From: Peter Senna Tschudin <peter.senna@linux.intel.com>
To: "Zbigniew Kempczyński" <zbigniew.kempczynski@intel.com>,
igt-dev@lists.freedesktop.org
Cc: Andi Shyti <andi.shyti@linux.intel.com>
Subject: Re: [PATCH i-g-t] lib|tests: Don't dereference engine outside for_each_ctx_engine macro
Date: Wed, 14 Aug 2024 09:10:40 +0200 [thread overview]
Message-ID: <a1e22517-77be-4c9d-8115-bc7ce201867e@linux.intel.com> (raw)
In-Reply-To: <20240813124536.66551-1-zbigniew.kempczynski@intel.com>
Please ignore my previous emails on this topic. I underestimated the effort and the potential consequences of the changes I want to make. I have reviewed the code and compile tested it.
On 13.08.2024 14:45, Zbigniew Kempczyński wrote:
> Engine used inside for_each_ctx_engine() macro is valid only inside its
> block as it is accessing local stack memory. Using outside the block works
> by an accident if luckily nothing will overwrite this stack before use.
> Fix found risky usages by using engine only inside the block or do its
> copy.
>
> Signed-off-by: Zbigniew Kempczyński <zbigniew.kempczynski@intel.com>
> Cc: Andi Shyti <andi.shyti@linux.intel.com>
Reviewed-by: Peter Senna Tschudin <peter.senna@linux.intel.com>
> ---
> lib/igt_fb.c | 8 ++++----
> tests/intel/gem_lmem_swapping.c | 14 ++++++++++----
> 2 files changed, 14 insertions(+), 8 deletions(-)
>
> diff --git a/lib/igt_fb.c b/lib/igt_fb.c
> index ab162a5b74..addff2cfba 100644
> --- a/lib/igt_fb.c
> +++ b/lib/igt_fb.c
> @@ -3050,13 +3050,13 @@ static void blitcopy(const struct igt_fb *dst_fb,
> dst_fb->size);
> } else if (ahnd && block_copy_ok(src_fb) && block_copy_ok(dst_fb)) {
> for_each_ctx_engine(src_fb->fd, ictx, e) {
> - if (gem_engine_can_block_copy(src_fb->fd, e))
> + if (gem_engine_can_block_copy(src_fb->fd, e)) {
> + do_block_copy(src_fb, dst_fb, mem_region, i, ahnd,
> + bb, bb_size, ictx, e);
> break;
> + }
> }
> igt_assert_f(e, "No block copy capable engine found!\n");
> -
> - do_block_copy(src_fb, dst_fb, mem_region, i, ahnd,
> - bb, bb_size, ictx, e);
> } else {
> igt_blitter_src_copy(dst_fb->fd,
> ahnd, ctx, NULL,
> diff --git a/tests/intel/gem_lmem_swapping.c b/tests/intel/gem_lmem_swapping.c
> index b8f24742b8..b125261519 100644
> --- a/tests/intel/gem_lmem_swapping.c
> +++ b/tests/intel/gem_lmem_swapping.c
> @@ -189,6 +189,7 @@ init_object_ccs(int i915, struct object *obj, struct blt_copy_object *tmp,
> {
> struct blt_block_copy_data_ext ext = {}, *pext = &ext;
> const struct intel_execution_engine2 *e;
> + struct intel_execution_engine2 ec;
> struct blt_copy_data blt = {};
> struct blt_copy_batch *cmd;
> uint64_t size = 4096;
> @@ -196,8 +197,10 @@ init_object_ccs(int i915, struct object *obj, struct blt_copy_object *tmp,
>
> obj->seed = seed;
> for_each_ctx_engine(i915, ctx, e) {
> - if (gem_engine_can_block_copy(i915, e))
> + if (gem_engine_can_block_copy(i915, e)) {
> + ec = *e;
> break;
> + }
> }
> igt_assert_f(e, "Ctx don't have blt engine\n");
>
> @@ -224,7 +227,7 @@ init_object_ccs(int i915, struct object *obj, struct blt_copy_object *tmp,
> blt_set_object_ext(&ext.dst, 0, obj->blt_obj->x2, obj->blt_obj->y2,
> SURFACE_TYPE_2D);
>
> - blt_block_copy(i915, ctx, e, ahnd, &blt, pext);
> + blt_block_copy(i915, ctx, &ec, ahnd, &blt, pext);
> free(cmd);
> }
>
> @@ -255,14 +258,17 @@ verify_object_ccs(int i915, const struct object *obj,
> {
> struct blt_block_copy_data_ext ext = {}, *pext = &ext;
> const struct intel_execution_engine2 *e;
> + struct intel_execution_engine2 ec;
> struct blt_copy_data blt = {};
> struct blt_copy_batch *cmd;
> uint64_t size = 4096;
> unsigned long j, val, *buf;
>
> for_each_ctx_engine(i915, ctx, e) {
> - if (gem_engine_can_block_copy(i915, e))
> + if (gem_engine_can_block_copy(i915, e)) {
> + ec = *e;
> break;
> + }
> }
> igt_assert_f(e, "Ctx don't have blt engine\n");
>
> @@ -284,7 +290,7 @@ verify_object_ccs(int i915, const struct object *obj,
> blt_set_object_ext(&ext.src, 0, obj->blt_obj->x2, obj->blt_obj->y2,
> SURFACE_TYPE_2D);
> blt_set_object_ext(&ext.dst, 0, tmp->x2, tmp->y2, SURFACE_TYPE_2D);
> - blt_block_copy(i915, ctx, e, ahnd, &blt, pext);
> + blt_block_copy(i915, ctx, &ec, ahnd, &blt, pext);
>
> buf = gem_mmap__device_coherent(i915, tmp->handle, 0,
> obj->size, PROT_READ);
prev parent reply other threads:[~2024-08-14 7:10 UTC|newest]
Thread overview: 19+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-08-13 12:45 [PATCH i-g-t] lib|tests: Don't dereference engine outside for_each_ctx_engine macro Zbigniew Kempczyński
2024-08-13 13:00 ` ✗ GitLab.Pipeline: warning for " Patchwork
2024-08-13 13:13 ` ✗ CI.xeBAT: failure " Patchwork
2024-08-13 13:16 ` ✓ Fi.CI.BAT: success " Patchwork
2024-08-13 13:36 ` [PATCH i-g-t] " Peter Senna Tschudin
2024-08-13 13:56 ` Zbigniew Kempczyński
2024-08-13 14:12 ` Peter Senna Tschudin
2024-08-13 19:51 ` Andi Shyti
2024-08-14 7:11 ` Peter Senna Tschudin
2024-08-14 9:34 ` Andi Shyti
2024-08-13 14:55 ` ✗ CI.xeBAT: failure for lib|tests: Don't dereference engine outside for_each_ctx_engine macro (rev2) Patchwork
2024-08-13 15:06 ` ✓ Fi.CI.BAT: success " Patchwork
2024-08-13 16:18 ` ✗ CI.xeFULL: failure for lib|tests: Don't dereference engine outside for_each_ctx_engine macro Patchwork
2024-08-14 8:52 ` Zbigniew Kempczyński
2024-08-13 17:14 ` ✗ CI.xeFULL: failure for lib|tests: Don't dereference engine outside for_each_ctx_engine macro (rev2) Patchwork
2024-08-13 19:49 ` ✓ Fi.CI.IGT: success for lib|tests: Don't dereference engine outside for_each_ctx_engine macro Patchwork
2024-08-13 22:49 ` ✗ Fi.CI.IGT: failure for lib|tests: Don't dereference engine outside for_each_ctx_engine macro (rev2) Patchwork
2024-08-14 8:50 ` Zbigniew Kempczyński
2024-08-14 7:10 ` Peter Senna Tschudin [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=a1e22517-77be-4c9d-8115-bc7ce201867e@linux.intel.com \
--to=peter.senna@linux.intel.com \
--cc=andi.shyti@linux.intel.com \
--cc=igt-dev@lists.freedesktop.org \
--cc=zbigniew.kempczynski@intel.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox