| Message ID | 20201112230043.28987-1-aouledameur@baylibre.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | [v2] reset: make shared pulsed reset controls re-triggerable | expand |
Please find in this patchset [0] a use case of the reset_control_rearm() call on some of meson usb drivers: [0] https://lore.kernel.org/lkml/20201113000508.14702-1-aouledameur@baylibre.com/ On 13/11/2020 00:00, Amjad Ouled-Ameur wrote: > The current reset framework API does not allow to release what is done by > reset_control_reset(), IOW decrement triggered_count. Add the new > reset_control_rearm() call to do so. > > When reset_control_reset() has been called once, the counter > triggered_count, in the reset framework, is incremented i.e the resource > under the reset is in-use and the reset should not be done again. > reset_control_rearm() would be the way to state that the resource is > no longer used and, that from the caller's perspective, the reset can be > fired again if necessary. > > Signed-off-by: Amjad Ouled-Ameur <aouledameur@baylibre.com> > Reported-by: Jerome Brunet <jbrunet@baylibre.com> > --- > Change since v1: [0] > * Renamed the new call from reset_control_(array_)resettable to > reset_control_(array_)rearm > * Open-coded reset_control_array_rearm to check for errors before > decrementing triggered_count because we cannot roll back in case an > error occurs while decrementing one of the rstc. > * Reworded the new call's description. > > [0] https://lore.kernel.org/lkml/20201001132758.12280-1-aouledameur@baylibre.com > > drivers/reset/core.c | 73 +++++++++++++++++++++++++++++++++++++++++++ > include/linux/reset.h | 1 + > 2 files changed, 74 insertions(+) > > diff --git a/drivers/reset/core.c b/drivers/reset/core.c > index a2df88e90011..34e89aa0fb5e 100644 > --- a/drivers/reset/core.c > +++ b/drivers/reset/core.c > @@ -208,6 +208,39 @@ static int reset_control_array_reset(struct reset_control_array *resets) > return 0; > } > > +static int reset_control_array_rearm(struct reset_control_array *resets) > +{ > + struct reset_control *rstc; > + int i; > + > + for (i = 0; i < resets->num_rstcs; i++) { > + rstc = resets->rstc[i]; > + > + if (!rstc) > + continue; > + > + if (WARN_ON(IS_ERR(rstc))) > + return -EINVAL; > + > + if (rstc->shared) { > + if (WARN_ON(atomic_read(&rstc->deassert_count) != 0)) > + return -EINVAL; > + } else { > + if (!rstc->acquired) > + return -EPERM; > + } > + } > + > + for (i = 0; i < resets->num_rstcs; i++) { > + rstc = resets->rstc[i]; > + > + if (rstc && rstc->shared) > + WARN_ON(atomic_dec_return(&rstc->triggered_count) < 0); > + } > + > + return 0; > +} > + > static int reset_control_array_assert(struct reset_control_array *resets) > { > int ret, i; > @@ -325,6 +358,46 @@ int reset_control_reset(struct reset_control *rstc) > } > EXPORT_SYMBOL_GPL(reset_control_reset); > > +/** > + * reset_control_rearm - allow shared reset line to be re-triggered" > + * @rstc: reset controller > + * > + * On a shared reset line the actual reset pulse is only triggered once for the > + * lifetime of the reset_control instance, except if this call is used. > + * > + * Calls to this function must be balanced with calls to reset_control_reset, > + * a warning is thrown in case triggered_count ever dips below 0. > + * > + * Consumers must not use reset_control_(de)assert on shared reset lines when > + * reset_control_reset or reset_control_rearm have been used. > + * > + * If rstc is NULL the function will just return 0. > + */ > +int reset_control_rearm(struct reset_control *rstc) > +{ > + if (!rstc) > + return 0; > + > + if (WARN_ON(IS_ERR(rstc))) > + return -EINVAL; > + > + if (reset_control_is_array(rstc)) > + return reset_control_array_rearm(rstc_to_array(rstc)); > + > + if (rstc->shared) { > + if (WARN_ON(atomic_read(&rstc->deassert_count) != 0)) > + return -EINVAL; > + > + WARN_ON(atomic_dec_return(&rstc->triggered_count) < 0); > + } else { > + if (!rstc->acquired) > + return -EPERM; > + } > + > + return 0; > +} > +EXPORT_SYMBOL_GPL(reset_control_rearm); > + > /** > * reset_control_assert - asserts the reset line > * @rstc: reset controller > diff --git a/include/linux/reset.h b/include/linux/reset.h > index 05aa9f440f48..439fec7112a9 100644 > --- a/include/linux/reset.h > +++ b/include/linux/reset.h > @@ -13,6 +13,7 @@ struct reset_control; > #ifdef CONFIG_RESET_CONTROLLER > > int reset_control_reset(struct reset_control *rstc); > +int reset_control_rearm(struct reset_control *rstc); > int reset_control_assert(struct reset_control *rstc); > int reset_control_deassert(struct reset_control *rstc); > int reset_control_status(struct reset_control *rstc);
On Fri, 2020-11-13 at 00:00 +0100, Amjad Ouled-Ameur wrote: > The current reset framework API does not allow to release what is done by > reset_control_reset(), IOW decrement triggered_count. Add the new > reset_control_rearm() call to do so. > > When reset_control_reset() has been called once, the counter > triggered_count, in the reset framework, is incremented i.e the resource > under the reset is in-use and the reset should not be done again. > reset_control_rearm() would be the way to state that the resource is > no longer used and, that from the caller's perspective, the reset can be > fired again if necessary. > > Signed-off-by: Amjad Ouled-Ameur <aouledameur@baylibre.com> > Reported-by: Jerome Brunet <jbrunet@baylibre.com> > --- > Change since v1: [0] > * Renamed the new call from reset_control_(array_)resettable to > reset_control_(array_)rearm > * Open-coded reset_control_array_rearm to check for errors before > decrementing triggered_count because we cannot roll back in case an > error occurs while decrementing one of the rstc. > * Reworded the new call's description. Thank you, applied to reset/next. regards Philipp
On Fri 13 Nov 2020 at 16:04, Philipp Zabel <p.zabel@pengutronix.de> wrote: > On Fri, 2020-11-13 at 00:00 +0100, Amjad Ouled-Ameur wrote: >> The current reset framework API does not allow to release what is done by >> reset_control_reset(), IOW decrement triggered_count. Add the new >> reset_control_rearm() call to do so. >> >> When reset_control_reset() has been called once, the counter >> triggered_count, in the reset framework, is incremented i.e the resource >> under the reset is in-use and the reset should not be done again. >> reset_control_rearm() would be the way to state that the resource is >> no longer used and, that from the caller's perspective, the reset can be >> fired again if necessary. >> >> Signed-off-by: Amjad Ouled-Ameur <aouledameur@baylibre.com> >> Reported-by: Jerome Brunet <jbrunet@baylibre.com> >> --- >> Change since v1: [0] >> * Renamed the new call from reset_control_(array_)resettable to >> reset_control_(array_)rearm >> * Open-coded reset_control_array_rearm to check for errors before >> decrementing triggered_count because we cannot roll back in case an >> error occurs while decrementing one of the rstc. >> * Reworded the new call's description. > > Thank you, applied to reset/next. Hi Philipp, Would it be possible to get an immutable branch/tag with this ? It would allow to move forward on the USB side, without waiting for the next rc1. Thx Jerome > > regards > Philipp
On 13/11/2020 16:04, Philipp Zabel wrote: > On Fri, 2020-11-13 at 00:00 +0100, Amjad Ouled-Ameur wrote: >> The current reset framework API does not allow to release what is done by >> reset_control_reset(), IOW decrement triggered_count. Add the new >> reset_control_rearm() call to do so. >> >> When reset_control_reset() has been called once, the counter >> triggered_count, in the reset framework, is incremented i.e the resource >> under the reset is in-use and the reset should not be done again. >> reset_control_rearm() would be the way to state that the resource is >> no longer used and, that from the caller's perspective, the reset can be >> fired again if necessary. >> >> Signed-off-by: Amjad Ouled-Ameur <aouledameur@baylibre.com> >> Reported-by: Jerome Brunet <jbrunet@baylibre.com> >> --- >> Change since v1: [0] >> * Renamed the new call from reset_control_(array_)resettable to >> reset_control_(array_)rearm >> * Open-coded reset_control_array_rearm to check for errors before >> decrementing triggered_count because we cannot roll back in case an >> error occurs while decrementing one of the rstc. >> * Reworded the new call's description. > Thank you, applied to reset/next. > > regards > Philipp Thank you for reviewing and approving my patch ! Furthermore, I think your idea of open coding reset_control_array_rearm was accurate, and should be also applied to reset_control_array_reset() and reset_control_array_(de)assert() What do you think ? In case you agree it is necessary to do so, I can work out an upcoming patchset to fix this matter. Best, Amjad
On Fri, 2020-11-13 at 16:13 +0100, Jerome Brunet wrote: > On Fri 13 Nov 2020 at 16:04, Philipp Zabel <p.zabel@pengutronix.de> wrote: > > > On Fri, 2020-11-13 at 00:00 +0100, Amjad Ouled-Ameur wrote: > > > The current reset framework API does not allow to release what is done by > > > reset_control_reset(), IOW decrement triggered_count. Add the new > > > reset_control_rearm() call to do so. > > > > > > When reset_control_reset() has been called once, the counter > > > triggered_count, in the reset framework, is incremented i.e the resource > > > under the reset is in-use and the reset should not be done again. > > > reset_control_rearm() would be the way to state that the resource is > > > no longer used and, that from the caller's perspective, the reset can be > > > fired again if necessary. > > > > > > Signed-off-by: Amjad Ouled-Ameur <aouledameur@baylibre.com> > > > Reported-by: Jerome Brunet <jbrunet@baylibre.com> > > > --- > > > Change since v1: [0] > > > * Renamed the new call from reset_control_(array_)resettable to > > > reset_control_(array_)rearm > > > * Open-coded reset_control_array_rearm to check for errors before > > > decrementing triggered_count because we cannot roll back in case an > > > error occurs while decrementing one of the rstc. > > > * Reworded the new call's description. > > > > Thank you, applied to reset/next. > > Hi Philipp, > > Would it be possible to get an immutable branch/tag with this ? > It would allow to move forward on the USB side, without waiting for the > next rc1. Here you go, The following changes since commit 3650b228f83adda7e5ee532e2b90429c03f7b9ec: Linux 5.10-rc1 (2020-10-25 15:14:11 -0700) are available in the Git repository at: git://git.pengutronix.de/git/pza/linux.git reset/shared-retrigger for you to fetch changes up to 557acb3d2cd9c82de19f944f6cc967a347735385: reset: make shared pulsed reset controls re-triggerable (2020-11-16 17:05:28 +0100) ---------------------------------------------------------------- Amjad Ouled-Ameur (1): reset: make shared pulsed reset controls re-triggerable drivers/reset/core.c | 73 +++++++++++++++++++++++++++++++++++++++++++++++++++ include/linux/reset.h | 1 + 2 files changed, 74 insertions(+) regards Philipp
On Mon 16 Nov 2020 at 17:36, Philipp Zabel <p.zabel@pengutronix.de> wrote: > On Fri, 2020-11-13 at 16:13 +0100, Jerome Brunet wrote: >> On Fri 13 Nov 2020 at 16:04, Philipp Zabel <p.zabel@pengutronix.de> wrote: >> >> > On Fri, 2020-11-13 at 00:00 +0100, Amjad Ouled-Ameur wrote: >> > > The current reset framework API does not allow to release what is done by >> > > reset_control_reset(), IOW decrement triggered_count. Add the new >> > > reset_control_rearm() call to do so. >> > > >> > > When reset_control_reset() has been called once, the counter >> > > triggered_count, in the reset framework, is incremented i.e the resource >> > > under the reset is in-use and the reset should not be done again. >> > > reset_control_rearm() would be the way to state that the resource is >> > > no longer used and, that from the caller's perspective, the reset can be >> > > fired again if necessary. >> > > >> > > Signed-off-by: Amjad Ouled-Ameur <aouledameur@baylibre.com> >> > > Reported-by: Jerome Brunet <jbrunet@baylibre.com> >> > > --- >> > > Change since v1: [0] >> > > * Renamed the new call from reset_control_(array_)resettable to >> > > reset_control_(array_)rearm >> > > * Open-coded reset_control_array_rearm to check for errors before >> > > decrementing triggered_count because we cannot roll back in case an >> > > error occurs while decrementing one of the rstc. >> > > * Reworded the new call's description. >> > >> > Thank you, applied to reset/next. >> >> Hi Philipp, >> >> Would it be possible to get an immutable branch/tag with this ? >> It would allow to move forward on the USB side, without waiting for the >> next rc1. > > Here you go, > > The following changes since commit 3650b228f83adda7e5ee532e2b90429c03f7b9ec: > > Linux 5.10-rc1 (2020-10-25 15:14:11 -0700) > > are available in the Git repository at: > > git://git.pengutronix.de/git/pza/linux.git reset/shared-retrigger > > for you to fetch changes up to 557acb3d2cd9c82de19f944f6cc967a347735385: > > reset: make shared pulsed reset controls re-triggerable (2020-11-16 17:05:28 +0100) > > ---------------------------------------------------------------- > Amjad Ouled-Ameur (1): > reset: make shared pulsed reset controls re-triggerable > > drivers/reset/core.c | 73 +++++++++++++++++++++++++++++++++++++++++++++++++++ > include/linux/reset.h | 1 + > 2 files changed, 74 insertions(+) > Thx Philipp ! > regards > Philipp
diff --git a/drivers/reset/core.c b/drivers/reset/core.c index a2df88e90011..34e89aa0fb5e 100644 --- a/drivers/reset/core.c +++ b/drivers/reset/core.c @@ -208,6 +208,39 @@ static int reset_control_array_reset(struct reset_control_array *resets) return 0; } +static int reset_control_array_rearm(struct reset_control_array *resets) +{ + struct reset_control *rstc; + int i; + + for (i = 0; i < resets->num_rstcs; i++) { + rstc = resets->rstc[i]; + + if (!rstc) + continue; + + if (WARN_ON(IS_ERR(rstc))) + return -EINVAL; + + if (rstc->shared) { + if (WARN_ON(atomic_read(&rstc->deassert_count) != 0)) + return -EINVAL; + } else { + if (!rstc->acquired) + return -EPERM; + } + } + + for (i = 0; i < resets->num_rstcs; i++) { + rstc = resets->rstc[i]; + + if (rstc && rstc->shared) + WARN_ON(atomic_dec_return(&rstc->triggered_count) < 0); + } + + return 0; +} + static int reset_control_array_assert(struct reset_control_array *resets) { int ret, i; @@ -325,6 +358,46 @@ int reset_control_reset(struct reset_control *rstc) } EXPORT_SYMBOL_GPL(reset_control_reset); +/** + * reset_control_rearm - allow shared reset line to be re-triggered" + * @rstc: reset controller + * + * On a shared reset line the actual reset pulse is only triggered once for the + * lifetime of the reset_control instance, except if this call is used. + * + * Calls to this function must be balanced with calls to reset_control_reset, + * a warning is thrown in case triggered_count ever dips below 0. + * + * Consumers must not use reset_control_(de)assert on shared reset lines when + * reset_control_reset or reset_control_rearm have been used. + * + * If rstc is NULL the function will just return 0. + */ +int reset_control_rearm(struct reset_control *rstc) +{ + if (!rstc) + return 0; + + if (WARN_ON(IS_ERR(rstc))) + return -EINVAL; + + if (reset_control_is_array(rstc)) + return reset_control_array_rearm(rstc_to_array(rstc)); + + if (rstc->shared) { + if (WARN_ON(atomic_read(&rstc->deassert_count) != 0)) + return -EINVAL; + + WARN_ON(atomic_dec_return(&rstc->triggered_count) < 0); + } else { + if (!rstc->acquired) + return -EPERM; + } + + return 0; +} +EXPORT_SYMBOL_GPL(reset_control_rearm); + /** * reset_control_assert - asserts the reset line * @rstc: reset controller diff --git a/include/linux/reset.h b/include/linux/reset.h index 05aa9f440f48..439fec7112a9 100644 --- a/include/linux/reset.h +++ b/include/linux/reset.h @@ -13,6 +13,7 @@ struct reset_control; #ifdef CONFIG_RESET_CONTROLLER int reset_control_reset(struct reset_control *rstc); +int reset_control_rearm(struct reset_control *rstc); int reset_control_assert(struct reset_control *rstc); int reset_control_deassert(struct reset_control *rstc); int reset_control_status(struct reset_control *rstc);
The current reset framework API does not allow to release what is done by reset_control_reset(), IOW decrement triggered_count. Add the new reset_control_rearm() call to do so. When reset_control_reset() has been called once, the counter triggered_count, in the reset framework, is incremented i.e the resource under the reset is in-use and the reset should not be done again. reset_control_rearm() would be the way to state that the resource is no longer used and, that from the caller's perspective, the reset can be fired again if necessary. Signed-off-by: Amjad Ouled-Ameur <aouledameur@baylibre.com> Reported-by: Jerome Brunet <jbrunet@baylibre.com> --- Change since v1: [0] * Renamed the new call from reset_control_(array_)resettable to reset_control_(array_)rearm * Open-coded reset_control_array_rearm to check for errors before decrementing triggered_count because we cannot roll back in case an error occurs while decrementing one of the rstc. * Reworded the new call's description. [0] https://lore.kernel.org/lkml/20201001132758.12280-1-aouledameur@baylibre.com drivers/reset/core.c | 73 +++++++++++++++++++++++++++++++++++++++++++ include/linux/reset.h | 1 + 2 files changed, 74 insertions(+)