| Message ID | 20210730025238.22456-5-yong.wu@mediatek.com |
|---|---|
| State | New |
| Headers | show |
| Series | Clean up "mediatek,larb" | expand |
On 30.07.21 04:52, Yong Wu wrote: > MediaTek IOMMU-SMI diagram is like below. all the consumer connect with > smi-larb, then connect with smi-common. > > M4U > | > smi-common > | > ------------- > | | ... > | | > larb1 larb2 > | | > vdec venc > > When the consumer works, it should enable the smi-larb's power which > also need enable the smi-common's power firstly. > > Thus, First of all, use the device link connect the consumer and the > smi-larbs. then add device link between the smi-larb and smi-common. > > This patch adds device_link between the consumer and the larbs. > > When device_link_add, I add the flag DL_FLAG_STATELESS to avoid calling > pm_runtime_xx to keep the original status of clocks. It can avoid two > issues: > 1) Display HW show fastlogo abnormally reported in [1]. At the beggining, > all the clocks are enabled before entering kernel, but the clocks for > display HW(always in larb0) will be gated after clk_enable and clk_disable > called from device_link_add(->pm_runtime_resume) and rpm_idle. The clock > operation happened before display driver probe. At that time, the display > HW will be abnormal. > > 2) A deadlock issue reported in [2]. Use DL_FLAG_STATELESS to skip > pm_runtime_xx to avoid the deadlock. > > Corresponding, DL_FLAG_AUTOREMOVE_CONSUMER can't be added, then > device_link_removed should be added explicitly. > > [1] https://lore.kernel.org/linux-mediatek/1564213888.22908.4.camel@mhfsdcap03/ > [2] https://lore.kernel.org/patchwork/patch/1086569/ > > Suggested-by: Tomasz Figa <tfiga@chromium.org> > Signed-off-by: Yong Wu <yong.wu@mediatek.com> > Tested-by: Dafna Hirschfeld <dafna.hirschfeld@collabora.com> # on mt8173 Hi, unfortunately, I have to take back the Tested-by tag. I am now testing the mtk-vcodec with latest kernel + patches sent from the mailing list: https://gitlab.collabora.com/eballetbo/linux/-/commits/topic/chromeos/chromeos-5.14 which includes this patchset. On chromeos I open a video conference with googl-meet which cause the mtk-vcodec vp8 encoder to run. If I kill it with `killall -9 chrome` I get some page fault messages from the iommu: [ 837.255952] mtk-iommu 10205000.iommu: fault type=0x5 iova=0xfcff0001 pa=0x0 larb=0 port=0 layer=1 read [ 837.265696] mtk-iommu 10205000.iommu: fault type=0x5 iova=0xfcff0001 pa=0x0 larb=0 port=0 layer=1 read [ 837.282367] mtk-iommu 10205000.iommu: fault type=0x5 iova=0xfcff0001 pa=0x0 larb=0 port=0 layer=1 read [ 837.299028] mtk-iommu 10205000.iommu: fault type=0x5 iova=0xfcff0001 pa=0x0 larb=0 port=0 layer=1 read [ 837.315683] mtk-iommu 10205000.iommu: fault type=0x5 iova=0xfcff0001 pa=0x0 larb=0 port=0 layer=1 read [ 837.332345] mtk-iommu 10205000.iommu: fault type=0x5 iova=0xfcff0001 pa=0x0 larb=0 port=0 layer=1 read [ 837.349004] mtk-iommu 10205000.iommu: fault type=0x5 iova=0xfcff0001 pa=0x0 larb=0 port=0 layer=1 read [ 837.365665] mtk-iommu 10205000.iommu: fault type=0x5 iova=0xfcff0001 pa=0x0 larb=0 port=0 layer=1 read [ 837.382329] mtk-iommu 10205000.iommu: fault type=0x5 iova=0xfcff0001 pa=0x0 larb=0 port=0 layer=1 read [ 837.400002] mtk-iommu 10205000.iommu: fault type=0x5 iova=0xfcff0001 pa=0x0 larb=0 port=0 layer=1 read In addition, running the encoder tests from the shell: sudo --user=#1000 /usr/local/libexec/chrome-binary-tests/video_encode_accelerator_tests --gtest_filter=VideoEncoderTest.FlushAtEndOfStream_Multiple* --codec=vp8 /usr/local/share/tast/data/chromiumos/tast/local/bundles/cros/video/data/tulip2-320x180.yuv --disable_validator At some point it fails with the error [ 5472.161821] [MTK_V4L2][ERROR] mtk_vcodec_wait_for_done_ctx:32: [290] ctx->type=1, cmd=1, wait_event_interruptible_timeout time=1000ms out 0 0! [ 5472.174678] [MTK_VCODEC][ERROR][290]: vp8_enc_encode_frame() irq_status=0 failed [ 5472.182687] [MTK_V4L2][ERROR] mtk_venc_worker:1239: venc_if_encode failed=-5 If you have any idea of what might be the problem or how to debug? Thanks, Dafna > --- > drivers/iommu/mtk_iommu.c | 22 ++++++++++++++++++++++ > drivers/iommu/mtk_iommu_v1.c | 20 +++++++++++++++++++- > 2 files changed, 41 insertions(+), 1 deletion(-) > > diff --git a/drivers/iommu/mtk_iommu.c b/drivers/iommu/mtk_iommu.c > index a02dde094788..ee742900cf4b 100644 > --- a/drivers/iommu/mtk_iommu.c > +++ b/drivers/iommu/mtk_iommu.c > @@ -571,22 +571,44 @@ static struct iommu_device *mtk_iommu_probe_device(struct device *dev) > { > struct iommu_fwspec *fwspec = dev_iommu_fwspec_get(dev); > struct mtk_iommu_data *data; > + struct device_link *link; > + struct device *larbdev; > + unsigned int larbid; > > if (!fwspec || fwspec->ops != &mtk_iommu_ops) > return ERR_PTR(-ENODEV); /* Not a iommu client device */ > > data = dev_iommu_priv_get(dev); > > + /* > + * Link the consumer device with the smi-larb device(supplier) > + * The device in each a larb is a independent HW. thus only link > + * one larb here. > + */ > + larbid = MTK_M4U_TO_LARB(fwspec->ids[0]); > + larbdev = data->larb_imu[larbid].dev; > + link = device_link_add(dev, larbdev, > + DL_FLAG_PM_RUNTIME | DL_FLAG_STATELESS); > + if (!link) > + dev_err(dev, "Unable to link %s\n", dev_name(larbdev)); > return &data->iommu; > } > > static void mtk_iommu_release_device(struct device *dev) > { > struct iommu_fwspec *fwspec = dev_iommu_fwspec_get(dev); > + struct mtk_iommu_data *data; > + struct device *larbdev; > + unsigned int larbid; > > if (!fwspec || fwspec->ops != &mtk_iommu_ops) > return; > > + data = dev_iommu_priv_get(dev); > + larbid = MTK_M4U_TO_LARB(fwspec->ids[0]); > + larbdev = data->larb_imu[larbid].dev; > + device_link_remove(dev, larbdev); > + > iommu_fwspec_free(dev); > } > > diff --git a/drivers/iommu/mtk_iommu_v1.c b/drivers/iommu/mtk_iommu_v1.c > index c259433f1130..806d4200665b 100644 > --- a/drivers/iommu/mtk_iommu_v1.c > +++ b/drivers/iommu/mtk_iommu_v1.c > @@ -424,7 +424,9 @@ static struct iommu_device *mtk_iommu_probe_device(struct device *dev) > struct iommu_fwspec *fwspec = dev_iommu_fwspec_get(dev); > struct of_phandle_args iommu_spec; > struct mtk_iommu_data *data; > - int err, idx = 0; > + int err, idx = 0, larbid; > + struct device_link *link; > + struct device *larbdev; > > /* > * In the deferred case, free the existed fwspec if the dev already has, > @@ -454,6 +456,14 @@ static struct iommu_device *mtk_iommu_probe_device(struct device *dev) > > data = dev_iommu_priv_get(dev); > > + /* Link the consumer device with the smi-larb device(supplier) */ > + larbid = mt2701_m4u_to_larb(fwspec->ids[0]); > + larbdev = data->larb_imu[larbid].dev; > + link = device_link_add(dev, larbdev, > + DL_FLAG_PM_RUNTIME | DL_FLAG_STATELESS); > + if (!link) > + dev_err(dev, "Unable to link %s\n", dev_name(larbdev)); > + > return &data->iommu; > } > > @@ -474,10 +484,18 @@ static void mtk_iommu_probe_finalize(struct device *dev) > static void mtk_iommu_release_device(struct device *dev) > { > struct iommu_fwspec *fwspec = dev_iommu_fwspec_get(dev); > + struct mtk_iommu_data *data; > + struct device *larbdev; > + unsigned int larbid; > > if (!fwspec || fwspec->ops != &mtk_iommu_ops) > return; > > + data = dev_iommu_priv_get(dev); > + larbid = mt2701_m4u_to_larb(fwspec->ids[0]); > + larbdev = data->larb_imu[larbid].dev; > + device_link_remove(dev, larbdev); > + > iommu_fwspec_free(dev); > } > >
On Thu, 2021-08-05 at 15:22 +0200, Dafna Hirschfeld wrote: > > On 30.07.21 04:52, Yong Wu wrote: > > MediaTek IOMMU-SMI diagram is like below. all the consumer connect > > with > > smi-larb, then connect with smi-common. > > > > M4U > > | > > smi-common > > | > > ------------- > > | | ... > > | | > > larb1 larb2 > > | | > > vdec venc > > > > When the consumer works, it should enable the smi-larb's power > > which > > also need enable the smi-common's power firstly. > > > > Thus, First of all, use the device link connect the consumer and > > the > > smi-larbs. then add device link between the smi-larb and smi- > > common. > > > > This patch adds device_link between the consumer and the larbs. > > > > When device_link_add, I add the flag DL_FLAG_STATELESS to avoid > > calling > > pm_runtime_xx to keep the original status of clocks. It can avoid > > two > > issues: > > 1) Display HW show fastlogo abnormally reported in [1]. At the > > beggining, > > all the clocks are enabled before entering kernel, but the clocks > > for > > display HW(always in larb0) will be gated after clk_enable and > > clk_disable > > called from device_link_add(->pm_runtime_resume) and rpm_idle. The > > clock > > operation happened before display driver probe. At that time, the > > display > > HW will be abnormal. > > > > 2) A deadlock issue reported in [2]. Use DL_FLAG_STATELESS to skip > > pm_runtime_xx to avoid the deadlock. > > > > Corresponding, DL_FLAG_AUTOREMOVE_CONSUMER can't be added, then > > device_link_removed should be added explicitly. > > > > [1] > > https://lore.kernel.org/linux-mediatek/1564213888.22908.4.camel@mhfsdcap03/ > > [2] https://lore.kernel.org/patchwork/patch/1086569/ > > > > Suggested-by: Tomasz Figa <tfiga@chromium.org> > > Signed-off-by: Yong Wu <yong.wu@mediatek.com> > > Tested-by: Dafna Hirschfeld <dafna.hirschfeld@collabora.com> # on > > mt8173 > > Hi, unfortunately, I have to take back the Tested-by tag. sorry for inconvenience you. (and sorry for reply late, there is something wrong about my local mail server.) > I am now testing the mtk-vcodec with latest kernel + patches sent > from the mailing list: > https://gitlab.collabora.com/eballetbo/linux/-/commits/topic/chromeos/chromeos-5.14 > which includes this patchset. > > On chromeos I open a video conference with googl-meet which cause the > mtk-vcodec vp8 encoder to run. > If I kill it with `killall -9 chrome` I get some page fault messages > from the iommu: Does the "git bisect" point to this patch? If you don't kill it, Does it also have these error below? I don't know what happen about "killall -9 chrome', Does it cause freeing some buffer? > > [ 837.255952] mtk-iommu 10205000.iommu: fault type=0x5 > iova=0xfcff0001 pa=0x0 larb=0 port=0 layer=1 read This means "larb0 port0" translation fault. If I am not wrong, you work at mt8173, from [0], this is DISP_OVL0. May be "killall -9 chrome" free the buffer(iova:0xfcff0000) that DISP_OVL is accessing, then iommu complain it is not a valid iova. [0] https://elixir.bootlin.com/linux/v5.14-rc1/source/include/dt-bindings/memory/mt8173-larb-port.h#L19 > [ 837.265696] mtk-iommu 10205000.iommu: fault type=0x5 > iova=0xfcff0001 pa=0x0 larb=0 port=0 layer=1 read > [ 837.282367] mtk-iommu 10205000.iommu: fault type=0x5 > iova=0xfcff0001 pa=0x0 larb=0 port=0 layer=1 read > [ 837.299028] mtk-iommu 10205000.iommu: fault type=0x5 > iova=0xfcff0001 pa=0x0 larb=0 port=0 layer=1 read > [ 837.315683] mtk-iommu 10205000.iommu: fault type=0x5 > iova=0xfcff0001 pa=0x0 larb=0 port=0 layer=1 read > [ 837.332345] mtk-iommu 10205000.iommu: fault type=0x5 > iova=0xfcff0001 pa=0x0 larb=0 port=0 layer=1 read > [ 837.349004] mtk-iommu 10205000.iommu: fault type=0x5 > iova=0xfcff0001 pa=0x0 larb=0 port=0 layer=1 read > [ 837.365665] mtk-iommu 10205000.iommu: fault type=0x5 > iova=0xfcff0001 pa=0x0 larb=0 port=0 layer=1 read > [ 837.382329] mtk-iommu 10205000.iommu: fault type=0x5 > iova=0xfcff0001 pa=0x0 larb=0 port=0 layer=1 read > [ 837.400002] mtk-iommu 10205000.iommu: fault type=0x5 > iova=0xfcff0001 pa=0x0 larb=0 port=0 layer=1 read > > In addition, running the encoder tests from the shell: > > sudo --user=#1000 /usr/local/libexec/chrome-binary- > tests/video_encode_accelerator_tests -- > gtest_filter=VideoEncoderTest.FlushAtEndOfStream_Multiple* > --codec=vp8 > /usr/local/share/tast/data/chromiumos/tast/local/bundles/cros/video/d > ata/tulip2-320x180.yuv --disable_validator > > At some point it fails with the error > > [ 5472.161821] [MTK_V4L2][ERROR] mtk_vcodec_wait_for_done_ctx:32: > [290] ctx->type=1, cmd=1, wait_event_interruptible_timeout > time=1000ms out 0 0! > [ 5472.174678] [MTK_VCODEC][ERROR][290]: vp8_enc_encode_frame() > irq_status=0 failed > [ 5472.182687] [MTK_V4L2][ERROR] mtk_venc_worker:1239: venc_if_encode > failed=-5 +our venc guy Irui. This looks VENC HW don't start to work. Does this caused by this patchset? this patchset only change the flow of power. I guess we should check if the power/clock for venc here is enable or not? > > > If you have any idea of what might be the problem or how to debug? > > Thanks, > Dafna > > > --
diff --git a/drivers/iommu/mtk_iommu.c b/drivers/iommu/mtk_iommu.c index a02dde094788..ee742900cf4b 100644 --- a/drivers/iommu/mtk_iommu.c +++ b/drivers/iommu/mtk_iommu.c @@ -571,22 +571,44 @@ static struct iommu_device *mtk_iommu_probe_device(struct device *dev) { struct iommu_fwspec *fwspec = dev_iommu_fwspec_get(dev); struct mtk_iommu_data *data; + struct device_link *link; + struct device *larbdev; + unsigned int larbid; if (!fwspec || fwspec->ops != &mtk_iommu_ops) return ERR_PTR(-ENODEV); /* Not a iommu client device */ data = dev_iommu_priv_get(dev); + /* + * Link the consumer device with the smi-larb device(supplier) + * The device in each a larb is a independent HW. thus only link + * one larb here. + */ + larbid = MTK_M4U_TO_LARB(fwspec->ids[0]); + larbdev = data->larb_imu[larbid].dev; + link = device_link_add(dev, larbdev, + DL_FLAG_PM_RUNTIME | DL_FLAG_STATELESS); + if (!link) + dev_err(dev, "Unable to link %s\n", dev_name(larbdev)); return &data->iommu; } static void mtk_iommu_release_device(struct device *dev) { struct iommu_fwspec *fwspec = dev_iommu_fwspec_get(dev); + struct mtk_iommu_data *data; + struct device *larbdev; + unsigned int larbid; if (!fwspec || fwspec->ops != &mtk_iommu_ops) return; + data = dev_iommu_priv_get(dev); + larbid = MTK_M4U_TO_LARB(fwspec->ids[0]); + larbdev = data->larb_imu[larbid].dev; + device_link_remove(dev, larbdev); + iommu_fwspec_free(dev); } diff --git a/drivers/iommu/mtk_iommu_v1.c b/drivers/iommu/mtk_iommu_v1.c index c259433f1130..806d4200665b 100644 --- a/drivers/iommu/mtk_iommu_v1.c +++ b/drivers/iommu/mtk_iommu_v1.c @@ -424,7 +424,9 @@ static struct iommu_device *mtk_iommu_probe_device(struct device *dev) struct iommu_fwspec *fwspec = dev_iommu_fwspec_get(dev); struct of_phandle_args iommu_spec; struct mtk_iommu_data *data; - int err, idx = 0; + int err, idx = 0, larbid; + struct device_link *link; + struct device *larbdev; /* * In the deferred case, free the existed fwspec if the dev already has, @@ -454,6 +456,14 @@ static struct iommu_device *mtk_iommu_probe_device(struct device *dev) data = dev_iommu_priv_get(dev); + /* Link the consumer device with the smi-larb device(supplier) */ + larbid = mt2701_m4u_to_larb(fwspec->ids[0]); + larbdev = data->larb_imu[larbid].dev; + link = device_link_add(dev, larbdev, + DL_FLAG_PM_RUNTIME | DL_FLAG_STATELESS); + if (!link) + dev_err(dev, "Unable to link %s\n", dev_name(larbdev)); + return &data->iommu; } @@ -474,10 +484,18 @@ static void mtk_iommu_probe_finalize(struct device *dev) static void mtk_iommu_release_device(struct device *dev) { struct iommu_fwspec *fwspec = dev_iommu_fwspec_get(dev); + struct mtk_iommu_data *data; + struct device *larbdev; + unsigned int larbid; if (!fwspec || fwspec->ops != &mtk_iommu_ops) return; + data = dev_iommu_priv_get(dev); + larbid = mt2701_m4u_to_larb(fwspec->ids[0]); + larbdev = data->larb_imu[larbid].dev; + device_link_remove(dev, larbdev); + iommu_fwspec_free(dev); }
MediaTek IOMMU-SMI diagram is like below. all the consumer connect with smi-larb, then connect with smi-common. M4U | smi-common | ------------- | | ... | | larb1 larb2 | | vdec venc When the consumer works, it should enable the smi-larb's power which also need enable the smi-common's power firstly. Thus, First of all, use the device link connect the consumer and the smi-larbs. then add device link between the smi-larb and smi-common. This patch adds device_link between the consumer and the larbs. When device_link_add, I add the flag DL_FLAG_STATELESS to avoid calling pm_runtime_xx to keep the original status of clocks. It can avoid two issues: 1) Display HW show fastlogo abnormally reported in [1]. At the beggining, all the clocks are enabled before entering kernel, but the clocks for display HW(always in larb0) will be gated after clk_enable and clk_disable called from device_link_add(->pm_runtime_resume) and rpm_idle. The clock operation happened before display driver probe. At that time, the display HW will be abnormal. 2) A deadlock issue reported in [2]. Use DL_FLAG_STATELESS to skip pm_runtime_xx to avoid the deadlock. Corresponding, DL_FLAG_AUTOREMOVE_CONSUMER can't be added, then device_link_removed should be added explicitly. [1] https://lore.kernel.org/linux-mediatek/1564213888.22908.4.camel@mhfsdcap03/ [2] https://lore.kernel.org/patchwork/patch/1086569/ Suggested-by: Tomasz Figa <tfiga@chromium.org> Signed-off-by: Yong Wu <yong.wu@mediatek.com> Tested-by: Dafna Hirschfeld <dafna.hirschfeld@collabora.com> # on mt8173 --- drivers/iommu/mtk_iommu.c | 22 ++++++++++++++++++++++ drivers/iommu/mtk_iommu_v1.c | 20 +++++++++++++++++++- 2 files changed, 41 insertions(+), 1 deletion(-)