* [RFC PATCH V1 1/2] clk: use atomic runtime pm api in clk_core_is_enabled @ 2017-12-22 9:46 Dong Aisheng 2017-12-22 9:46 ` [RFC PATCH V1 2/2] clk: add lock for clk_core_is_enabled Dong Aisheng ` (2 more replies) 0 siblings, 3 replies; 6+ messages in thread From: Dong Aisheng @ 2017-12-22 9:46 UTC (permalink / raw) To: linux-clk Cc: linux-kernel, linux-arm-kernel, sboyd, mturquette, Dong Aisheng, Ulf Hansson, Marek Szyprowski Current clk_pm_runtime_put is using pm_runtime_put_sync which is not safe to be called in clk_core_is_enabled as it should be able to run in atomic context. Thus use pm_runtime_put instead which is atomic safe. Cc: Stephen Boyd <sboyd@codeaurora.org> Cc: Michael Turquette <mturquette@baylibre.com> Cc: Ulf Hansson <ulf.hansson@linaro.org> Cc: Marek Szyprowski <m.szyprowski@samsung.com> Fixes: 9a34b45397e5 ("clk: Add support for runtime PM") Signed-off-by: Dong Aisheng <aisheng.dong@nxp.com> --- drivers/clk/clk.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/clk/clk.c b/drivers/clk/clk.c index 5ec5809..e24968f 100644 --- a/drivers/clk/clk.c +++ b/drivers/clk/clk.c @@ -227,7 +227,8 @@ static bool clk_core_is_enabled(struct clk_core *core) ret = core->ops->is_enabled(core->hw); done: - clk_pm_runtime_put(core); + if (core->dev) + pm_runtime_put(core->dev); return ret; } -- 2.7.4 ^ permalink raw reply related [flat|nested] 6+ messages in thread
* [RFC PATCH V1 2/2] clk: add lock for clk_core_is_enabled 2017-12-22 9:46 [RFC PATCH V1 1/2] clk: use atomic runtime pm api in clk_core_is_enabled Dong Aisheng @ 2017-12-22 9:46 ` Dong Aisheng 2017-12-27 1:29 ` Stephen Boyd 2017-12-22 12:51 ` [RFC PATCH V1 1/2] clk: use atomic runtime pm api in clk_core_is_enabled Ulf Hansson 2017-12-27 1:38 ` Stephen Boyd 2 siblings, 1 reply; 6+ messages in thread From: Dong Aisheng @ 2017-12-22 9:46 UTC (permalink / raw) To: linux-clk; +Cc: linux-kernel, linux-arm-kernel, sboyd, mturquette, Dong Aisheng According to design doc, .is_enabled should be protected by enable lock. Then users don't have to protect it against enable/disable operation in clock drivers. See: Documentation/clk.txt "The enable lock is a spinlock and is held across calls to the .enable, .disable and .is_enabled operations." Cc: Stephen Boyd <sboyd@codeaurora.org> Cc: Michael Turquette <mturquette@baylibre.com> Signed-off-by: Dong Aisheng <aisheng.dong@nxp.com> --- drivers/clk/clk.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-) diff --git a/drivers/clk/clk.c b/drivers/clk/clk.c index e24968f..d6e2d5c 100644 --- a/drivers/clk/clk.c +++ b/drivers/clk/clk.c @@ -198,14 +198,19 @@ static bool clk_core_is_prepared(struct clk_core *core) static bool clk_core_is_enabled(struct clk_core *core) { + unsigned long flags; bool ret = false; + flags = clk_enable_lock(); + /* * .is_enabled is only mandatory for clocks that gate * fall back to software usage counter if .is_enabled is missing */ - if (!core->ops->is_enabled) + if (!core->ops->is_enabled) { + clk_enable_unlock(flags); return core->enable_count; + } /* * Check if clock controller's device is runtime active before @@ -230,6 +235,8 @@ static bool clk_core_is_enabled(struct clk_core *core) if (core->dev) pm_runtime_put(core->dev); + clk_enable_unlock(flags); + return ret; } -- 2.7.4 ^ permalink raw reply related [flat|nested] 6+ messages in thread
* Re: [RFC PATCH V1 2/2] clk: add lock for clk_core_is_enabled 2017-12-22 9:46 ` [RFC PATCH V1 2/2] clk: add lock for clk_core_is_enabled Dong Aisheng @ 2017-12-27 1:29 ` Stephen Boyd 2018-01-17 2:57 ` Dong Aisheng 0 siblings, 1 reply; 6+ messages in thread From: Stephen Boyd @ 2017-12-27 1:29 UTC (permalink / raw) To: Dong Aisheng; +Cc: linux-clk, linux-kernel, linux-arm-kernel, mturquette On 12/22, Dong Aisheng wrote: > According to design doc, .is_enabled should be protected by enable lock. > Then users don't have to protect it against enable/disable operation > in clock drivers. > > See: Documentation/clk.txt > "The enable lock is a spinlock and is held across calls to the .enable, > .disable and .is_enabled operations." > > Cc: Stephen Boyd <sboyd@codeaurora.org> > Cc: Michael Turquette <mturquette@baylibre.com> > Signed-off-by: Dong Aisheng <aisheng.dong@nxp.com> > --- > drivers/clk/clk.c | 9 ++++++++- > 1 file changed, 8 insertions(+), 1 deletion(-) > > diff --git a/drivers/clk/clk.c b/drivers/clk/clk.c > index e24968f..d6e2d5c 100644 > --- a/drivers/clk/clk.c > +++ b/drivers/clk/clk.c > @@ -198,14 +198,19 @@ static bool clk_core_is_prepared(struct clk_core *core) > > static bool clk_core_is_enabled(struct clk_core *core) > { > + unsigned long flags; > bool ret = false; > > + flags = clk_enable_lock(); > + > /* > * .is_enabled is only mandatory for clocks that gate > * fall back to software usage counter if .is_enabled is missing > */ > - if (!core->ops->is_enabled) > + if (!core->ops->is_enabled) { > + clk_enable_unlock(flags); > return core->enable_count; > + } > > /* > * Check if clock controller's device is runtime active before > @@ -230,6 +235,8 @@ static bool clk_core_is_enabled(struct clk_core *core) > if (core->dev) > pm_runtime_put(core->dev); > > + clk_enable_unlock(flags); > + > return ret; > } It doesn't really make any sense to hold the enable lock inside the clk_core_is_enabled() function, unless you want to do something else with the information of the enable state with that lock held. Otherwise, seeing if a clk is enabled is a one-shot read of the enabled state, which could just as easily change after the function returns because the lock is released. We should update the documentation. -- Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, a Linux Foundation Collaborative Project ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [RFC PATCH V1 2/2] clk: add lock for clk_core_is_enabled 2017-12-27 1:29 ` Stephen Boyd @ 2018-01-17 2:57 ` Dong Aisheng 0 siblings, 0 replies; 6+ messages in thread From: Dong Aisheng @ 2018-01-17 2:57 UTC (permalink / raw) To: Stephen Boyd Cc: Dong Aisheng, linux-clk, linux-kernel, linux-arm-kernel, mturquette, linux-imx On Tue, Dec 26, 2017 at 05:29:18PM -0800, Stephen Boyd wrote: > On 12/22, Dong Aisheng wrote: > > According to design doc, .is_enabled should be protected by enable lock. > > Then users don't have to protect it against enable/disable operation > > in clock drivers. > > > > See: Documentation/clk.txt > > "The enable lock is a spinlock and is held across calls to the .enable, > > .disable and .is_enabled operations." > > > > Cc: Stephen Boyd <sboyd@codeaurora.org> > > Cc: Michael Turquette <mturquette@baylibre.com> > > Signed-off-by: Dong Aisheng <aisheng.dong@nxp.com> > > --- > > drivers/clk/clk.c | 9 ++++++++- > > 1 file changed, 8 insertions(+), 1 deletion(-) > > > > diff --git a/drivers/clk/clk.c b/drivers/clk/clk.c > > index e24968f..d6e2d5c 100644 > > --- a/drivers/clk/clk.c > > +++ b/drivers/clk/clk.c > > @@ -198,14 +198,19 @@ static bool clk_core_is_prepared(struct clk_core *core) > > > > static bool clk_core_is_enabled(struct clk_core *core) > > { > > + unsigned long flags; > > bool ret = false; > > > > + flags = clk_enable_lock(); > > + > > /* > > * .is_enabled is only mandatory for clocks that gate > > * fall back to software usage counter if .is_enabled is missing > > */ > > - if (!core->ops->is_enabled) > > + if (!core->ops->is_enabled) { > > + clk_enable_unlock(flags); > > return core->enable_count; > > + } > > > > /* > > * Check if clock controller's device is runtime active before > > @@ -230,6 +235,8 @@ static bool clk_core_is_enabled(struct clk_core *core) > > if (core->dev) > > pm_runtime_put(core->dev); > > > > + clk_enable_unlock(flags); > > + > > return ret; > > } > > It doesn't really make any sense to hold the enable lock inside > the clk_core_is_enabled() function, unless you want to do > something else with the information of the enable state with that > lock held. Otherwise, seeing if a clk is enabled is a one-shot > read of the enabled state, which could just as easily change > after the function returns because the lock is released. > > We should update the documentation. > Yes, you're absolutely right. I could draft a patch to update it later. Thanks Regards Dong Aisheng > -- > Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, > a Linux Foundation Collaborative Project > -- > To unsubscribe from this list: send the line "unsubscribe linux-clk" in > the body of a message to majordomo@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [RFC PATCH V1 1/2] clk: use atomic runtime pm api in clk_core_is_enabled 2017-12-22 9:46 [RFC PATCH V1 1/2] clk: use atomic runtime pm api in clk_core_is_enabled Dong Aisheng 2017-12-22 9:46 ` [RFC PATCH V1 2/2] clk: add lock for clk_core_is_enabled Dong Aisheng @ 2017-12-22 12:51 ` Ulf Hansson 2017-12-27 1:38 ` Stephen Boyd 2 siblings, 0 replies; 6+ messages in thread From: Ulf Hansson @ 2017-12-22 12:51 UTC (permalink / raw) To: Dong Aisheng Cc: linux-clk, Linux Kernel Mailing List, linux-arm-kernel, Stephen Boyd, Mike Turquette, Marek Szyprowski On 22 December 2017 at 10:46, Dong Aisheng <aisheng.dong@nxp.com> wrote: > Current clk_pm_runtime_put is using pm_runtime_put_sync which > is not safe to be called in clk_core_is_enabled as it should > be able to run in atomic context. > > Thus use pm_runtime_put instead which is atomic safe. > > Cc: Stephen Boyd <sboyd@codeaurora.org> > Cc: Michael Turquette <mturquette@baylibre.com> > Cc: Ulf Hansson <ulf.hansson@linaro.org> > Cc: Marek Szyprowski <m.szyprowski@samsung.com> > Fixes: 9a34b45397e5 ("clk: Add support for runtime PM") > Signed-off-by: Dong Aisheng <aisheng.dong@nxp.com> Reviewed-by: Ulf Hansson <ulf.hansson@linaro.org> Kind regards Uffe > --- > drivers/clk/clk.c | 3 ++- > 1 file changed, 2 insertions(+), 1 deletion(-) > > diff --git a/drivers/clk/clk.c b/drivers/clk/clk.c > index 5ec5809..e24968f 100644 > --- a/drivers/clk/clk.c > +++ b/drivers/clk/clk.c > @@ -227,7 +227,8 @@ static bool clk_core_is_enabled(struct clk_core *core) > > ret = core->ops->is_enabled(core->hw); > done: > - clk_pm_runtime_put(core); > + if (core->dev) > + pm_runtime_put(core->dev); > > return ret; > } > -- > 2.7.4 > ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [RFC PATCH V1 1/2] clk: use atomic runtime pm api in clk_core_is_enabled 2017-12-22 9:46 [RFC PATCH V1 1/2] clk: use atomic runtime pm api in clk_core_is_enabled Dong Aisheng 2017-12-22 9:46 ` [RFC PATCH V1 2/2] clk: add lock for clk_core_is_enabled Dong Aisheng 2017-12-22 12:51 ` [RFC PATCH V1 1/2] clk: use atomic runtime pm api in clk_core_is_enabled Ulf Hansson @ 2017-12-27 1:38 ` Stephen Boyd 2 siblings, 0 replies; 6+ messages in thread From: Stephen Boyd @ 2017-12-27 1:38 UTC (permalink / raw) To: Dong Aisheng Cc: linux-clk, linux-kernel, linux-arm-kernel, mturquette, Ulf Hansson, Marek Szyprowski On 12/22, Dong Aisheng wrote: > Current clk_pm_runtime_put is using pm_runtime_put_sync which > is not safe to be called in clk_core_is_enabled as it should > be able to run in atomic context. > > Thus use pm_runtime_put instead which is atomic safe. > > Cc: Stephen Boyd <sboyd@codeaurora.org> > Cc: Michael Turquette <mturquette@baylibre.com> > Cc: Ulf Hansson <ulf.hansson@linaro.org> > Cc: Marek Szyprowski <m.szyprowski@samsung.com> > Fixes: 9a34b45397e5 ("clk: Add support for runtime PM") > Signed-off-by: Dong Aisheng <aisheng.dong@nxp.com> > --- Applied to clk-fixes -- Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, a Linux Foundation Collaborative Project ^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2018-01-17 2:57 UTC | newest] Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2017-12-22 9:46 [RFC PATCH V1 1/2] clk: use atomic runtime pm api in clk_core_is_enabled Dong Aisheng 2017-12-22 9:46 ` [RFC PATCH V1 2/2] clk: add lock for clk_core_is_enabled Dong Aisheng 2017-12-27 1:29 ` Stephen Boyd 2018-01-17 2:57 ` Dong Aisheng 2017-12-22 12:51 ` [RFC PATCH V1 1/2] clk: use atomic runtime pm api in clk_core_is_enabled Ulf Hansson 2017-12-27 1:38 ` Stephen Boyd
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).