assumption all the time. >> Yeah, completely agree :) But that's not my problem. > I guess I'll do another rfc series and poke a pile of people ... seems > to be a habit I'm developing :-) > -Daniel > _______________________________________________ dri-devel mailing list dri-devel@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/dri-devel ^ permalink raw reply [flat|nested] 27+ messages in thread
end of thread, other threads:[~2020-11-26 8:24 UTC | newest] Thread overview: 27+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2020-11-25 16:25 [PATCH v4 0/3] mmu_notifier vs fs_reclaim lockdep annotations Daniel Vetter 2020-11-25 16:25 ` [PATCH v4 1/3] mm: Track mmu notifiers in fs_reclaim_acquire/release Daniel Vetter 2020-11-25 16:25 ` [PATCH v4 2/3] mm: Extract might_alloc() debug check Daniel Vetter 2020-11-25 16:25 ` [PATCH v4 3/3] locking/selftests: Add testcases for fs_reclaim Daniel Vetter 2020-11-25 16:25 ` [PATCH] drm/ttm: don't set page->mapping Daniel Vetter 2020-11-25 16:28 ` Daniel Vetter 2020-11-25 18:06 ` Jason Gunthorpe 2020-11-25 18:16 ` Daniel Stone [not found] ` <20201125181129.GA1858@infradead.org> 2020-11-25 23:57 ` Jason Gunthorpe -- strict thread matches above, loose matches on Subject: below -- 2020-11-20 9:54 [PATCH 0/3] mmu_notifier fs fs_reclaim lockdep annotations Daniel Vetter 2020-11-20 9:54 ` [PATCH] drm/ttm: don't set page->mapping Daniel Vetter 2020-11-20 10:04 ` Christian König 2020-11-20 10:05 ` Daniel Vetter 2020-11-20 10:08 ` Christian König 2020-11-20 15:01 ` Daniel Vetter 2020-11-04 16:50 Daniel Vetter 2020-11-05 7:59 ` Christian König 2020-11-05 9:11 ` Daniel Vetter 2020-11-05 12:29 ` Christian König 2020-11-05 12:50 ` Daniel Vetter 2020-11-05 12:56 ` Christian König 2020-11-05 13:20 ` Daniel Vetter 2020-11-05 13:22 ` Christian König 2020-11-05 14:31 ` Daniel Vetter 2020-11-05 14:38 ` Daniel Vetter 2020-11-05 15:15 ` Christian König 2020-11-05 16:37 ` Daniel Vetter 2020-11-06 8:30 ` Christian König
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).