[Intel-gfx] [PATCH v2] drm/i915/ttm: fixup the mock_bo

* [Intel-gfx] [PATCH v2] drm/i915/ttm: fixup the mock_bo
@ 2022-02-21 12:11 ` Matthew Auld
  0 siblings, 0 replies; 7+ messages in thread
From: Matthew Auld @ 2022-02-21 12:11 UTC (permalink / raw)
  To: intel-gfx; +Cc: Thomas Hellström, Christian König, dri-devel

When running the mock selftests we currently blow up with:

<6> [299.836278] i915: Running i915_gem_huge_page_mock_selftests/igt_mock_memory_region_huge_pages
<1> [299.836356] BUG: kernel NULL pointer dereference, address: 00000000000000c8
<1> [299.836361] #PF: supervisor read access in kernel mode
<1> [299.836364] #PF: error_code(0x0000) - not-present page
<6> [299.836367] PGD 0 P4D 0
<4> [299.836369] Oops: 0000 [#1] PREEMPT SMP NOPTI
<4> [299.836372] CPU: 1 PID: 1429 Comm: i915_selftest Tainted: G     U            5.17.0-rc4-CI-CI_DRM_11227+ #1
<4> [299.836376] Hardware name: Intel(R) Client Systems NUC11TNHi5/NUC11TNBi5, BIOS TNTGL357.0042.2020.1221.1743 12/21/2020
<4> [299.836380] RIP: 0010:ttm_resource_init+0x57/0x90 [ttm]
<4> [299.836392] RSP: 0018:ffffc90001e4f680 EFLAGS: 00010203
<4> [299.836395] RAX: 0000000000000000 RBX: ffffc90001e4f708 RCX: 0000000000000000
<4> [299.836398] RDX: ffff888116172528 RSI: ffffc90001e4f6f8 RDI: 0000000000000000
<4> [299.836401] RBP: ffffc90001e4f6f8 R08: 00000000000001b0 R09: ffff888116172528
<4> [299.836403] R10: 0000000000000001 R11: 00000000a4cb2e51 R12: ffffc90001e4fa90
<4> [299.836406] R13: ffff888116172528 R14: ffff888130d7f4b0 R15: ffff888130d7f400
<4> [299.836409] FS:  00007ff241684500(0000) GS:ffff88849fe80000(0000) knlGS:0000000000000000
<4> [299.836412] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
<4> [299.836416] CR2: 00000000000000c8 CR3: 0000000107b80001 CR4: 0000000000770ee0
<4> [299.836418] PKRU: 55555554
<4> [299.836420] Call Trace:
<4> [299.836422]  <TASK>
<4> [299.836423]  i915_ttm_buddy_man_alloc+0x68/0x240 [i915]

ttm_resource_init() now needs to access the bo->bdev, and also wants to
store the bo reference. Try to keep both working. The mock_bo is a hack
so we can interface directly with the ttm managers alloc() and free() hooks for
our mock testing, without invoking other TTM features like eviction,
moves, etc.

v2: make sure we only touch res->bo if the alloc() returns successfully

Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/5123
Fixes: 0e05fc49c358 ("drm/ttm: add common accounting to the resource mgr v3")
Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Cc: Christian König <christian.koenig@amd.com>
Cc: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Acked-by: Christian König <christian.koenig@amd.com>
---
 drivers/gpu/drm/i915/intel_region_ttm.c | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/drivers/gpu/drm/i915/intel_region_ttm.c b/drivers/gpu/drm/i915/intel_region_ttm.c
index f2b888c16958..7dea07c579aa 100644
--- a/drivers/gpu/drm/i915/intel_region_ttm.c
+++ b/drivers/gpu/drm/i915/intel_region_ttm.c
@@ -200,11 +200,14 @@ intel_region_ttm_resource_alloc(struct intel_memory_region *mem,
 	int ret;
 
 	mock_bo.base.size = size;
+	mock_bo.bdev = &mem->i915->bdev;
 	place.flags = flags;
 
 	ret = man->func->alloc(man, &mock_bo, &place, &res);
 	if (ret == -ENOSPC)
 		ret = -ENXIO;
+	if (!ret)
+		res->bo = NULL; /* Rather blow up, then some uaf */
 	return ret ? ERR_PTR(ret) : res;
 }
 
@@ -219,6 +222,11 @@ void intel_region_ttm_resource_free(struct intel_memory_region *mem,
 				    struct ttm_resource *res)
 {
 	struct ttm_resource_manager *man = mem->region_private;
+	struct ttm_buffer_object mock_bo = {};
+
+	mock_bo.base.size = res->num_pages << PAGE_SHIFT;
+	mock_bo.bdev = &mem->i915->bdev;
+	res->bo = &mock_bo;
 
 	man->func->free(man, res);
 }
-- 
2.34.1


^ permalink raw reply related	[flat|nested] 7+ messages in thread