| Message ID | 20240222010517.1573931-1-cyndis@kapsi.fi (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | gpu: host1x: Skip reset assert on Tegra186 | expand |
On Thu Feb 22, 2024 at 2:05 AM CET, Mikko Perttunen wrote: > From: Mikko Perttunen <mperttunen@nvidia.com> > > On Tegra186, secure world applications may need to access host1x > during suspend/resume, and rely on the kernel to keep Host1x out > of reset during the suspend cycle. As such, as a quirk, > skip asserting Host1x's reset on Tegra186. > > We don't need to keep the clocks enabled, as BPMP ensures the clock > stays on while Host1x is being used. On newer SoC's, the reset line > is inaccessible, so there is no need for the quirk. > > Signed-off-by: Mikko Perttunen <mperttunen@nvidia.com> > --- > drivers/gpu/host1x/dev.c | 15 +++++++++------ > drivers/gpu/host1x/dev.h | 6 ++++++ > 2 files changed, 15 insertions(+), 6 deletions(-) Applied to drm-misc-fixes, though I added the Fixes: tag that Jon mentioned in reply to v1 of this as well as his Reviewed-by and Tested-by as well, since this is pretty much the same patch except for the comments. Thanks, Thierry
On 2/23/24 03:06, Thierry Reding wrote: > On Thu Feb 22, 2024 at 2:05 AM CET, Mikko Perttunen wrote: >> From: Mikko Perttunen <mperttunen@nvidia.com> >> >> On Tegra186, secure world applications may need to access host1x >> during suspend/resume, and rely on the kernel to keep Host1x out >> of reset during the suspend cycle. As such, as a quirk, >> skip asserting Host1x's reset on Tegra186. >> >> We don't need to keep the clocks enabled, as BPMP ensures the clock >> stays on while Host1x is being used. On newer SoC's, the reset line >> is inaccessible, so there is no need for the quirk. >> >> Signed-off-by: Mikko Perttunen <mperttunen@nvidia.com> >> --- >> drivers/gpu/host1x/dev.c | 15 +++++++++------ >> drivers/gpu/host1x/dev.h | 6 ++++++ >> 2 files changed, 15 insertions(+), 6 deletions(-) > > Applied to drm-misc-fixes, though I added the Fixes: tag that Jon > mentioned in reply to v1 of this as well as his Reviewed-by and > Tested-by as well, since this is pretty much the same patch except > for the comments. Thanks. Sorry for messing it up :p Mikko > > Thanks, > Thierry
diff --git a/drivers/gpu/host1x/dev.c b/drivers/gpu/host1x/dev.c index 42fd504abbcd..89983d7d73ca 100644 --- a/drivers/gpu/host1x/dev.c +++ b/drivers/gpu/host1x/dev.c @@ -169,6 +169,7 @@ static const struct host1x_info host1x06_info = { .num_sid_entries = ARRAY_SIZE(tegra186_sid_table), .sid_table = tegra186_sid_table, .reserve_vblank_syncpts = false, + .skip_reset_assert = true, }; static const struct host1x_sid_entry tegra194_sid_table[] = { @@ -680,13 +681,15 @@ static int __maybe_unused host1x_runtime_suspend(struct device *dev) host1x_intr_stop(host); host1x_syncpt_save(host); - err = reset_control_bulk_assert(host->nresets, host->resets); - if (err) { - dev_err(dev, "failed to assert reset: %d\n", err); - goto resume_host1x; - } + if (!host->info->skip_reset_assert) { + err = reset_control_bulk_assert(host->nresets, host->resets); + if (err) { + dev_err(dev, "failed to assert reset: %d\n", err); + goto resume_host1x; + } - usleep_range(1000, 2000); + usleep_range(1000, 2000); + } clk_disable_unprepare(host->clk); reset_control_bulk_release(host->nresets, host->resets); diff --git a/drivers/gpu/host1x/dev.h b/drivers/gpu/host1x/dev.h index c8e302de7625..6143c2a61d70 100644 --- a/drivers/gpu/host1x/dev.h +++ b/drivers/gpu/host1x/dev.h @@ -116,6 +116,12 @@ struct host1x_info { * the display driver disables VBLANK increments. */ bool reserve_vblank_syncpts; + /* + * On T186, secure world applications may require access to host1x + * during suspend/resume. To allow this, we need to leave host1x + * not in reset. + */ + bool skip_reset_assert; }; struct host1x {