Message ID | 20230911-topic-mars-v2-20-3dac84b88c4b@linaro.org |
---|---|
State | Superseded |
Headers | show |
Series | Venus cleanups | expand |
Hi Konrad, On Fr, 2024-02-09 at 22:10 +0100, Konrad Dybcio wrote: > All of the resets are toggled together. Use the bulk api to save on some > code complexity. > > The delay between resets is now correctly determined by the reset > framework. If this is a recent change, could you reference the commit? > Signed-off-by: Konrad Dybcio <konrad.dybcio@linaro.org> > --- > drivers/media/platform/qcom/venus/core.c | 15 ++++++++++----- > drivers/media/platform/qcom/venus/core.h | 4 ++-- > drivers/media/platform/qcom/venus/pm_helpers.c | 15 +++------------ > 3 files changed, 15 insertions(+), 19 deletions(-) > > diff --git a/drivers/media/platform/qcom/venus/core.c b/drivers/media/platform/qcom/venus/core.c > index 873affe17537..ff5601a5ce77 100644 > --- a/drivers/media/platform/qcom/venus/core.c > +++ b/drivers/media/platform/qcom/venus/core.c > @@ -328,11 +328,16 @@ static int venus_probe(struct platform_device *pdev) > if (ret) > return ret; > > - for (i = 0; i < res->resets_num; i++) { > - core->resets[i] = devm_reset_control_get_exclusive(dev, res->resets[i]); > - if (IS_ERR(core->resets[i])) > - return PTR_ERR(core->resets[i]); > - } > + core->resets = devm_kcalloc(dev, res->resets_num, sizeof(*core->resets), GFP_KERNEL); Since VIDC_RESETS_NUM_MAX is only 2, I don't think a separate allocation is worth it. > + if (res->resets_num && !core->resets) > + return -ENOMEM; > + > + for (i = 0; i < res->resets_num; i++) > + core->resets[i].id = res->resets[i]; > + > + ret = devm_reset_control_bulk_get_exclusive(dev, res->resets_num, core->resets); > + if (ret) > + return dev_err_probe(dev, ret, "Failed to get resets\n"); > > ret = venus_get_resources(core); > if (ret) > diff --git a/drivers/media/platform/qcom/venus/core.h b/drivers/media/platform/qcom/venus/core.h > index 6ecaa3e38cac..2376b9cbdf2c 100644 > --- a/drivers/media/platform/qcom/venus/core.h > +++ b/drivers/media/platform/qcom/venus/core.h > @@ -130,7 +130,7 @@ struct venus_format { > * @pmdomains: a pointer to a list of pmdomains > * @opp_dl_venus: an device-link for device OPP > * @opp_pmdomain: an OPP power-domain > - * @resets: an array of reset signals > + * @resets: a reset_control_bulk_data array of hardware reset signals > * @vdev_dec: a reference to video device structure for decoder instances > * @vdev_enc: a reference to video device structure for encoder instances > * @v4l2_dev: a holder for v4l2 device structure > @@ -183,7 +183,7 @@ struct venus_core { > struct dev_pm_domain_list *pmdomains; > struct device_link *opp_dl_venus; > struct device *opp_pmdomain; > - struct reset_control *resets[VIDC_RESETS_NUM_MAX]; > + struct reset_control_bulk_data *resets; Any reason not to just keep this as an array[VIDC_RESETS_NUM_MAX]? > struct video_device *vdev_dec; > struct video_device *vdev_enc; > struct v4l2_device v4l2_dev; > diff --git a/drivers/media/platform/qcom/venus/pm_helpers.c b/drivers/media/platform/qcom/venus/pm_helpers.c > index 9df8f2292c17..170fb131cb1e 100644 > --- a/drivers/media/platform/qcom/venus/pm_helpers.c > +++ b/drivers/media/platform/qcom/venus/pm_helpers.c > @@ -865,21 +865,12 @@ void vcodec_domains_put(struct venus_core *core) > static int core_resets_reset(struct venus_core *core) > { > const struct venus_resources *res = core->res; > - unsigned int i; > int ret; > > - for (i = 0; i < res->resets_num; i++) { > - ret = reset_control_assert(core->resets[i]); > - if (ret) > - goto err; > - > - usleep_range(150, 250); > - ret = reset_control_deassert(core->resets[i]); > - if (ret) > - goto err; > - } > + ret = reset_control_bulk_reset(res->resets_num, core->resets); > + if (ret) > + dev_err(core->dev, "Failed to toggle resets: %d\n", ret); > > -err: > return ret; Could be simplified to: return reset_control_bulk_reset(res->resets_num, core- >resets); regards Philipp
On 21.02.2024 14:34, Philipp Zabel wrote: > On Mi, 2024-02-14 at 22:20 +0100, Konrad Dybcio wrote: >> On 14.02.2024 14:31, Philipp Zabel wrote: >>> Hi Konrad, >>> >>> On Fr, 2024-02-09 at 22:10 +0100, Konrad Dybcio wrote: >>>> All of the resets are toggled together. Use the bulk api to save on some >>>> code complexity. >>>> >>>> The delay between resets is now correctly determined by the reset >>>> framework. >>> >>> If this is a recent change, could you reference the commit? >> >> It's a series that recently landed in -next [1] > > Missing link? Yes, sorry! Konrad [1] https://lore.kernel.org/linux-arm-msm/20240105-topic-venus_reset-v2-0-c37eba13b5ce@linaro.org/
diff --git a/drivers/media/platform/qcom/venus/core.c b/drivers/media/platform/qcom/venus/core.c index 873affe17537..ff5601a5ce77 100644 --- a/drivers/media/platform/qcom/venus/core.c +++ b/drivers/media/platform/qcom/venus/core.c @@ -328,11 +328,16 @@ static int venus_probe(struct platform_device *pdev) if (ret) return ret; - for (i = 0; i < res->resets_num; i++) { - core->resets[i] = devm_reset_control_get_exclusive(dev, res->resets[i]); - if (IS_ERR(core->resets[i])) - return PTR_ERR(core->resets[i]); - } + core->resets = devm_kcalloc(dev, res->resets_num, sizeof(*core->resets), GFP_KERNEL); + if (res->resets_num && !core->resets) + return -ENOMEM; + + for (i = 0; i < res->resets_num; i++) + core->resets[i].id = res->resets[i]; + + ret = devm_reset_control_bulk_get_exclusive(dev, res->resets_num, core->resets); + if (ret) + return dev_err_probe(dev, ret, "Failed to get resets\n"); ret = venus_get_resources(core); if (ret) diff --git a/drivers/media/platform/qcom/venus/core.h b/drivers/media/platform/qcom/venus/core.h index 6ecaa3e38cac..2376b9cbdf2c 100644 --- a/drivers/media/platform/qcom/venus/core.h +++ b/drivers/media/platform/qcom/venus/core.h @@ -130,7 +130,7 @@ struct venus_format { * @pmdomains: a pointer to a list of pmdomains * @opp_dl_venus: an device-link for device OPP * @opp_pmdomain: an OPP power-domain - * @resets: an array of reset signals + * @resets: a reset_control_bulk_data array of hardware reset signals * @vdev_dec: a reference to video device structure for decoder instances * @vdev_enc: a reference to video device structure for encoder instances * @v4l2_dev: a holder for v4l2 device structure @@ -183,7 +183,7 @@ struct venus_core { struct dev_pm_domain_list *pmdomains; struct device_link *opp_dl_venus; struct device *opp_pmdomain; - struct reset_control *resets[VIDC_RESETS_NUM_MAX]; + struct reset_control_bulk_data *resets; struct video_device *vdev_dec; struct video_device *vdev_enc; struct v4l2_device v4l2_dev; diff --git a/drivers/media/platform/qcom/venus/pm_helpers.c b/drivers/media/platform/qcom/venus/pm_helpers.c index 9df8f2292c17..170fb131cb1e 100644 --- a/drivers/media/platform/qcom/venus/pm_helpers.c +++ b/drivers/media/platform/qcom/venus/pm_helpers.c @@ -865,21 +865,12 @@ void vcodec_domains_put(struct venus_core *core) static int core_resets_reset(struct venus_core *core) { const struct venus_resources *res = core->res; - unsigned int i; int ret; - for (i = 0; i < res->resets_num; i++) { - ret = reset_control_assert(core->resets[i]); - if (ret) - goto err; - - usleep_range(150, 250); - ret = reset_control_deassert(core->resets[i]); - if (ret) - goto err; - } + ret = reset_control_bulk_reset(res->resets_num, core->resets); + if (ret) + dev_err(core->dev, "Failed to toggle resets: %d\n", ret); -err: return ret; }
All of the resets are toggled together. Use the bulk api to save on some code complexity. The delay between resets is now correctly determined by the reset framework. Signed-off-by: Konrad Dybcio <konrad.dybcio@linaro.org> --- drivers/media/platform/qcom/venus/core.c | 15 ++++++++++----- drivers/media/platform/qcom/venus/core.h | 4 ++-- drivers/media/platform/qcom/venus/pm_helpers.c | 15 +++------------ 3 files changed, 15 insertions(+), 19 deletions(-)