Re: [PATCH v2] acpi/ghes: Prevent sleeping with spinlock held

[Dan Williams]

Ira Weiny wrote:
> Ira Weiny wrote:
> > Jonathan Cameron wrote:
> > > On Wed, 14 Feb 2024 10:23:10 -0500
> > > Steven Rostedt <rostedt@xxxxxxxxxxx> wrote:
> > > 
> > > > On Wed, 14 Feb 2024 12:11:53 +0000
> > > > Jonathan Cameron <Jonathan.Cameron@xxxxxxxxxx> wrote:
> > > > 
> > > > > So I'm thinking this is a won't fix - wait for the printk rework to land and
> > > > > assume this will be resolved as well?  
> > > > 
> > > > That pretty much sums up what I was about to say ;-)
> > > > 
> > > > tp_printk is more of a hack and not to be used sparingly. With the right
> > > > trace events it can hang the machine.
> > > > 
> > > > So, you can use your internal patch locally, but I would recommend waiting
> > > > for the new printk changes to land.
> > 
> > Steven, Do you think that will land in 6.9?
> > 
> > > >
> > > > I'm really hoping that will be soon!
> > > > 
> > > > -- Steve
> > > 
> > > Thanks Steve,
> > > 
> > > Ira's fix is needed for other valid locking reasons - this was 'just another'
> > > lock debugging report that came up whilst testing it.
> > > 
> > > For this patch (not a potential additional one that we aren't going to do ;)
> > > 
> > > Tested-by: Jonathan Cameron <Jonathan.Cameron@xxxxxxxxxx>
> > > Reviewed-by: Jonathan Cameron <Jonathan.Cameron@xxxxxxxxxx>
> > 
> > Jonathan,
> > 
> > Again thanks for the testing!  However, Dan and I just discussed this and
> > he has an uneasy feeling about going forward with this for 6.8 final.
> > 
> > If we revert the following patch I can squash this fix and wait for the
> > tp_printk() fix to land in 6.9 and resubmit.
> > 
> > Dan here is the patch which backs out the actual bug:
> > 
> > 	Fixes: 671a794c33c6 ("acpi/ghes: Process CXL Component Events") 
> 
> Unfortunately this is not the only patch.
> 
> We need to revert this too:
> 
> Fixes: dc97f6344f20 ("cxl/pci: Register for and process CPER events") 
> 
> And then revert ...
> 	Fixes: 671a794c33c6 ("acpi/ghes: Process CXL Component Events") 
> 
> ... but there is a conflict.
> 
> Dan, below is the correct revert patch.  Let me know if you need more.
> 
> Ira
> 
> commit 807fbe9cac9b190dab83e3ff377a30d18859c8ab
> Author: Ira Weiny <ira.weiny@xxxxxxxxx>
> Date:   Wed Feb 14 15:25:24 2024 -0800
> 
>     Revert "acpi/ghes: Process CXL Component Events"
>     
>     This reverts commit 671a794c33c6e048ca5cedd5ad6af44d52d5d7e5.

Even reverts need changelogs, I can add one. I got conflicts trying to
apply this to current fixes branch. I think I am going to just
surgically backout the drivers/acpi/apei/ghes.c changes.