[RESEND PATCH v2 0/3] Fix use-after-free bug in mtk_jpeg_dec_device_run and fix schedule error in mtk_jpegdec_worker
Hello,
This v2 series fixes the use-after-free bug in mtk_jpeg_dec_device_run. This patch fixes the security bug in chrome-os. It inclues reverting the incomplete fix before and make the right fix. Also,it fixes the error of timeout-worker-schedule in multiple-core devices.
1. Remove cancel worker in mtk_jpeg_remove for the worker is only registered in single-core device but we try to cacnel it in both single-core and multiple-core devices.
2. Fix use-after-free bug by delay the schedule_delayed_work only if mtk_jpeg_set_dec_dst runs successfully.
3. Delay the schedule_delayed_work in mtk_jpegdec_worker as it has same code logic in mtk_jpeg_dec_device_run.
version 2 changes
-put the patches into on series suggested by Dmitry
Zheng Wang (3): media: mtk-jpeg: Remove cancel worker in mtk_jpeg_remove to avoid the crash of multi-core JPEG devices media: mtk-jpeg: Fix use after free bug due to error path handling in mtk_jpeg_dec_device_run media: mtk-jpeg: Fix timeout schedule error in mtk_jpegdec_worker.
.../media/platform/mediatek/jpeg/mtk_jpeg_core.c | 13 ++++++------- 1 file changed, 6 insertions(+), 7 deletions(-)
This patch reverts commit c677d7ae8314 ("media: mtk-jpeg: Fix use after free bug due to uncanceled work"). The job_timeout_work is initialized only for the single-core JPEG device so it will cause the crash for multi-core JPEG devices.
Fix it by removing the cancel_delayed_work_sync function.
Fixes: c677d7ae8314 ("media: mtk-jpeg: Fix use after free bug due to uncanceled work") Signed-off-by: Zheng Wang zyytlz.wz@163.com Signed-off-by: Dmitry Osipenko dmitry.osipenko@collabora.com --- v2: - put the patches into a single series suggested by Dmitry --- drivers/media/platform/mediatek/jpeg/mtk_jpeg_core.c | 1 - 1 file changed, 1 deletion(-)
diff --git a/drivers/media/platform/mediatek/jpeg/mtk_jpeg_core.c b/drivers/media/platform/mediatek/jpeg/mtk_jpeg_core.c index 7194f88edc0f..60425c99a2b8 100644 --- a/drivers/media/platform/mediatek/jpeg/mtk_jpeg_core.c +++ b/drivers/media/platform/mediatek/jpeg/mtk_jpeg_core.c @@ -1403,7 +1403,6 @@ static void mtk_jpeg_remove(struct platform_device *pdev) { struct mtk_jpeg_dev *jpeg = platform_get_drvdata(pdev);
- cancel_delayed_work_sync(&jpeg->job_timeout_work); pm_runtime_disable(&pdev->dev); video_unregister_device(jpeg->vdev); v4l2_m2m_release(jpeg->m2m_dev);
Hi,
Thanks for your patch.
FYI: kernel test robot notices the stable kernel rule is not satisfied.
The check is based on https://www.kernel.org/doc/html/latest/process/stable-kernel-rules.html#opti...
Rule: add the tag "Cc: stable@vger.kernel.org" in the sign-off area to have the patch automatically included in the stable tree. Subject: [RESEND PATCH v2 1/3] media: mtk-jpeg: Remove cancel worker in mtk_jpeg_remove to avoid the crash of multi-core JPEG devices Link: https://lore.kernel.org/stable/20231106144811.868127-2-zyytlz.wz%40163.com
In mtk_jpeg_probe, &jpeg->job_timeout_work is bound with mtk_jpeg_job_timeout_work.
In mtk_jpeg_dec_device_run, if error happens in mtk_jpeg_set_dec_dst, it will finally start the worker while mark the job as finished by invoking v4l2_m2m_job_finish.
There are two methods to trigger the bug. If we remove the module, it which will call mtk_jpeg_remove to make cleanup. The possible sequence is as follows, which will cause a use-after-free bug.
CPU0 CPU1 mtk_jpeg_dec_... | start worker | |mtk_jpeg_job_timeout_work mtk_jpeg_remove | v4l2_m2m_release | kfree(m2m_dev); | | | v4l2_m2m_get_curr_priv | m2m_dev->curr_ctx //use
If we close the file descriptor, which will call mtk_jpeg_release, it will have a similar sequence.
Fix this bug by starting timeout worker only if started jpegdec worker successfully. Then v4l2_m2m_job_finish will only be called in either mtk_jpeg_job_timeout_work or mtk_jpeg_dec_device_run. Fixes: b2f0d2724ba4 ("[media] vcodec: mediatek: Add Mediatek JPEG Decoder Driver") Signed-off-by: Zheng Wang zyytlz.wz@163.com Signed-off-by: Dmitry Osipenko dmitry.osipenko@collabora.com Cc: stable@vger.kernel.org --- v2: - put the patches into a single series suggested by Dmitry --- drivers/media/platform/mediatek/jpeg/mtk_jpeg_core.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/drivers/media/platform/mediatek/jpeg/mtk_jpeg_core.c b/drivers/media/platform/mediatek/jpeg/mtk_jpeg_core.c index 60425c99a2b8..a39acde2724a 100644 --- a/drivers/media/platform/mediatek/jpeg/mtk_jpeg_core.c +++ b/drivers/media/platform/mediatek/jpeg/mtk_jpeg_core.c @@ -1021,13 +1021,13 @@ static void mtk_jpeg_dec_device_run(void *priv) if (ret < 0) goto dec_end;
- schedule_delayed_work(&jpeg->job_timeout_work, - msecs_to_jiffies(MTK_JPEG_HW_TIMEOUT_MSEC)); - mtk_jpeg_set_dec_src(ctx, &src_buf->vb2_buf, &bs); if (mtk_jpeg_set_dec_dst(ctx, &jpeg_src_buf->dec_param, &dst_buf->vb2_buf, &fb)) goto dec_end;
+ schedule_delayed_work(&jpeg->job_timeout_work, + msecs_to_jiffies(MTK_JPEG_HW_TIMEOUT_MSEC)); + spin_lock_irqsave(&jpeg->hw_lock, flags); mtk_jpeg_dec_reset(jpeg->reg_base); mtk_jpeg_dec_set_config(jpeg->reg_base,
In mtk_jpegdec_worker, if error occurs in mtk_jpeg_set_dec_dst, it will start the timeout worker and invoke v4l2_m2m_job_finish at the same time. This will break the logic of design for there should be only one function to call v4l2_m2m_job_finish. But now the timeout handler and mtk_jpegdec_worker will both invoke it.
Fix it by start the worker only if mtk_jpeg_set_dec_dst successfully finished.
Fixes: da4ede4b7fd6 ("media: mtk-jpeg: move data/code inside CONFIG_OF blocks") Signed-off-by: Zheng Wang zyytlz.wz@163.com Signed-off-by: Dmitry Osipenko dmitry.osipenko@collabora.com Cc: stable@vger.kernel.org --- v2: - put the patches into a single series suggested by Dmitry --- drivers/media/platform/mediatek/jpeg/mtk_jpeg_core.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/drivers/media/platform/mediatek/jpeg/mtk_jpeg_core.c b/drivers/media/platform/mediatek/jpeg/mtk_jpeg_core.c index a39acde2724a..c3456c700c07 100644 --- a/drivers/media/platform/mediatek/jpeg/mtk_jpeg_core.c +++ b/drivers/media/platform/mediatek/jpeg/mtk_jpeg_core.c @@ -1749,9 +1749,6 @@ static void mtk_jpegdec_worker(struct work_struct *work) v4l2_m2m_src_buf_remove(ctx->fh.m2m_ctx); v4l2_m2m_dst_buf_remove(ctx->fh.m2m_ctx);
- schedule_delayed_work(&comp_jpeg[hw_id]->job_timeout_work, - msecs_to_jiffies(MTK_JPEG_HW_TIMEOUT_MSEC)); - mtk_jpeg_set_dec_src(ctx, &src_buf->vb2_buf, &bs); if (mtk_jpeg_set_dec_dst(ctx, &jpeg_src_buf->dec_param, @@ -1761,6 +1758,9 @@ static void mtk_jpegdec_worker(struct work_struct *work) goto setdst_end; }
+ schedule_delayed_work(&comp_jpeg[hw_id]->job_timeout_work, + msecs_to_jiffies(MTK_JPEG_HW_TIMEOUT_MSEC)); + spin_lock_irqsave(&comp_jpeg[hw_id]->hw_lock, flags); ctx->total_frame_num++; mtk_jpeg_dec_reset(comp_jpeg[hw_id]->reg_base);
On 06/11/2023 15:48, Zheng Wang wrote:
Hello,
This v2 series fixes the use-after-free bug in mtk_jpeg_dec_device_run. This patch fixes the security bug in chrome-os. It inclues reverting the incomplete fix before and make the right fix. Also,it fixes the error of timeout-worker-schedule in multiple-core devices.
- Remove cancel worker in mtk_jpeg_remove for the worker is only
registered in single-core device but we try to cacnel it in both single-core and multiple-core devices.
- Fix use-after-free bug by delay the schedule_delayed_work only if
mtk_jpeg_set_dec_dst runs successfully.
- Delay the schedule_delayed_work in mtk_jpegdec_worker as it has same
code logic in mtk_jpeg_dec_device_run.
version 2 changes
-put the patches into on series suggested by Dmitry
Zheng Wang (3): media: mtk-jpeg: Remove cancel worker in mtk_jpeg_remove to avoid the crash of multi-core JPEG devices media: mtk-jpeg: Fix use after free bug due to error path handling in mtk_jpeg_dec_device_run media: mtk-jpeg: Fix timeout schedule error in mtk_jpegdec_worker.
.../media/platform/mediatek/jpeg/mtk_jpeg_core.c | 13 ++++++------- 1 file changed, 6 insertions(+), 7 deletions(-)
Please don't resend. If it is in patchwork.linuxtv.org then it is good. Fixes (unless they fix something really nasty) can take several weeks before they are picked up. I usually schedule 2 or 3 rounds per kernel cycle where I go through all the pending patches with fixes.
Regards,
Hans
Hans Verkuil hverkuil@xs4all.nl 于2023年11月7日周二 18:14写道：
On 06/11/2023 15:48, Zheng Wang wrote:
Hello,
This v2 series fixes the use-after-free bug in mtk_jpeg_dec_device_run. This patch fixes the security bug in chrome-os. It inclues reverting the incomplete fix before and make the right fix. Also,it fixes the error of timeout-worker-schedule in multiple-core devices.
- Remove cancel worker in mtk_jpeg_remove for the worker is only
registered in single-core device but we try to cacnel it in both single-core and multiple-core devices.
- Fix use-after-free bug by delay the schedule_delayed_work only if
mtk_jpeg_set_dec_dst runs successfully.
- Delay the schedule_delayed_work in mtk_jpegdec_worker as it has same
code logic in mtk_jpeg_dec_device_run.
version 2 changes
-put the patches into on series suggested by Dmitry
Zheng Wang (3): media: mtk-jpeg: Remove cancel worker in mtk_jpeg_remove to avoid the crash of multi-core JPEG devices media: mtk-jpeg: Fix use after free bug due to error path handling in mtk_jpeg_dec_device_run media: mtk-jpeg: Fix timeout schedule error in mtk_jpegdec_worker.
.../media/platform/mediatek/jpeg/mtk_jpeg_core.c | 13 ++++++------- 1 file changed, 6 insertions(+), 7 deletions(-)
Please don't resend. If it is in patchwork.linuxtv.org then it is good. Fixes (unless they fix something really nasty) can take several weeks before they are picked up. I usually schedule 2 or 3 rounds per kernel cycle where I go through all the pending patches with fixes.
Oh Sorry for my rudeness. And I'm deeply appreciative of your contributions to the community.
Best regards, Zheng
Regards,
Hans
linux-stable-mirror@lists.linaro.org
-
Hans Verkuil -
kernel test robot -
Zheng Hacker -
Zheng Wang