Re: [PATCH 1/1] platform/x86/amd: Fix pmc compile dependency errors.

[Borislav Petkov]

On Mon, Nov 07, 2022 at 06:27:52AM -0600, Mario Limonciello wrote:
> On 11/7/22 04:28, Hans de Goede wrote:
> > Hi,
> > 
> > On 10/26/22 09:25, Yupeng Li wrote:
> > > When disabled CONFIG_SUSPEND and CONFIG_DEBUG_FS, get_metrics_table
> > > and amd_pmc_idlemask_read is defined under two conditions of this,
> > > pmc build with implicit declaration of function error.Some build error
> > > messages are as follows:
> > > 
> > >   CC [M]  drivers/platform/x86/amd/pmc.o
> > > drivers/platform/x86/amd/pmc.c: In function ‘smu_fw_info_show’:
> > > drivers/platform/x86/amd/pmc.c:436:6: error: implicit declaration of function ‘get_metrics_table’ [-Werror=implicit-function-declaration]
> > >    436 |  if (get_metrics_table(dev, &table))
> > >        |      ^~~~~~~~~~~~~~~~~
> > > drivers/platform/x86/amd/pmc.c: In function ‘amd_pmc_idlemask_show’:
> > > drivers/platform/x86/amd/pmc.c:508:8: error: implicit declaration of function ‘amd_pmc_idlemask_read’; did you mean ‘amd_pmc_idlemask_show’? [-Werror=implicit-function-declaration]
> > >    508 |   rc = amd_pmc_idlemask_read(dev, NULL, s);
> > >        |        ^~~~~~~~~~~~~~~~~~~~~
> > >        |        amd_pmc_idlemask_show
> > > cc1: some warnings being treated as errors
> > > 
> > > Signed-off-by: Yupeng Li <liyupeng@xxxxxxxxxx>
> > > Reviewed-by: Caicai <caizp2008@xxxxxxx>
> > 
> > Thank you for your patch, however I do not believe that this is the correct fix:
> > 
> > > ---
> > >   drivers/platform/x86/amd/pmc.c | 5 ++++-
> > >   1 file changed, 4 insertions(+), 1 deletion(-)
> > > 
> > > diff --git a/drivers/platform/x86/amd/pmc.c b/drivers/platform/x86/amd/pmc.c
> > > index ce859b300712..2b08039921b8 100644
> > > --- a/drivers/platform/x86/amd/pmc.c
> > > +++ b/drivers/platform/x86/amd/pmc.c
> > > @@ -433,8 +433,10 @@ static int smu_fw_info_show(struct seq_file *s, void *unused)
> > >   	struct smu_metrics table;
> > >   	int idx;
> > > +#if defined(CONFIG_SUSPEND) || defined(CONFIG_DEBUG_FS)
> > >   	if (get_metrics_table(dev, &table))
> > >   		return -EINVAL;
> > > +#endif
> > >   	seq_puts(s, "\n=== SMU Statistics ===\n");
> > >   	seq_printf(s, "Table Version: %d\n", table.table_version);
> > 
> > The table variable gets used here, but now it no longer has been initialized by
> > get_metrics_table() so just #ifdef-ing out the get_metrics_table() call is wrong.
> > 
> > > @@ -503,11 +505,12 @@ static int amd_pmc_idlemask_show(struct seq_file *s, void *unused)
> > >   		if (rc)
> > >   			return rc;
> > >   	}
> > > -
> > >   	if (dev->major > 56 || (dev->major >= 55 && dev->minor >= 37)) {
> > > +#if defined(CONFIG_SUSPEND) || defined(CONFIG_DEBUG_FS)
> > >   		rc = amd_pmc_idlemask_read(dev, NULL, s);
> > >   		if (rc)
> > >   			return rc;
> > > +#endif
> > >   	} else {
> > >   		seq_puts(s, "Unsupported SMU version for Idlemask\n");
> > >   	}
> > 
> > Same thing except now nothing gets send to the struct seq_file *s,
> > so this is wrong too.
> > 
> > Shyam / Mario, can you take a look at the reported compile errors please?
> > 
> > Regards,
> > 
> > Hans
> > 
> 
> What kernel is this?  Shyam removed CONFIG_DEBUGFS in b37fe34c83.  So is
> this failure still happening in 6.1-rc4?

It managed to trickle into stable too:

https://bugzilla.kernel.org/show_bug.cgi?id=216679

Adding bug reporter to Cc too.

>From a quick glance, I'd drop all that ifdeffery and add empty stubs.
IMNSVHO.

Thx.

-- 
Regards/Gruss,
    Boris.

https://people.kernel.org/tglx/notes-about-netiquette