When suspending/resuming an interface which was not up, we saw mdiobus
related PM handling despite 'mac_managed_pm' being set for RAVB/SH_ETH.
Heiner kindly suggested the fix to set this flag at probe time, not at
init/open time. I implemented his suggestion and it works fine on these
two Renesas drivers. These are patches 1+2.
I then looked which other drivers could be affected by the same problem.
I could only identify two where I am quite sure. Because I don't have
any HW, I opted to at least add a FIXME and send this out as patches
3+4. Ideally, they will never need to go upstream because the relevant
people will see their patch and do something like I did for patches 1+2.
Looking forward to comments. Thanks and happy hacking!
Wolfram Sang (4):
ravb: avoid PHY being resumed when interface is not up
sh_eth: avoid PHY being resumed when interface is not up
fec: add FIXME to move 'mac_managed_pm' to probe
smsc911x: add FIXME to move 'mac_managed_pm' to probe
drivers/net/ethernet/freescale/fec_main.c | 1 +
drivers/net/ethernet/renesas/ravb_main.c | 12 ++++++++++--
drivers/net/ethernet/renesas/sh_eth.c | 12 ++++++++++--
drivers/net/ethernet/smsc/smsc911x.c | 1 +
4 files changed, 22 insertions(+), 4 deletions(-)
--
2.30.2
On Tue, Mar 14, 2023 at 02:14:38PM +0100, Wolfram Sang wrote:
> When suspending/resuming an interface which was not up, we saw mdiobus
> related PM handling despite 'mac_managed_pm' being set for RAVB/SH_ETH.
> Heiner kindly suggested the fix to set this flag at probe time, not at
> init/open time. I implemented his suggestion and it works fine on these
> two Renesas drivers. These are patches 1+2.
>
> I then looked which other drivers could be affected by the same problem.
> I could only identify two where I am quite sure. Because I don't have
> any HW, I opted to at least add a FIXME and send this out as patches
> 3+4. Ideally, they will never need to go upstream because the relevant
> people will see their patch and do something like I did for patches 1+2.
>
> Looking forward to comments. Thanks and happy hacking!
>
>
> Wolfram Sang (4):
> ravb: avoid PHY being resumed when interface is not up
> sh_eth: avoid PHY being resumed when interface is not up
> fec: add FIXME to move 'mac_managed_pm' to probe
> smsc911x: add FIXME to move 'mac_managed_pm' to probe
>
> drivers/net/ethernet/freescale/fec_main.c | 1 +
> drivers/net/ethernet/renesas/ravb_main.c | 12 ++++++++++--
> drivers/net/ethernet/renesas/sh_eth.c | 12 ++++++++++--
> drivers/net/ethernet/smsc/smsc911x.c | 1 +
> 4 files changed, 22 insertions(+), 4 deletions(-)
>
> --
> 2.30.2
>
Unfortunately, I wasn't able to check the series in terms of content, but
I have no objections to the logic and coding style.
For the series.
Reviewed-by: Michal Kubiak <[email protected]>
Thanks,
Michal
On Tue, Mar 14, 2023 at 02:14:38PM +0100, Wolfram Sang wrote:
> When suspending/resuming an interface which was not up, we saw mdiobus
> related PM handling despite 'mac_managed_pm' being set for RAVB/SH_ETH.
> Heiner kindly suggested the fix to set this flag at probe time, not at
> init/open time. I implemented his suggestion and it works fine on these
> two Renesas drivers. These are patches 1+2.
>
> I then looked which other drivers could be affected by the same problem.
> I could only identify two where I am quite sure. Because I don't have
> any HW, I opted to at least add a FIXME and send this out as patches
> 3+4. Ideally, they will never need to go upstream because the relevant
> people will see their patch and do something like I did for patches 1+2.
>
> Looking forward to comments. Thanks and happy hacking!
Hi Wolfram,
an initial comment is, and I know this is a bit of a pain,
that 'net' and 'net-next' patches need to be split into separate serries.
And if there are dependencies, the 'net' series should go first, and then
the 'net-next' series should go out once the 'net' patches have been
accepted and 'net' has been merged into 'net-next', which I believe occurs
after Linus has pulled the weekly 'net' pull request, which typically
occurs on Thursdays.
> Wolfram Sang (4):
> ravb: avoid PHY being resumed when interface is not up
> sh_eth: avoid PHY being resumed when interface is not up
> fec: add FIXME to move 'mac_managed_pm' to probe
> smsc911x: add FIXME to move 'mac_managed_pm' to probe
>
> drivers/net/ethernet/freescale/fec_main.c | 1 +
> drivers/net/ethernet/renesas/ravb_main.c | 12 ++++++++++--
> drivers/net/ethernet/renesas/sh_eth.c | 12 ++++++++++--
> drivers/net/ethernet/smsc/smsc911x.c | 1 +
> 4 files changed, 22 insertions(+), 4 deletions(-)
>
> --
> 2.30.2
>
Hi Simon!
> an initial comment is, and I know this is a bit of a pain,
> that 'net' and 'net-next' patches need to be split into separate serries.
Thanks for the heads up. I will refactor the series. First a part which
shall go into "net" and I will make the second part RFC with more
explanations how to reproduce the issue.
All the best,
Wolfram