[Skiboot] [PATCH 2/2] Remove skiboot 'trace' infrastructure

Stewart Smith stewart at linux.ibm.com
Wed May 22 13:25:03 AEST 2019


Oliver <oohall at gmail.com> writes:
> On Wed, Dec 12, 2018 at 3:07 PM Joel Stanley <joel at jms.id.au> wrote:
>>
>> On Wed, 12 Dec 2018 at 14:28, Stewart Smith <stewart at linux.ibm.com> wrote:
>> >
>> > Joel Stanley <joel at jms.id.au> writes:
>> > > This feature has bit rotted. For instance, the FSP TCE mapping was
>> > > broken in 2015 by commit f2a7a5c23d81 ("Move boot_tracebuf to BSS"),
>> > > and no one seems to have noticed.
>> >
>> > Oliver, you were showing signs in Slack of "volunteering" of becoming
>> > the maintainer of this stuff and cleaning up.
>> >
>> > Your thoughts on ripping it out for the moment versus just ripping out
>> > the FSP bit in patch 1?
>>
>> The linker script bit is a noop, as f2a7a5c23d81 stopped using the
>> section defined in the linker script.
>>
>> The comment about FSP was from map_debug_areas in
>> platforms/ibm-fsp/common.c. I assumed that Ben had the traces in a
>> defined area for it to be mapped properly. But I could be interpreting
>> the intent of that wrong.
>
> Yes, it was placed in it's own section so that the linker script could
> force it to a 4K alignment. Moving it to .bss would be fine if we
> fixed the linker script rather than removing the section annotation.
>
> Anyway, I have plans to use this stuff and as soon as I do it's going
> to require either:
> a) Reverting this patch to remove it, or
> b) Rewriting it entirely
>
> There's not much point in either, so leave this alone.

Jordan's series on updating things recently went in, so there's some
amount of care for the trace infrastructure. I'm going to hold off on
the removal of it, and maybe if nobody uses it for more than the 5 years
it's been between uses then we should look at ripping it out?

-- 
Stewart Smith
OPAL Architect, IBM.



More information about the Skiboot mailing list