| Message ID | 20230921122641.RFT.v2.9.Iea33274908b6b258955f45a8aaf6f5bba24ad6cd@changeid (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | drm: call drm_atomic_helper_shutdown() at the right times | expand |
On 21.09.2023 21:26, Douglas Anderson wrote: > Based on grepping through the source code this driver appears to be > missing a call to drm_atomic_helper_shutdown() at system shutdown time > and at driver unbind time. Among other things, this means that if a > panel is in use that it won't be cleanly powered off at system > shutdown time. > > The fact that we should call drm_atomic_helper_shutdown() in the case > of OS shutdown/restart and at driver remove (or unbind) time comes > straight out of the kernel doc "driver instance overview" in > drm_drv.c. > > A few notes about this fix: > - When adding drm_atomic_helper_shutdown() to the unbind path, I added > it after drm_kms_helper_poll_fini() since that's when other drivers > seemed to have it. > - Technically with a previous patch, ("drm/atomic-helper: > drm_atomic_helper_shutdown(NULL) should be a noop"), we don't > actually need to check to see if our "drm" pointer is NULL before > calling drm_atomic_helper_shutdown(). We'll leave the "if" test in, > though, so that this patch can land without any dependencies. It > could potentially be removed later. > - This patch also makes sure to set the drvdata to NULL in the case of > bind errors to make sure that shutdown can't access freed data. > > Suggested-by: Maxime Ripard <mripard@kernel.org> > Reviewed-by: Maxime Ripard <mripard@kernel.org> > Signed-off-by: Douglas Anderson <dianders@chromium.org> Seems to be working fine on all my test Exynos-based boards with display. Tested-by: Marek Szyprowski <m.szyprowski@samsung.com> Reviewed-by: Marek Szyprowski <m.szyprowski@samsung.com> > --- > This commit is only compile-time tested. > > (no changes since v1) > > drivers/gpu/drm/exynos/exynos_drm_drv.c | 11 +++++++++++ > 1 file changed, 11 insertions(+) > > diff --git a/drivers/gpu/drm/exynos/exynos_drm_drv.c b/drivers/gpu/drm/exynos/exynos_drm_drv.c > index 8399256cb5c9..5380fb6c55ae 100644 > --- a/drivers/gpu/drm/exynos/exynos_drm_drv.c > +++ b/drivers/gpu/drm/exynos/exynos_drm_drv.c > @@ -300,6 +300,7 @@ static int exynos_drm_bind(struct device *dev) > drm_mode_config_cleanup(drm); > exynos_drm_cleanup_dma(drm); > kfree(private); > + dev_set_drvdata(dev, NULL); > err_free_drm: > drm_dev_put(drm); > > @@ -313,6 +314,7 @@ static void exynos_drm_unbind(struct device *dev) > drm_dev_unregister(drm); > > drm_kms_helper_poll_fini(drm); > + drm_atomic_helper_shutdown(drm); > > component_unbind_all(drm->dev, drm); > drm_mode_config_cleanup(drm); > @@ -350,9 +352,18 @@ static int exynos_drm_platform_remove(struct platform_device *pdev) > return 0; > } > > +static void exynos_drm_platform_shutdown(struct platform_device *pdev) > +{ > + struct drm_device *drm = platform_get_drvdata(pdev); > + > + if (drm) > + drm_atomic_helper_shutdown(drm); > +} > + > static struct platform_driver exynos_drm_platform_driver = { > .probe = exynos_drm_platform_probe, > .remove = exynos_drm_platform_remove, > + .shutdown = exynos_drm_platform_shutdown, > .driver = { > .name = "exynos-drm", > .pm = &exynos_drm_pm_ops, Best regards
Thanks for testing. :) Acked-by : Inki Dae <inki.dae@samsung.com> 2023년 9월 22일 (금) 오후 3:00, Marek Szyprowski <m.szyprowski@samsung.com>님이 작성: > > > On 21.09.2023 21:26, Douglas Anderson wrote: > > Based on grepping through the source code this driver appears to be > > missing a call to drm_atomic_helper_shutdown() at system shutdown time > > and at driver unbind time. Among other things, this means that if a > > panel is in use that it won't be cleanly powered off at system > > shutdown time. > > > > The fact that we should call drm_atomic_helper_shutdown() in the case > > of OS shutdown/restart and at driver remove (or unbind) time comes > > straight out of the kernel doc "driver instance overview" in > > drm_drv.c. > > > > A few notes about this fix: > > - When adding drm_atomic_helper_shutdown() to the unbind path, I added > > it after drm_kms_helper_poll_fini() since that's when other drivers > > seemed to have it. > > - Technically with a previous patch, ("drm/atomic-helper: > > drm_atomic_helper_shutdown(NULL) should be a noop"), we don't > > actually need to check to see if our "drm" pointer is NULL before > > calling drm_atomic_helper_shutdown(). We'll leave the "if" test in, > > though, so that this patch can land without any dependencies. It > > could potentially be removed later. > > - This patch also makes sure to set the drvdata to NULL in the case of > > bind errors to make sure that shutdown can't access freed data. > > > > Suggested-by: Maxime Ripard <mripard@kernel.org> > > Reviewed-by: Maxime Ripard <mripard@kernel.org> > > Signed-off-by: Douglas Anderson <dianders@chromium.org> > > Seems to be working fine on all my test Exynos-based boards with display. > > Tested-by: Marek Szyprowski <m.szyprowski@samsung.com> > > Reviewed-by: Marek Szyprowski <m.szyprowski@samsung.com> > > > --- > > This commit is only compile-time tested. > > > > (no changes since v1) > > > > drivers/gpu/drm/exynos/exynos_drm_drv.c | 11 +++++++++++ > > 1 file changed, 11 insertions(+) > > > > diff --git a/drivers/gpu/drm/exynos/exynos_drm_drv.c b/drivers/gpu/drm/exynos/exynos_drm_drv.c > > index 8399256cb5c9..5380fb6c55ae 100644 > > --- a/drivers/gpu/drm/exynos/exynos_drm_drv.c > > +++ b/drivers/gpu/drm/exynos/exynos_drm_drv.c > > @@ -300,6 +300,7 @@ static int exynos_drm_bind(struct device *dev) > > drm_mode_config_cleanup(drm); > > exynos_drm_cleanup_dma(drm); > > kfree(private); > > + dev_set_drvdata(dev, NULL); > > err_free_drm: > > drm_dev_put(drm); > > > > @@ -313,6 +314,7 @@ static void exynos_drm_unbind(struct device *dev) > > drm_dev_unregister(drm); > > > > drm_kms_helper_poll_fini(drm); > > + drm_atomic_helper_shutdown(drm); > > > > component_unbind_all(drm->dev, drm); > > drm_mode_config_cleanup(drm); > > @@ -350,9 +352,18 @@ static int exynos_drm_platform_remove(struct platform_device *pdev) > > return 0; > > } > > > > +static void exynos_drm_platform_shutdown(struct platform_device *pdev) > > +{ > > + struct drm_device *drm = platform_get_drvdata(pdev); > > + > > + if (drm) > > + drm_atomic_helper_shutdown(drm); > > +} > > + > > static struct platform_driver exynos_drm_platform_driver = { > > .probe = exynos_drm_platform_probe, > > .remove = exynos_drm_platform_remove, > > + .shutdown = exynos_drm_platform_shutdown, > > .driver = { > > .name = "exynos-drm", > > .pm = &exynos_drm_pm_ops, > > Best regards > -- > Marek Szyprowski, PhD > Samsung R&D Institute Poland >
Hi, On Thu, Oct 5, 2023 at 7:20 PM Inki Dae <daeinki@gmail.com> wrote: > > Thanks for testing. :) > > Acked-by : Inki Dae <inki.dae@samsung.com> Inki: does that mean you'd like this to go through drm-misc? I'm happy to do that, but there are no dependencies here so this could easily go through your tree. > 2023년 9월 22일 (금) 오후 3:00, Marek Szyprowski <m.szyprowski@samsung.com>님이 작성: > > > > > > On 21.09.2023 21:26, Douglas Anderson wrote: > > > Based on grepping through the source code this driver appears to be > > > missing a call to drm_atomic_helper_shutdown() at system shutdown time > > > and at driver unbind time. Among other things, this means that if a > > > panel is in use that it won't be cleanly powered off at system > > > shutdown time. > > > > > > The fact that we should call drm_atomic_helper_shutdown() in the case > > > of OS shutdown/restart and at driver remove (or unbind) time comes > > > straight out of the kernel doc "driver instance overview" in > > > drm_drv.c. > > > > > > A few notes about this fix: > > > - When adding drm_atomic_helper_shutdown() to the unbind path, I added > > > it after drm_kms_helper_poll_fini() since that's when other drivers > > > seemed to have it. > > > - Technically with a previous patch, ("drm/atomic-helper: > > > drm_atomic_helper_shutdown(NULL) should be a noop"), we don't > > > actually need to check to see if our "drm" pointer is NULL before > > > calling drm_atomic_helper_shutdown(). We'll leave the "if" test in, > > > though, so that this patch can land without any dependencies. It > > > could potentially be removed later. > > > - This patch also makes sure to set the drvdata to NULL in the case of > > > bind errors to make sure that shutdown can't access freed data. > > > > > > Suggested-by: Maxime Ripard <mripard@kernel.org> > > > Reviewed-by: Maxime Ripard <mripard@kernel.org> > > > Signed-off-by: Douglas Anderson <dianders@chromium.org> > > > > Seems to be working fine on all my test Exynos-based boards with display. > > > > Tested-by: Marek Szyprowski <m.szyprowski@samsung.com> > > > > Reviewed-by: Marek Szyprowski <m.szyprowski@samsung.com> > > > > > --- > > > This commit is only compile-time tested. > > > > > > (no changes since v1) > > > > > > drivers/gpu/drm/exynos/exynos_drm_drv.c | 11 +++++++++++ > > > 1 file changed, 11 insertions(+) > > > > > > diff --git a/drivers/gpu/drm/exynos/exynos_drm_drv.c b/drivers/gpu/drm/exynos/exynos_drm_drv.c > > > index 8399256cb5c9..5380fb6c55ae 100644 > > > --- a/drivers/gpu/drm/exynos/exynos_drm_drv.c > > > +++ b/drivers/gpu/drm/exynos/exynos_drm_drv.c > > > @@ -300,6 +300,7 @@ static int exynos_drm_bind(struct device *dev) > > > drm_mode_config_cleanup(drm); > > > exynos_drm_cleanup_dma(drm); > > > kfree(private); > > > + dev_set_drvdata(dev, NULL); > > > err_free_drm: > > > drm_dev_put(drm); > > > > > > @@ -313,6 +314,7 @@ static void exynos_drm_unbind(struct device *dev) > > > drm_dev_unregister(drm); > > > > > > drm_kms_helper_poll_fini(drm); > > > + drm_atomic_helper_shutdown(drm); > > > > > > component_unbind_all(drm->dev, drm); > > > drm_mode_config_cleanup(drm); > > > @@ -350,9 +352,18 @@ static int exynos_drm_platform_remove(struct platform_device *pdev) > > > return 0; > > > } > > > > > > +static void exynos_drm_platform_shutdown(struct platform_device *pdev) > > > +{ > > > + struct drm_device *drm = platform_get_drvdata(pdev); > > > + > > > + if (drm) > > > + drm_atomic_helper_shutdown(drm); > > > +} > > > + > > > static struct platform_driver exynos_drm_platform_driver = { > > > .probe = exynos_drm_platform_probe, > > > .remove = exynos_drm_platform_remove, > > > + .shutdown = exynos_drm_platform_shutdown, > > > .driver = { > > > .name = "exynos-drm", > > > .pm = &exynos_drm_pm_ops, > > > > Best regards > > -- > > Marek Szyprowski, PhD > > Samsung R&D Institute Poland > >
Hi, 2023년 10월 6일 (금) 오후 10:51, Doug Anderson <dianders@chromium.org>님이 작성: > > Hi, > > On Thu, Oct 5, 2023 at 7:20 PM Inki Dae <daeinki@gmail.com> wrote: > > > > Thanks for testing. :) > > > > Acked-by : Inki Dae <inki.dae@samsung.com> > > Inki: does that mean you'd like this to go through drm-misc? I'm happy > to do that, but there are no dependencies here so this could easily go > through your tree. Ah, you are right. No dependency here. I will pick it up. Thanks, Inki Dae > > > > 2023년 9월 22일 (금) 오후 3:00, Marek Szyprowski <m.szyprowski@samsung.com>님이 작성: > > > > > > > > > On 21.09.2023 21:26, Douglas Anderson wrote: > > > > Based on grepping through the source code this driver appears to be > > > > missing a call to drm_atomic_helper_shutdown() at system shutdown time > > > > and at driver unbind time. Among other things, this means that if a > > > > panel is in use that it won't be cleanly powered off at system > > > > shutdown time. > > > > > > > > The fact that we should call drm_atomic_helper_shutdown() in the case > > > > of OS shutdown/restart and at driver remove (or unbind) time comes > > > > straight out of the kernel doc "driver instance overview" in > > > > drm_drv.c. > > > > > > > > A few notes about this fix: > > > > - When adding drm_atomic_helper_shutdown() to the unbind path, I added > > > > it after drm_kms_helper_poll_fini() since that's when other drivers > > > > seemed to have it. > > > > - Technically with a previous patch, ("drm/atomic-helper: > > > > drm_atomic_helper_shutdown(NULL) should be a noop"), we don't > > > > actually need to check to see if our "drm" pointer is NULL before > > > > calling drm_atomic_helper_shutdown(). We'll leave the "if" test in, > > > > though, so that this patch can land without any dependencies. It > > > > could potentially be removed later. > > > > - This patch also makes sure to set the drvdata to NULL in the case of > > > > bind errors to make sure that shutdown can't access freed data. > > > > > > > > Suggested-by: Maxime Ripard <mripard@kernel.org> > > > > Reviewed-by: Maxime Ripard <mripard@kernel.org> > > > > Signed-off-by: Douglas Anderson <dianders@chromium.org> > > > > > > Seems to be working fine on all my test Exynos-based boards with display. > > > > > > Tested-by: Marek Szyprowski <m.szyprowski@samsung.com> > > > > > > Reviewed-by: Marek Szyprowski <m.szyprowski@samsung.com> > > > > > > > --- > > > > This commit is only compile-time tested. > > > > > > > > (no changes since v1) > > > > > > > > drivers/gpu/drm/exynos/exynos_drm_drv.c | 11 +++++++++++ > > > > 1 file changed, 11 insertions(+) > > > > > > > > diff --git a/drivers/gpu/drm/exynos/exynos_drm_drv.c b/drivers/gpu/drm/exynos/exynos_drm_drv.c > > > > index 8399256cb5c9..5380fb6c55ae 100644 > > > > --- a/drivers/gpu/drm/exynos/exynos_drm_drv.c > > > > +++ b/drivers/gpu/drm/exynos/exynos_drm_drv.c > > > > @@ -300,6 +300,7 @@ static int exynos_drm_bind(struct device *dev) > > > > drm_mode_config_cleanup(drm); > > > > exynos_drm_cleanup_dma(drm); > > > > kfree(private); > > > > + dev_set_drvdata(dev, NULL); > > > > err_free_drm: > > > > drm_dev_put(drm); > > > > > > > > @@ -313,6 +314,7 @@ static void exynos_drm_unbind(struct device *dev) > > > > drm_dev_unregister(drm); > > > > > > > > drm_kms_helper_poll_fini(drm); > > > > + drm_atomic_helper_shutdown(drm); > > > > > > > > component_unbind_all(drm->dev, drm); > > > > drm_mode_config_cleanup(drm); > > > > @@ -350,9 +352,18 @@ static int exynos_drm_platform_remove(struct platform_device *pdev) > > > > return 0; > > > > } > > > > > > > > +static void exynos_drm_platform_shutdown(struct platform_device *pdev) > > > > +{ > > > > + struct drm_device *drm = platform_get_drvdata(pdev); > > > > + > > > > + if (drm) > > > > + drm_atomic_helper_shutdown(drm); > > > > +} > > > > + > > > > static struct platform_driver exynos_drm_platform_driver = { > > > > .probe = exynos_drm_platform_probe, > > > > .remove = exynos_drm_platform_remove, > > > > + .shutdown = exynos_drm_platform_shutdown, > > > > .driver = { > > > > .name = "exynos-drm", > > > > .pm = &exynos_drm_pm_ops, > > > > > > Best regards > > > -- > > > Marek Szyprowski, PhD > > > Samsung R&D Institute Poland > > >
diff --git a/drivers/gpu/drm/exynos/exynos_drm_drv.c b/drivers/gpu/drm/exynos/exynos_drm_drv.c index 8399256cb5c9..5380fb6c55ae 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_drv.c +++ b/drivers/gpu/drm/exynos/exynos_drm_drv.c @@ -300,6 +300,7 @@ static int exynos_drm_bind(struct device *dev) drm_mode_config_cleanup(drm); exynos_drm_cleanup_dma(drm); kfree(private); + dev_set_drvdata(dev, NULL); err_free_drm: drm_dev_put(drm); @@ -313,6 +314,7 @@ static void exynos_drm_unbind(struct device *dev) drm_dev_unregister(drm); drm_kms_helper_poll_fini(drm); + drm_atomic_helper_shutdown(drm); component_unbind_all(drm->dev, drm); drm_mode_config_cleanup(drm); @@ -350,9 +352,18 @@ static int exynos_drm_platform_remove(struct platform_device *pdev) return 0; } +static void exynos_drm_platform_shutdown(struct platform_device *pdev) +{ + struct drm_device *drm = platform_get_drvdata(pdev); + + if (drm) + drm_atomic_helper_shutdown(drm); +} + static struct platform_driver exynos_drm_platform_driver = { .probe = exynos_drm_platform_probe, .remove = exynos_drm_platform_remove, + .shutdown = exynos_drm_platform_shutdown, .driver = { .name = "exynos-drm", .pm = &exynos_drm_pm_ops,
Based on grepping through the source code this driver appears to be missing a call to drm_atomic_helper_shutdown() at system shutdown time and at driver unbind time. Among other things, this means that if a panel is in use that it won't be cleanly powered off at system shutdown time. The fact that we should call drm_atomic_helper_shutdown() in the case of OS shutdown/restart and at driver remove (or unbind) time comes straight out of the kernel doc "driver instance overview" in drm_drv.c. A few notes about this fix: - When adding drm_atomic_helper_shutdown() to the unbind path, I added it after drm_kms_helper_poll_fini() since that's when other drivers seemed to have it. - Technically with a previous patch, ("drm/atomic-helper: drm_atomic_helper_shutdown(NULL) should be a noop"), we don't actually need to check to see if our "drm" pointer is NULL before calling drm_atomic_helper_shutdown(). We'll leave the "if" test in, though, so that this patch can land without any dependencies. It could potentially be removed later. - This patch also makes sure to set the drvdata to NULL in the case of bind errors to make sure that shutdown can't access freed data. Suggested-by: Maxime Ripard <mripard@kernel.org> Reviewed-by: Maxime Ripard <mripard@kernel.org> Signed-off-by: Douglas Anderson <dianders@chromium.org> --- This commit is only compile-time tested. (no changes since v1) drivers/gpu/drm/exynos/exynos_drm_drv.c | 11 +++++++++++ 1 file changed, 11 insertions(+)