* [PATCH 0/1] NET: mcan: Move runtime PM enable/disable to m_can_platform
@ 2020-08-25 10:17 Patrik Flykt
2020-08-25 10:17 ` [PATCH 1/1] " Patrik Flykt
0 siblings, 1 reply; 6+ messages in thread
From: Patrik Flykt @ 2020-08-25 10:17 UTC (permalink / raw)
To: linux-can; +Cc: wg, mkl, dmurphy, sriram.dash, jarkko.nikula
Hi,
This is a preparatory patch for modifying PM enabling due to PCI based
M_CAN devices. As the functionality is right now, adding a PCI based
M_CAN driver causes the core PCI functionality to call
pm_runtime_enable() from pci_pm_init() when the PCI device is added.
When the device is added, it is registered with m_can_class_register(),
which ends up calling pm_runtime_enable() once more and causes the
kernel to log an angry 'Unbalanced pm_runtime_enable!' message, as
dev->power.disable_depth has gone down to zero.
To resolve this situation, I have added a patch that moves runtime PM
enabling from m_can.c to the m_can_platform.c driver, which currently
is the only driver that ends up enabling runtime PM. Would this
approach be appropriate, or should PM be enabled in some other way
with PCI based M_CAN devices?
Now, the actual sticky point is that there aren't yet any PCI based
M_CAN devices upstream, but one is in the works. So if the change
proposed works out, should this patch actually be sent in the patch
set providing a PCI based M_CAN device?
Thanks,
Patrik
Patrik Flykt (1):
NET: mcan: Move runtime PM enable/disable to m_can_platform
drivers/net/can/m_can/m_can.c | 6 +-----
drivers/net/can/m_can/m_can_platform.c | 3 +++
2 files changed, 4 insertions(+), 5 deletions(-)
--
2.27.0
^ permalink raw reply [flat|nested] 6+ messages in thread
* [PATCH 1/1] NET: mcan: Move runtime PM enable/disable to m_can_platform
2020-08-25 10:17 [PATCH 0/1] NET: mcan: Move runtime PM enable/disable to m_can_platform Patrik Flykt
@ 2020-08-25 10:17 ` Patrik Flykt
2020-09-01 18:22 ` Dan Murphy
2020-10-19 15:47 ` Marc Kleine-Budde
0 siblings, 2 replies; 6+ messages in thread
From: Patrik Flykt @ 2020-08-25 10:17 UTC (permalink / raw)
To: linux-can; +Cc: wg, mkl, dmurphy, sriram.dash, jarkko.nikula
This is a preparatory patch for upcoming PCI based M_CAN devices.
The current PM implementation would cause PCI based drivers to
enable PM twice, once when the pci device is added and a second
time in m_can_class_register(). This will cause 'Unbalanced
pm_runtime_enable!' to be logged, and is a situation that should
be avoided.
Therefore, in anticipation of PCI devices, move PM enabling out
from M_CAN class registration to its currently only user, the
m_can_platform driver.
Signed-off-by: Patrik Flykt <patrik.flykt@linux.intel.com>
---
drivers/net/can/m_can/m_can.c | 6 +-----
drivers/net/can/m_can/m_can_platform.c | 3 +++
2 files changed, 4 insertions(+), 5 deletions(-)
diff --git a/drivers/net/can/m_can/m_can.c b/drivers/net/can/m_can/m_can.c
index 02c5795b7393..2c4d74113443 100644
--- a/drivers/net/can/m_can/m_can.c
+++ b/drivers/net/can/m_can/m_can.c
@@ -1817,7 +1817,6 @@ int m_can_class_register(struct m_can_classdev *m_can_dev)
int ret;
if (m_can_dev->pm_clock_support) {
- pm_runtime_enable(m_can_dev->dev);
ret = m_can_clk_start(m_can_dev);
if (ret)
goto pm_runtime_fail;
@@ -1847,11 +1846,8 @@ int m_can_class_register(struct m_can_classdev *m_can_dev)
clk_disable:
m_can_clk_stop(m_can_dev);
pm_runtime_fail:
- if (ret) {
- if (m_can_dev->pm_clock_support)
- pm_runtime_disable(m_can_dev->dev);
+ if (ret)
free_candev(m_can_dev->net);
- }
return ret;
}
diff --git a/drivers/net/can/m_can/m_can_platform.c b/drivers/net/can/m_can/m_can_platform.c
index 38ea5e600fb8..1260e99b9322 100644
--- a/drivers/net/can/m_can/m_can_platform.c
+++ b/drivers/net/can/m_can/m_can_platform.c
@@ -111,7 +111,10 @@ static int m_can_plat_probe(struct platform_device *pdev)
m_can_init_ram(mcan_class);
+ pm_runtime_enable(mcan_class->dev);
ret = m_can_class_register(mcan_class);
+ if (ret)
+ pm_runtime_disable(mcan_class->dev);
failed_ret:
return ret;
--
2.27.0
^ permalink raw reply related [flat|nested] 6+ messages in thread* Re: [PATCH 1/1] NET: mcan: Move runtime PM enable/disable to m_can_platform
2020-08-25 10:17 ` [PATCH 1/1] " Patrik Flykt
@ 2020-09-01 18:22 ` Dan Murphy
2020-10-19 15:01 ` Patrik Flykt
2020-10-19 15:47 ` Marc Kleine-Budde
1 sibling, 1 reply; 6+ messages in thread
From: Dan Murphy @ 2020-09-01 18:22 UTC (permalink / raw)
To: Patrik Flykt, linux-can; +Cc: wg, mkl, sriram.dash, jarkko.nikula
Patrik
On 8/25/20 5:17 AM, Patrik Flykt wrote:
> This is a preparatory patch for upcoming PCI based M_CAN devices.
> The current PM implementation would cause PCI based drivers to
> enable PM twice, once when the pci device is added and a second
> time in m_can_class_register(). This will cause 'Unbalanced
> pm_runtime_enable!' to be logged, and is a situation that should
> be avoided.
>
> Therefore, in anticipation of PCI devices, move PM enabling out
> from M_CAN class registration to its currently only user, the
> m_can_platform driver.
>
>
> Signed-off-by: Patrik Flykt <patrik.flykt@linux.intel.com>
> ---
> drivers/net/can/m_can/m_can.c | 6 +-----
> drivers/net/can/m_can/m_can_platform.c | 3 +++
> 2 files changed, 4 insertions(+), 5 deletions(-)
>
> diff --git a/drivers/net/can/m_can/m_can.c b/drivers/net/can/m_can/m_can.c
> index 02c5795b7393..2c4d74113443 100644
> --- a/drivers/net/can/m_can/m_can.c
> +++ b/drivers/net/can/m_can/m_can.c
> @@ -1817,7 +1817,6 @@ int m_can_class_register(struct m_can_classdev *m_can_dev)
> int ret;
>
> if (m_can_dev->pm_clock_support) {
> - pm_runtime_enable(m_can_dev->dev);
> ret = m_can_clk_start(m_can_dev);
> if (ret)
> goto pm_runtime_fail;
> @@ -1847,11 +1846,8 @@ int m_can_class_register(struct m_can_classdev *m_can_dev)
> clk_disable:
> m_can_clk_stop(m_can_dev);
> pm_runtime_fail:
> - if (ret) {
> - if (m_can_dev->pm_clock_support)
> - pm_runtime_disable(m_can_dev->dev);
> + if (ret)
> free_candev(m_can_dev->net);
> - }
>
> return ret;
> }
> diff --git a/drivers/net/can/m_can/m_can_platform.c b/drivers/net/can/m_can/m_can_platform.c
> index 38ea5e600fb8..1260e99b9322 100644
> --- a/drivers/net/can/m_can/m_can_platform.c
> +++ b/drivers/net/can/m_can/m_can_platform.c
> @@ -111,7 +111,10 @@ static int m_can_plat_probe(struct platform_device *pdev)
>
> m_can_init_ram(mcan_class);
>
> + pm_runtime_enable(mcan_class->dev);
> ret = m_can_class_register(mcan_class);
> + if (ret)
> + pm_runtime_disable(mcan_class->dev);
>
> failed_ret:
> return ret;
Reviewed-by: Dan Murphy <dmurphy@ti.com>
^ permalink raw reply [flat|nested] 6+ messages in thread* Re: [PATCH 1/1] NET: mcan: Move runtime PM enable/disable to m_can_platform
2020-09-01 18:22 ` Dan Murphy
@ 2020-10-19 15:01 ` Patrik Flykt
0 siblings, 0 replies; 6+ messages in thread
From: Patrik Flykt @ 2020-10-19 15:01 UTC (permalink / raw)
To: Dan Murphy, linux-can; +Cc: wg, mkl, sriram.dash, jarkko.nikula
On Tue, 2020-09-01 at 13:22 -0500, Dan Murphy wrote:
> Patrik
>
> On 8/25/20 5:17 AM, Patrik Flykt wrote:
> > This is a preparatory patch for upcoming PCI based M_CAN devices.
> > The current PM implementation would cause PCI based drivers to
> > enable PM twice, once when the pci device is added and a second
> > time in m_can_class_register(). This will cause 'Unbalanced
> > pm_runtime_enable!' to be logged, and is a situation that should
> > be avoided.
> >
> > Therefore, in anticipation of PCI devices, move PM enabling out
> > from M_CAN class registration to its currently only user, the
> > m_can_platform driver.
> Reviewed-by: Dan Murphy <dmurphy@ti.com>
Ping?
Can this patch be added as-is or should it be sent in the patch set
providing a PCI based M_CAN device?
Cheers,
Patrik
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH 1/1] NET: mcan: Move runtime PM enable/disable to m_can_platform
2020-08-25 10:17 ` [PATCH 1/1] " Patrik Flykt
2020-09-01 18:22 ` Dan Murphy
@ 2020-10-19 15:47 ` Marc Kleine-Budde
2020-10-23 11:58 ` Patrik Flykt
1 sibling, 1 reply; 6+ messages in thread
From: Marc Kleine-Budde @ 2020-10-19 15:47 UTC (permalink / raw)
To: Patrik Flykt, linux-can; +Cc: wg, dmurphy, sriram.dash, jarkko.nikula
[-- Attachment #1.1: Type: text/plain, Size: 2625 bytes --]
On 8/25/20 12:17 PM, Patrik Flykt wrote:
> This is a preparatory patch for upcoming PCI based M_CAN devices.
> The current PM implementation would cause PCI based drivers to
> enable PM twice, once when the pci device is added and a second
> time in m_can_class_register(). This will cause 'Unbalanced
> pm_runtime_enable!' to be logged, and is a situation that should
> be avoided.
>
> Therefore, in anticipation of PCI devices, move PM enabling out
> from M_CAN class registration to its currently only user, the
> m_can_platform driver.
Makes sense.
> Signed-off-by: Patrik Flykt <patrik.flykt@linux.intel.com>
Please rebase this patch onto linux-can/m_can
https://git.kernel.org/pub/scm/linux/kernel/git/mkl/linux-can.git/log/?h=m_can
> ---
> drivers/net/can/m_can/m_can.c | 6 +-----
> drivers/net/can/m_can/m_can_platform.c | 3 +++
> 2 files changed, 4 insertions(+), 5 deletions(-)
>
> diff --git a/drivers/net/can/m_can/m_can.c b/drivers/net/can/m_can/m_can.c
> index 02c5795b7393..2c4d74113443 100644
> --- a/drivers/net/can/m_can/m_can.c
> +++ b/drivers/net/can/m_can/m_can.c
> @@ -1817,7 +1817,6 @@ int m_can_class_register(struct m_can_classdev *m_can_dev)
> int ret;
>
> if (m_can_dev->pm_clock_support) {
I think pm_clock_support can be removed altogether now.
> - pm_runtime_enable(m_can_dev->dev);
> ret = m_can_clk_start(m_can_dev);
> if (ret)
> goto pm_runtime_fail;
> @@ -1847,11 +1846,8 @@ int m_can_class_register(struct m_can_classdev *m_can_dev)
> clk_disable:
> m_can_clk_stop(m_can_dev);
> pm_runtime_fail:
> - if (ret) {
> - if (m_can_dev->pm_clock_support)
> - pm_runtime_disable(m_can_dev->dev);
> + if (ret)
> free_candev(m_can_dev->net);
> - }
>
> return ret;
> }
> diff --git a/drivers/net/can/m_can/m_can_platform.c b/drivers/net/can/m_can/m_can_platform.c
> index 38ea5e600fb8..1260e99b9322 100644
> --- a/drivers/net/can/m_can/m_can_platform.c
> +++ b/drivers/net/can/m_can/m_can_platform.c
> @@ -111,7 +111,10 @@ static int m_can_plat_probe(struct platform_device *pdev)
>
> m_can_init_ram(mcan_class);
>
> + pm_runtime_enable(mcan_class->dev);
> ret = m_can_class_register(mcan_class);
> + if (ret)
> + pm_runtime_disable(mcan_class->dev);
>
> failed_ret:
> return ret;
>
Marc
--
Pengutronix e.K. | Marc Kleine-Budde |
Embedded Linux | https://www.pengutronix.de |
Vertretung West/Dortmund | Phone: +49-231-2826-924 |
Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-5555 |
[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 488 bytes --]
^ permalink raw reply [flat|nested] 6+ messages in thread* Re: [PATCH 1/1] NET: mcan: Move runtime PM enable/disable to m_can_platform
2020-10-19 15:47 ` Marc Kleine-Budde
@ 2020-10-23 11:58 ` Patrik Flykt
0 siblings, 0 replies; 6+ messages in thread
From: Patrik Flykt @ 2020-10-23 11:58 UTC (permalink / raw)
To: Marc Kleine-Budde, linux-can; +Cc: wg, dmurphy, sriram.dash, jarkko.nikula
On Mon, 2020-10-19 at 17:47 +0200, Marc Kleine-Budde wrote:
> > Therefore, in anticipation of PCI devices, move PM enabling out
> > from M_CAN class registration to its currently only user, the
> > m_can_platform driver.
>
> Makes sense.
>
> > Signed-off-by: Patrik Flykt <patrik.flykt@linux.intel.com>
>
> Please rebase this patch onto linux-can/m_can
>
> https://git.kernel.org/pub/scm/linux/kernel/git/mkl/linux-can.git/log/?h=m_can
Rebased and sent as v2.
Thanks!
Patrik
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2020-10-23 11:59 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2020-08-25 10:17 [PATCH 0/1] NET: mcan: Move runtime PM enable/disable to m_can_platform Patrik Flykt
2020-08-25 10:17 ` [PATCH 1/1] " Patrik Flykt
2020-09-01 18:22 ` Dan Murphy
2020-10-19 15:01 ` Patrik Flykt
2020-10-19 15:47 ` Marc Kleine-Budde
2020-10-23 11:58 ` Patrik Flykt
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox