[PATCH] drm/xe: Drop useless d3cold allowed message

[Lucas De Marchi]

On Tue, Nov 19, 2024 at 02:03:34PM -0800, Matthew Brost wrote:
>On Tue, Nov 19, 2024 at 03:01:48PM -0600, Lucas De Marchi wrote:
>> On Tue, Nov 19, 2024 at 01:50:20PM -0500, Rodrigo Vivi wrote:
>> > On Fri, Nov 15, 2024 at 03:24:27PM -0800, Matthew Brost wrote:
>> > > On Fri, Nov 15, 2024 at 04:53:04PM -0500, Rodrigo Vivi wrote:
>> > > > On Fri, Nov 15, 2024 at 11:21:55AM -0800, Matthew Brost wrote:
>> > > > > This message just spams dmesg providing little benefit. Remove it.
>> > > >
>> > > > I don't believe it is useless, but I also hate the spammers...
>> > > >
>> > > > perhaps we could bring the status of it in some debugfs form
>> > > > as a replacement?
>> > > >
>> > >
>> > > ftrace could an option so if you really care about this enable it there?
>> >
>> > hmmm it looks like a good idea... but no strong preference from my side...
>>
>> as long as we are talking about adding a tracepoint rather than a plain
>> trace_printk() unconditionally.
>>
>
>Yes, trace point rather than trace_printk as that would just spam the
>trace buffer.
>
>> with dynamic debug we could use a verbose level that is only enabled per
>> file when needed. Then we can keep messages like this rather than
>> misusing the trace buffer.
>>
>
>'with dynamic debug' - Can you explain what you mean here?

normally used with pr_err/pr_debug() etc so you can enable it in runtime
as needed. See https://docs.kernel.org/admin-guide/dynamic-debug-howto.html#dynamic-debug
It's currently broken in drm, but I think the last bits are getting fixed
(see db62482e3242 ("drm/xe: Fix build error for XE_IOCTL_DBG
macro")).... I only use it with other parts of the kernel and find it
very nice.

Lucas De Marchi


>
>Matt
>
>> Lucas De Marchi