PCI: also apply D3 delay when leaving D3cold
PATCH 1/1 PCI: also apply D3 delay when leaving D3cold
This delay is needed to fix resume from s2idle of the XHCI controller on AMD Ryzen SoCs, where a 20ms delay is required (this will be quirked in a followup patch), to avoid this failure:
xhci_hcd 0000:03:00.4: WARN: xHC restore state timeout xhci_hcd 0000:03:00.4: PCI post-resume error -110!
The D3 delay is already being performed in a runtime resume from D3cold, through the following sequence of events:
pci_pm_runtime_resume
-> pci_restore_standard_config
-> pci_set_power_state(D0)
-> __pci_start_power_transition
-> pci_platform_power_transition
-> pci_update_current_state
At this point, the device has been set to D0 at the platform level, so pci_update_current_state() reads pmcsr and updates dev->current_state to D3hot. Now when we reach pci_raw_set_power_state() the D3 delay will be applied.
However, the D3cold resume from s2idle path is somewhat different, and we arrive at the same function without hitting pci_update_current_state() along the way:
pci_pm_resume_noirq
-> pci_pm_default_resume_early
-> pci_power_up
-> pci_raw_set_power_state
As dev->current_state is D3cold, the D3 delay is skipped and the XHCI controllers fail to be powered up.
Apply the D3 delay in the s2idle resume case too, in order to fix USB functionality after resume.
--- drivers/pci/pci.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/pci/pci.c +2 -1
@@ -883,7 +883,8 @@ static int pci_raw_set_power_state(struct pci_dev *dev, pci_power_t state)* Mandatory power management transition delays; see PCI PM 1.1 * 5.6.1 table 18 */ -if (state == PCI_D3hot || dev->current_state == PCI_D3hot) +if (state == PCI_D3hot || dev->current_state == PCI_D3hot + || dev->current_state == PCI_D3cold)pci_dev_d3_sleep (dev);else if (state ==PCI_D2 || dev->current_state ==PCI_D2 )udelay (PCI_PM_D2_DELAY );
Signature · 1 line
2.20.1
Quoted 48 lines · CEST Daniel Drake
This delay is needed to fix resume from s2idle of the XHCI controller on AMD Ryzen SoCs, where a 20ms delay is required (this will be quirked in a followup patch), to avoid this failure:
xhci_hcd0000:03:00.4: WARN: xHC restore state timeoutxhci_hcd0000:03:00.4: PCI post-resume error -110!The D3 delay is already being performed in a runtime resume from D3cold, through the following sequence of events:
pci_pm_runtime_resume
->pci_restore_standard_config
->pci_set_power_state(D0)
->__pci_start_power_transition
->pci_platform_power_transition
->pci_update_current_stateAt this point, the device has been set to D0 at the platform level, so
pci_update_current_state()reads pmcsr and updatesdev->current_stateto D3hot. Now when we reachpci_raw_set_power_state()the D3 delay will be applied.However, the D3cold resume from s2idle path is somewhat different, and we arrive at the same function without hitting
pci_update_current_state()along the way:
pci_pm_resume_noirq
->pci_pm_default_resume_early
->pci_power_up
->pci_raw_set_power_stateAs
dev->current_stateis D3cold, the D3 delay is skipped and the XHCI controllers fail to be powered up.Apply the D3 delay in the s2idle resume case too, in order to fix USB functionality after resume.
Link http://lkml.kernel.org/r/CAD8Lp47Vh69gQjROYG69=waJgL7hs1PwnLonL9+27S_TcRhixA@mail.gmail.com open hereSigned-off-by Daniel Drake <[email protected]>--- drivers/pci/pci.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)drivers/pci/pci.c +0
-0
drivers/pci/pci.c static int pci_raw_set_power_state(struct pci_dev *dev, pci_power_t state) · line 890
Quoted 9 lines
drivers/pci/pci.c +2
-1@@ -883,7 +883,8 @@ static int pci_raw_set_power_state(struct pci_dev *dev, pci_power_t state)* Mandatory power management transition delays; see PCI PM 1.1 * 5.6.1 table 18 */ -if (state == PCI_D3hot || dev->current_state == PCI_D3hot) +if (state == PCI_D3hot || dev->current_state == PCI_D3hot + || dev->current_state == PCI_D3cold)pci_dev_d3_sleep (dev);else if (state ==PCI_D2 || dev->current_state ==PCI_D2 )udelay (PCI_PM_D2_DELAY );
So I think that we can use pci_restore_standard_config() in the system resume patch too, which should address the issue as well.
Basically, there is no reason for the PM-runtime and system-wide resume code paths to be different in that respect.
Something like this (untested):
--- drivers/pci/pci-driver.c | 11 ++--------- drivers/pci/pci.c | 13 ------------- drivers/pci/pci.h | 1 - 3 files changed, 2 insertions(+), 23 deletions(-)
drivers/pci/pci-driver.c +2 -9
@@ -521,13 +521,6 @@ static int pci_restore_standard_config(s#ifdef CONFIG_PM_SLEEP -static void pci_pm_default_resume_early (struct pci_dev *pci_dev) -{ -pci_power_up (pci_dev); -pci_restore_state (pci_dev); -pci_pme_restore (pci_dev); -} -/* * Default "suspend" method for devices that have no driver provided suspend, * or not even a driver at all (second part).
@@ -938,7 +931,7 @@ static int pci_pm_resume_noirq(struct de* pointless, so avoid doing that. */ if (!(pci_dev->skip_bus_pm &&pm_suspend_no_platform ())) -pci_pm_default_resume_early (pci_dev); +pci_restore_standard_config (pci_dev);pci_fixup_device (pci_fixup_resume_early, pci_dev);
@@ -1213,7 +1206,7 @@ static int pci_pm_restore_noirq(struct dreturn error; } -pci_pm_default_resume_early (pci_dev); +pci_restore_standard_config (pci_dev);pci_fixup_device (pci_fixup_resume_early, pci_dev);if (pci_has_legacy_pm_support (pci_dev))
drivers/pci/pci.c +0 -13
@@ -959,19 +959,6 @@ void pci_refresh_power_state(struct pci_ }/** -* pci_power_up - Put the given device into D0 forcibly -* @dev: PCI device to power up -*/ -void pci_power_up (struct pci_dev *dev) -{ -if (platform_pci_power_manageable (dev)) -platform_pci_set_power_state (dev,PCI_D0 ); - -pci_raw_set_power_state (dev,PCI_D0 ); -pci_update_current_state (dev,PCI_D0 ); -} - -/** * pci_platform_power_transition - Use platform to change device power state * @dev: PCI device to handle. * @state: State to put the device into.
drivers/pci/pci.h +0 -1
@@ -85,7 +85,6 @@ struct pci_platform_pm_ops {int pci_set_platform_pm (const struct pci_platform_pm_ops *ops);void pci_update_current_state (struct pci_dev *dev,pci_power_t state);void pci_refresh_power_state (struct pci_dev *dev); -void pci_power_up (struct pci_dev *dev);void pci_disable_enabled_device (struct pci_dev *dev);int pci_finish_runtime_suspend (struct pci_dev *dev);void pcie_clear_root_pme_status (struct pci_dev *dev);
Quoted 5 lines · Rafael J. Wysocki
So I think that we can use
pci_restore_standard_config()in the system resume patch too, which should address the issue as well.Basically, there is no reason for the PM-runtime and system-wide resume code paths to be different in that respect.
Your patch works without modification when combined with
https://patchwork.kernel.org/patch/11187815/
Can you push this directly or would it be helpful if I update the commit message and submit it by email?
Thanks!
Daniel
Quoted 9 lines · CEST Daniel Drake
Quoted 5 lines · Rafael J. Wysocki
So I think that we can use
pci_restore_standard_config()in the system resume patch too, which should address the issue as well.Basically, there is no reason for the PM-runtime and system-wide resume code paths to be different in that respect.
Your patch works without modification when combined with
https://patchwork.kernel.org/patch/11187815/
Great, thanks for the confirmation!
Quoted 2 lines
Can you push this directly or would it be helpful if I update the commit message and submit it by email?
It would be prudent if I submitted it properly with a changelog etc.
However, I noticed that it might cause the ACPI power state to get out of sync during resume from suspend-to-RAM in some (arguably theoretical only) cases.
Specifically, that may happen if the kernel puts a device into D3hot on suspend and the platform firmware powers it up during system wakeup, because in that case the pci_update_current_state() in pci_restore_standard_config() will notice D0 and the pci_set_power_state() in there will not be called, so the ACPI layer will still hold on to the stale D3hot power state value.
While I still think that both the system resume and runtime resume code paths should be as similar as reasonably possible, the above needs to be taken into account IMO, so it is better to retain pci_pm_default_resume_early(), but make it do a conditional "ACPI power state refresh" and then call pci_restore_standard_config().
So something like the patch below (can you please test it too?).
--- drivers/pci/pci-driver.c | 8 +++++--- drivers/pci/pci.c | 15 --------------- drivers/pci/pci.h | 1 - 3 files changed, 5 insertions(+), 19 deletions(-)
drivers/pci/pci-driver.c +5 -3
@@ -523,9 +523,10 @@ static int pci_restore_standard_config(sstatic void pci_pm_default_resume_early (struct pci_dev *pci_dev) { -pci_power_up (pci_dev); -pci_restore_state (pci_dev); -pci_pme_restore (pci_dev); +if (pm_resume_via_firmware ()) +pci_refresh_power_state (pci_dev); + +pci_restore_standard_config (pci_dev); }/*
@@ -713,6 +714,7 @@ static void pci_pm_complete(struct devicpci_power_t pre_sleep_state = pci_dev->current_state;pci_refresh_power_state (pci_dev); +pci_update_current_state (pci_dev, pci_dev->current_state);/* * On platforms with ACPI this check may also trigger for * devices sharing power resources if one of those power
drivers/pci/pci.c +0 -15
@@ -954,21 +954,6 @@ void pci_refresh_power_state(struct pci_ {if (platform_pci_power_manageable (dev))platform_pci_refresh_power_state (dev); - -pci_update_current_state (dev, dev->current_state); -} - -/** -* pci_power_up - Put the given device into D0 forcibly -* @dev: PCI device to power up -*/ -void pci_power_up (struct pci_dev *dev) -{ -if (platform_pci_power_manageable (dev)) -platform_pci_set_power_state (dev,PCI_D0 ); - -pci_raw_set_power_state (dev,PCI_D0 ); -pci_update_current_state (dev,PCI_D0 ); }/**
drivers/pci/pci.h +0 -1
@@ -85,7 +85,6 @@ struct pci_platform_pm_ops {int pci_set_platform_pm (const struct pci_platform_pm_ops *ops);void pci_update_current_state (struct pci_dev *dev,pci_power_t state);void pci_refresh_power_state (struct pci_dev *dev); -void pci_power_up (struct pci_dev *dev);void pci_disable_enabled_device (struct pci_dev *dev);int pci_finish_runtime_suspend (struct pci_dev *dev);void pcie_clear_root_pme_status (struct pci_dev *dev);
Quoted 7 lines · Rafael J. Wysocki
While I still think that both the system resume and runtime resume code paths should be as similar as reasonably possible, the above needs to be taken into account IMO, so it is better to retain
pci_pm_default_resume_early(), but make it do a conditional "ACPI power state refresh" and then callpci_restore_standard_config().So something like the patch below (can you please test it too?).
This is also working fine, thanks for your help!
Quoted 11 lines · Daniel Drake
Quoted 7 lines · Rafael J. Wysocki
While I still think that both the system resume and runtime resume code paths should be as similar as reasonably possible, the above needs to be taken into account IMO, so it is better to retain
pci_pm_default_resume_early(), but make it do a conditional "ACPI power state refresh" and then callpci_restore_standard_config().So something like the patch below (can you please test it too?).
This is also working fine, thanks for your help!
Thanks for the confirmation and let me submit the patch properly.
PATCH 1/1 PCI: PM: Consolidate runtime resume and system resume paths
In reply to Daniel DrakeFrom: Rafael J. Wysocki <[email protected]>
There is an arbitrary difference between the system resume and runtime resume code paths for PCI devices regarding the delay to apply when switching the devices from D3cold to D0.
Namely, pci_restore_standard_config() used in the runtime resume code path calls pci_set_power_state() which in turn invokes __pci_start_power_transition() to power up the device through the platform firmware and that function applies the transition delay (as per PCI Express Base Specification Revision 2.0, Section 6.6.1). However, pci_pm_default_resume_early() used in the system resume code path calls pci_power_up() which doesn't apply the delay at all and that causes issues to occur during resume from suspend-to-idle on some systems where the delay is required.
Since there is no reason for that difference to exist, modify pci_pm_default_resume_early() to invoke pci_restore_standard_config() instead of pci_power_up() and drop the latter, but in order to prevent the ACPI power state values (cached by the ACPI layer) from becoming stale in some cases during resume from suspend-to-RAM (ACPI S3), as per commit cc2893b6af52 ("PCI: Ensure we re-enable devices on resume"), refresh the ACPI power state information in pci_pm_default_resume_early() in that case.
[Note that while this change should take the issue originally
addressed by commit cc2893b6af52 ("PCI: Ensure we re-enable devices on resume") into account in a generally safer way, an alternative would be to make pci_power_up() use __pci_start_power_transition() instead of calling platform_pci_set_power_state() directly.]
--- drivers/pci/pci-driver.c | 8 +++++--- drivers/pci/pci.c | 15 --------------- drivers/pci/pci.h | 1 - 3 files changed, 5 insertions(+), 19 deletions(-)
drivers/pci/pci-driver.c +5 -3
@@ -523,9 +523,10 @@ static int pci_restore_standard_config(sstatic void pci_pm_default_resume_early (struct pci_dev *pci_dev) { -pci_power_up (pci_dev); -pci_restore_state (pci_dev); -pci_pme_restore (pci_dev); +if (pm_resume_via_firmware ()) +pci_refresh_power_state (pci_dev); + +pci_restore_standard_config (pci_dev); }/*
@@ -713,6 +714,7 @@ static void pci_pm_complete(struct devicpci_power_t pre_sleep_state = pci_dev->current_state;pci_refresh_power_state (pci_dev); +pci_update_current_state (pci_dev, pci_dev->current_state);/* * On platforms with ACPI this check may also trigger for * devices sharing power resources if one of those power
drivers/pci/pci.c +0 -15
@@ -954,21 +954,6 @@ void pci_refresh_power_state(struct pci_ {if (platform_pci_power_manageable (dev))platform_pci_refresh_power_state (dev); - -pci_update_current_state (dev, dev->current_state); -} - -/** -* pci_power_up - Put the given device into D0 forcibly -* @dev: PCI device to power up -*/ -void pci_power_up (struct pci_dev *dev) -{ -if (platform_pci_power_manageable (dev)) -platform_pci_set_power_state (dev,PCI_D0 ); - -pci_raw_set_power_state (dev,PCI_D0 ); -pci_update_current_state (dev,PCI_D0 ); }/**
drivers/pci/pci.h +0 -1
@@ -85,7 +85,6 @@ struct pci_platform_pm_ops {int pci_set_platform_pm (const struct pci_platform_pm_ops *ops);void pci_update_current_state (struct pci_dev *dev,pci_power_t state);void pci_refresh_power_state (struct pci_dev *dev); -void pci_power_up (struct pci_dev *dev);void pci_disable_enabled_device (struct pci_dev *dev);int pci_finish_runtime_suspend (struct pci_dev *dev);void pcie_clear_root_pme_status (struct pci_dev *dev);
Re: [PATCH] PCI: PM: Consolidate runtime resume and system resume paths
Quoted 47 lines · CEST Rafael J. Wysocki
From: Rafael J. Wysocki <[email protected]>
There is an arbitrary difference between the system resume and runtime resume code paths for PCI devices regarding the delay to apply when switching the devices from D3cold to D0.
Namely,
pci_restore_standard_config()used in the runtime resume code path callspci_set_power_state()which in turn invokes__pci_start_power_transition()to power up the device through the platform firmware and that function applies the transition delay (as per PCI Express Base Specification Revision 2.0, Section 6.6.1). However,pci_pm_default_resume_early()used in the system resume code path callspci_power_up()which doesn't apply the delay at all and that causes issues to occur during resume from suspend-to-idle on some systems where the delay is required.Since there is no reason for that difference to exist, modify
pci_pm_default_resume_early()to invokepci_restore_standard_config()instead ofpci_power_up()and drop the latter, but in order to prevent the ACPI power state values (cached by the ACPI layer) from becoming stale in some cases during resume from suspend-to-RAM (ACPI S3), as per commitcc2893b6af52("PCI: Ensure we re-enable devices on resume"), refresh the ACPI power state information inpci_pm_default_resume_early()in that case.[Note that while this change should take the issue originally
addressed by commitcc2893b6af52("PCI: Ensure we re-enable devices on resume") into account in a generally safer way, an alternative would be to makepci_power_up()use__pci_start_power_transition()instead of callingplatform_pci_set_power_state()directly.]Fixes db288c9c5f9d ("PCI / PM: restore the original behavior of pci_set_power_state()")Reported-by Daniel Drake <[email protected]>Tested-by Daniel Drake <[email protected]>Link https://lore.kernel.org/linux-pm/CAD8Lp44TYxrMgPLkHCqF9hv6smEurMXvmmvmtyFhZ6Q4SE+dig@mail.gmail.com/T/#m21be74af263c6a34f36e0fc5c77c5449d9406925 open hereSigned-off-by Rafael J. Wysocki <[email protected]>--- drivers/pci/pci-driver.c | 8 +++++--- drivers/pci/pci.c | 15 --------------- drivers/pci/pci.h | 1 - 3 files changed, 5 insertions(+), 19 deletions(-)Index:
linux-pm/drivers/pci/pci-driver.c
===================================================================
drivers/pci/pci-driver.c static int pci_restore_standard_config(s · line 527
Quoted 8 lines
drivers/pci/pci-driver.c +2
-3@@ -523,9 +523,10 @@ static int pci_restore_standard_config(sstatic void pci_pm_default_resume_early (struct pci_dev *pci_dev) { -pci_power_up (pci_dev); -pci_restore_state (pci_dev); -pci_pme_restore (pci_dev); +if (pm_resume_via_firmware ()) +pci_refresh_power_state (pci_dev);
Well, this is still not going to work if the ACPI power state after the update above is not D0, but the pci_update_current_state() in pci_restore_standard_config() returns D0, which was the case that triggered commit cc2893b6af52 IIRC.
So scratch this one, please, and I'll submit the safer option.
Quoted 13 lines
drivers/pci/pci-driver.c +3
-0+ +pci_restore_standard_config (pci_dev); }/* @@ -713,6 +714,7 @@ static void pci_pm_complete(struct devicpci_power_t pre_sleep_state = pci_dev->current_state;pci_refresh_power_state (pci_dev); +pci_update_current_state (pci_dev, pci_dev->current_state);/* * On platforms with ACPI this check may also trigger for * devices sharing power resources if one of those power
PATCH 1/1 PCI: PM: Fix pci_power_up()
In reply to Rafael J. WysockiFrom: Rafael J. Wysocki <[email protected]>
There is an arbitrary difference between the system resume and runtime resume code paths for PCI devices regarding the delay to apply when switching the devices from D3cold to D0.
Namely, pci_restore_standard_config() used in the runtime resume code path calls pci_set_power_state() which in turn invokes __pci_start_power_transition() to power up the device through the platform firmware and that function applies the transition delay (as per PCI Express Base Specification Revision 2.0, Section 6.6.1). However, pci_pm_default_resume_early() used in the system resume code path calls pci_power_up() which doesn't apply the delay at all and that causes issues to occur during resume from suspend-to-idle on some systems where the delay is required.
Since there is no reason for that difference to exist, modify pci_power_up() to follow pci_set_power_state() more closely and invoke __pci_start_power_transition() from there to call the platform firmware to power up the device (in case that's necessary).
---
Daniel, please test this one.
--- drivers/pci/pci.c | 24 +++++++++++------------- 1 file changed, 11 insertions(+), 13 deletions(-)
drivers/pci/pci.c +11 -13
@@ -959,19 +959,6 @@ void pci_refresh_power_state(struct pci_ }/** -* pci_power_up - Put the given device into D0 forcibly -* @dev: PCI device to power up -*/ -void pci_power_up (struct pci_dev *dev) -{ -if (platform_pci_power_manageable (dev)) -platform_pci_set_power_state (dev,PCI_D0 ); - -pci_raw_set_power_state (dev,PCI_D0 ); -pci_update_current_state (dev,PCI_D0 ); -} - -/** * pci_platform_power_transition - Use platform to change device power state * @dev: PCI device to handle. * @state: State to put the device into.
@@ -1154,6 +1141,17 @@ int pci_set_power_state(struct pci_dev *EXPORT_SYMBOL (pci_set_power_state);/** +* pci_power_up - Put the given device into D0 forcibly +* @dev: PCI device to power up +*/ +void pci_power_up (struct pci_dev *dev) +{ +__pci_start_power_transition (dev,PCI_D0 ); +pci_raw_set_power_state (dev,PCI_D0 ); +pci_update_current_state (dev,PCI_D0 ); +} + +/** * pci_choose_state - Choose the power state of a PCI device * @dev: PCI device to be suspended * @state: target sleep state for the whole system. This is the value
Re: [PATCH] PCI: PM: Fix pci_power_up()
Quoted 5 lines · Rafael J. Wysocki
Since there is no reason for that difference to exist, modify
pci_power_up()to followpci_set_power_state()more closely and invoke__pci_start_power_transition()from there to call the platform firmware to power up the device (in case that's necessary).
Quoted 7 lines
Fixes db288c9c5f9d ("PCI / PM: restore the original behavior of pci_set_power_state()")Reported-by Daniel Drake <[email protected]>Link https://lore.kernel.org/linux-pm/CAD8Lp44TYxrMgPLkHCqF9hv6smEurMXvmmvmtyFhZ6Q4SE+dig@mail.gmail.com/T/#m21be74af263c6a34f36e0fc5c77c5449d9406925 open hereSigned-off-by Rafael J. Wysocki <[email protected]>---
Daniel, please test this one.
This one is working too, thanks
Daniel
Re: [PATCH] PCI: PM: Fix pci_power_up()
Quoted 13 lines · Daniel Drake
Quoted 11 lines · Rafael J. Wysocki
Since there is no reason for that difference to exist, modify
pci_power_up()to followpci_set_power_state()more closely and invoke__pci_start_power_transition()from there to call the platform firmware to power up the device (in case that's necessary).Fixes db288c9c5f9d ("PCI / PM: restore the original behavior of pci_set_power_state()")Reported-by Daniel Drake <[email protected]>Link https://lore.kernel.org/linux-pm/CAD8Lp44TYxrMgPLkHCqF9hv6smEurMXvmmvmtyFhZ6Q4SE+dig@mail.gmail.com/T/#m21be74af263c6a34f36e0fc5c77c5449d9406925 open hereSigned-off-by Rafael J. Wysocki <[email protected]>---
Quoted 3 lines
Quoted 1 line
Daniel, please test this one.
This one is working too, thanks
Thank you!
Bjorn, any concerns?
Re: [PATCH] PCI: PM: Fix pci_power_up()
In reply to Rafael J. WysockiQuoted 58 lines · Rafael J. Wysocki
From: Rafael J. Wysocki <[email protected]>
There is an arbitrary difference between the system resume and runtime resume code paths for PCI devices regarding the delay to apply when switching the devices from D3cold to D0.
Namely,
pci_restore_standard_config()used in the runtime resume code path callspci_set_power_state()which in turn invokes__pci_start_power_transition()to power up the device through the platform firmware and that function applies the transition delay (as per PCI Express Base Specification Revision 2.0, Section 6.6.1). However,pci_pm_default_resume_early()used in the system resume code path callspci_power_up()which doesn't apply the delay at all and that causes issues to occur during resume from suspend-to-idle on some systems where the delay is required.Since there is no reason for that difference to exist, modify
pci_power_up()to followpci_set_power_state()more closely and invoke__pci_start_power_transition()from there to call the platform firmware to power up the device (in case that's necessary).Fixes db288c9c5f9d ("PCI / PM: restore the original behavior of pci_set_power_state()")Reported-by Daniel Drake <[email protected]>Link https://lore.kernel.org/linux-pm/CAD8Lp44TYxrMgPLkHCqF9hv6smEurMXvmmvmtyFhZ6Q4SE+dig@mail.gmail.com/T/#m21be74af263c6a34f36e0fc5c77c5449d9406925 open hereSigned-off-by Rafael J. Wysocki <[email protected]>---
Daniel, please test this one.
--- drivers/pci/pci.c | 24 +++++++++++------------- 1 file changed, 11 insertions(+), 13 deletions(-)Index:
linux-pm/drivers/pci/pci.c
===================================================================drivers/pci/pci.c +0
-13@@ -959,19 +959,6 @@ void pci_refresh_power_state(struct pci_ }/** -* pci_power_up - Put the given device into D0 forcibly -* @dev: PCI device to power up -*/ -void pci_power_up (struct pci_dev *dev) -{ -if (platform_pci_power_manageable (dev)) -platform_pci_set_power_state (dev,PCI_D0 ); - -pci_raw_set_power_state (dev,PCI_D0 ); -pci_update_current_state (dev,PCI_D0 ); -} - -/** * pci_platform_power_transition - Use platform to change device power state * @dev: PCI device to handle. * @state: State to put the device into.
drivers/pci/pci.c int pci_set_power_state(struct pci_dev * · line 1144
Quoted 4 lines
drivers/pci/pci.c +1
-0@@ -1154,6 +1141,17 @@ int pci_set_power_state(struct pci_dev *EXPORT_SYMBOL (pci_set_power_state);/** +* pci_power_up - Put the given device into D0 forcibly
Not specifically for this patch, but what does "forcibly" mean?
drivers/pci/pci.c int pci_set_power_state(struct pci_dev * · line 1151
Quoted 7 lines
drivers/pci/pci.c +7
-0+* @dev: PCI device to power up +*/ +void pci_power_up (struct pci_dev *dev) +{ +__pci_start_power_transition (dev,PCI_D0 ); +pci_raw_set_power_state (dev,PCI_D0 ); +pci_update_current_state (dev,PCI_D0 );
There's not very much difference between:
pci_power_up(dev);
and
pci_set_power_state(dev, PCI_D0);
It looks like the main difference is that pci_set_power_state() calls __pci_complete_power_transition(), which ultimately calls acpi_pci_set_power_state() (for ACPI systems).
So maybe "forcibly" means something like "ignoring any platform power management methods"? It's not obvious to me when we should skip the platform stuff or whether the skipping should be done at the high level (like calling either pci_power_up() or pci_set_power_state()) or at a lower level (e.g., if everybody called pci_set_power_state() and it could internally tell whether we're skipping the platform part).
If we could unify the paths as much as possible, that would be nice, but if it's not feasible, it's not feasible. If you'd like me to push this for v5.4, let me know, otherwise you can apply my:
Quoted 9 lines
drivers/pci/pci.c +3
-0+} + +/** * pci_choose_state - Choose the power state of a PCI device * @dev: PCI device to be suspended * @state: target sleep state for the whole system. This is the value
Re: [PATCH] PCI: PM: Fix pci_power_up()
Quoted 79 lines · Bjorn Helgaas
Quoted 58 lines · Rafael J. Wysocki
From: Rafael J. Wysocki <[email protected]>
There is an arbitrary difference between the system resume and runtime resume code paths for PCI devices regarding the delay to apply when switching the devices from D3cold to D0.
Namely,
pci_restore_standard_config()used in the runtime resume code path callspci_set_power_state()which in turn invokes__pci_start_power_transition()to power up the device through the platform firmware and that function applies the transition delay (as per PCI Express Base Specification Revision 2.0, Section 6.6.1). However,pci_pm_default_resume_early()used in the system resume code path callspci_power_up()which doesn't apply the delay at all and that causes issues to occur during resume from suspend-to-idle on some systems where the delay is required.Since there is no reason for that difference to exist, modify
pci_power_up()to followpci_set_power_state()more closely and invoke__pci_start_power_transition()from there to call the platform firmware to power up the device (in case that's necessary).Fixes db288c9c5f9d ("PCI / PM: restore the original behavior of pci_set_power_state()")Reported-by Daniel Drake <[email protected]>Link https://lore.kernel.org/linux-pm/CAD8Lp44TYxrMgPLkHCqF9hv6smEurMXvmmvmtyFhZ6Q4SE+dig@mail.gmail.com/T/#m21be74af263c6a34f36e0fc5c77c5449d9406925 open hereSigned-off-by Rafael J. Wysocki <[email protected]>---
Daniel, please test this one.
--- drivers/pci/pci.c | 24 +++++++++++------------- 1 file changed, 11 insertions(+), 13 deletions(-)Index:
linux-pm/drivers/pci/pci.c
===================================================================drivers/pci/pci.c +0
-13@@ -959,19 +959,6 @@ void pci_refresh_power_state(struct pci_ }/** -* pci_power_up - Put the given device into D0 forcibly -* @dev: PCI device to power up -*/ -void pci_power_up (struct pci_dev *dev) -{ -if (platform_pci_power_manageable (dev)) -platform_pci_set_power_state (dev,PCI_D0 ); - -pci_raw_set_power_state (dev,PCI_D0 ); -pci_update_current_state (dev,PCI_D0 ); -} - -/** * pci_platform_power_transition - Use platform to change device power state * @dev: PCI device to handle. * @state: State to put the device into. drivers/pci/pci.c int pci_set_power_state(struct pci_dev * · line 1144
Quoted 4 lines
drivers/pci/pci.c +1
-0@@ -1154,6 +1141,17 @@ int pci_set_power_state(struct pci_dev *EXPORT_SYMBOL (pci_set_power_state);/** +* pci_power_up - Put the given device into D0 forcibly Not specifically for this patch, but what does "forcibly" mean?
drivers/pci/pci.c int pci_set_power_state(struct pci_dev * · line 1151
Quoted 7 lines
drivers/pci/pci.c +7
-0+* @dev: PCI device to power up +*/ +void pci_power_up (struct pci_dev *dev) +{ +__pci_start_power_transition (dev,PCI_D0 ); +pci_raw_set_power_state (dev,PCI_D0 ); +pci_update_current_state (dev,PCI_D0 );There's not very much difference between:
pci_power_up(dev);
Quoted 7 lines
and
pci_set_power_state(dev, PCI_D0);It looks like the main difference is that
pci_set_power_state()calls__pci_complete_power_transition(), which ultimately callsacpi_pci_set_power_state()(for ACPI systems).
Yes, it does, for power states deeper than D0, which is not the case here.
The main difference is the dev->current_state == state check in pci_set_power_state(), but in the resume case specifically dev->current_state == PCI_D0 doesn't matter, because the real power state of the device may be different.
Quoted 2 lines
So maybe "forcibly" means something like "ignoring any platform power management methods"?
It means "go into D0 no matter what the current cached value is".
Quoted 5 lines
It's not obvious to me when we should skip the
platform stuff or whether the skipping should be done at the high level (like calling eitherpci_power_up()orpci_set_power_state()) or at a lower level (e.g., if everybody calledpci_set_power_state()and it could internally tell whether we're skipping the platform part).
For transitions into D0 __pci_start_power_transition() is the platform stuff, so we don't skip it and the other things that are present in pci_set_power_state() and are not there in pci_power_up() are simply unnecessary for transitions to D0.
Quoted 2 lines
If we could unify the paths as much as possible, that would be nice, but if it's not feasible, it's not feasible.
It kind of is, but I'd prefer to do it on top of this patch.
First, the pci_update_current_state() in pci_power_up() can be moved to pci_pm_default_resume_early() which is the only caller of pci_power_up(). [The role of that pci_update_current_state() is to change the current_state value to D3cold if the device is not accessible (or the platform firmware says that it is D3cold, which may be the case after a failing attempt to use it to switch the device over to D0).] Next, if pci_power_up() is modified to return the return value of pci_raw_set_power_state(), pci_set_power_state() can be implemented (roughly) as
sanitize the state argument
if (dev->current_state == state)
return 0;
if (state == PCI_D0)
return pci_power_up();
carry out a transition into a deeper power state.
And so pci_power_up() will be used by pci_set_power_state(), for transitions into D0, and (directly) by pci_pm_default_resume_early().
How does that sound?
Quoted 4 lines
If you'd like me to push this for v5.4, let me know, otherwise you can apply my:
Acked-by Bjorn Helgaas <[email protected]>
I will, thanks!