Re: RFC: CXL: How to handle trace event ABI break from CXL_HEADERLOG_SIZE fix?

[Mauro Carvalho Chehab]

On Thu, 21 May 2026 07:23:45 -0500
"Bowman, Terry" <terry.bowman@xxxxxxx> wrote:

> On 5/20/2026 5:18 PM, Dan Williams (nvidia) wrote:
> > Bowman, Terry wrote:  
> >> Hi Mauro and Everyone,
> >>
> >> During development of the CXL protocol series [1], the Sashiko tool 
> >> identified an issue in the CXL RAS UCE trace logging for Ports and Endpoints. 
> >> Specifically, the CXL RAS UCE header log size is incorrectly defined in 
> >> drivers/cxl/cxl.h.
> >>
> >> The UCE header log size is currently defined as 128 u32s (512 bytes),
> >> whereas it should be 16 u32s (64 bytes) per CXL r4.0 8.2.4.17.7. Correcting 
> >> this will change the trace format and break the existing ABI contract 
> >> with rasdaemon.
> >>
> >> How would you recommend proceeding to resolve this?  
> > 
> > Split the definitions into one that reads from the register space and
> > once that passes the parsed log to the trace buffer.
> > 
> > Something like:
> > 
> > u32 hl[CXL_HEADERLOG_TRACE_SIZE_U32];
> > 
> > ...zero-fill that and put a comment about how userspace has already
> > grown a dependency on the buggy size.
> > 
> > In other words, keep the userspace compatibility, but fix the iomem
> > overflow.  
> 
> Ok, that's straight forward. Thanks.

Such approach sounds good to me.

Regards,
Mauro

Thanks,
Mauro