Re: [PATCH v2 1/1] mfd: intel-lpss: Fix too early PM enablement in the ACPI ->probe()

From: Lee Jones
Date: Thu Nov 18 2021 - 03:42:39 EST

On Wed, 17 Nov 2021, Aditya Garg wrote:

> On 16-Nov-2021, at 9:23 PM, Lee Jones <lee.jones@xxxxxxxxxx<mailto:lee.jones@xxxxxxxxxx>> wrote:
> On Mon, 15 Nov 2021, Aditya Garg wrote:
> ________________________________
> From: Andy Shevchenko <andriy.shevchenko@xxxxxxxxxxxxxxx<mailto:andriy.shevchenko@xxxxxxxxxxxxxxx>>
> Sent: Monday, November 15, 2021 4:06 PM
> To: Aditya Garg
> Cc: Lee Jones; linux-kernel@xxxxxxxxxxxxxxx<mailto:linux-kernel@xxxxxxxxxxxxxxx>; Orlando Chamberlain
> Subject: Re: [PATCH v2 1/1] mfd: intel-lpss: Fix too early PM enablement in the ACPI ->probe()
> On Tue, Nov 02, 2021 at 05:42:35AM +0000, Aditya Garg wrote:
> On 02-Nov-2021, at 12:30 AM, Andy Shevchenko <andriy.shevchenko@xxxxxxxxxxxxxxx<mailto:andriy.shevchenko@xxxxxxxxxxxxxxx>> wrote:
> The runtime PM callback may be called as soon as the runtime PM facility
> is enabled and activated. It means that ->suspend() may be called before
> we finish probing the device in the ACPI case. Hence, NULL pointer
> dereference:
> intel-lpss INT34BA:00: IRQ index 0 not found
> BUG: kernel NULL pointer dereference, address: 0000000000000030
> ...
> Workqueue: pm pm_runtime_work
> RIP: 0010:intel_lpss_suspend+0xb/0x40 [intel_lpss]
> To fix this, first try to register the device and only after that enable
> runtime PM facility.
> Fixes: 4b45efe85263 ("mfd: Add support for Intel Sunrisepoint LPSS devices")
> Reported-by: Orlando Chamberlain <redecorating@xxxxxxxxxxxxxx<mailto:redecorating@xxxxxxxxxxxxxx>>
> Reported-by: Aditya Garg <gargaditya08@xxxxxxxx<mailto:gargaditya08@xxxxxxxx>>
> Signed-off-by: Andy Shevchenko <andriy.shevchenko@xxxxxxxxxxxxxxx<mailto:andriy.shevchenko@xxxxxxxxxxxxxxx>>
> Tested-by: Aditya Garg <gargaditya08@xxxxxxxx<mailto:gargaditya08@xxxxxxxx>>
> ---
> v2: added tag (Aditya), returned 0 explicitly at the end of ->probe()
> It works
> Thanks for testing again!
> Lee, can we have this as a fix material for v5.16-rcX?
> Generally not, no.
> We usually only push patches for the -rcs if they fix something that
> was broken during the merge window. Not 6+ years ago. :)
> However, if other valid fixes appear, I'll shove it into the PR too.
> I would like to have it backported to stable 5.15 too.
> Yes, once it's merged into mainline, it will be back-ported as far
> back as it applies cleanly.
> If you have a specific kernel in mind, you should indicate it on the
> end of the Fixes line.
> No specific kernel in mind. The issue was observed since 5.15.

That was really hard to read!

Please fix your mailer to set previous replies as quotes.

Not only did you not quote the mail you were replying to, you also
stripped out *all* previous quotes too. I didn't even know you could
do that!

Lee Jones [李琼斯]
Senior Technical Lead - Developer Services │ Open source software for Arm SoCs
Follow Linaro: Facebook | Twitter | Blog