[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CY5PR11MB62579D3B679BB437017529D79702A@CY5PR11MB6257.namprd11.prod.outlook.com>
Date: Mon, 24 Jul 2023 11:06:02 +0000
From: "Lu, Brent" <brent.lu@...el.com>
To: Andy Shevchenko <andriy.shevchenko@...ux.intel.com>,
Pierre-Louis Bossart <pierre-louis.bossart@...ux.intel.com>
CC: "alsa-devel@...a-project.org" <alsa-devel@...a-project.org>,
"Rojewski, Cezary" <cezary.rojewski@...el.com>,
Liam Girdwood <liam.r.girdwood@...ux.intel.com>,
Peter Ujfalusi <peter.ujfalusi@...ux.intel.com>,
Bard Liao <yung-chuan.liao@...ux.intel.com>,
Ranjani Sridharan <ranjani.sridharan@...ux.intel.com>,
Kai Vehmanen <kai.vehmanen@...ux.intel.com>,
Mark Brown <broonie@...nel.org>,
"Jaroslav Kysela" <perex@...ex.cz>, Takashi Iwai <tiwai@...e.com>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
"Zhi, Yong" <yong.zhi@...el.com>,
Ajye Huang <ajye_huang@...pal.corp-partner.google.com>,
"Bhat, Uday M" <uday.m.bhat@...el.com>,
Terry Cheong <htcheong@...omium.org>,
"Chiang, Mac" <mac.chiang@...el.com>,
"R, Dharageswari" <dharageswari.r@...el.com>,
Kuninori Morimoto <kuninori.morimoto.gx@...esas.com>
Subject: RE: [PATCH 1/2] ASoC: Intel: maxim-common: get codec number from ACPI
> > > +/* helper function to get the number of specific codec */
>
> ...and leak a lot of reference counts...
>
> > > +static int get_num_codecs(const char *hid) {
> > > + struct acpi_device *adev = NULL;
> > > + int dev_num = 0;
> > > +
> > > + do {
> > > + adev = acpi_dev_get_next_match_dev(adev, hid, NULL, -1);
> >
> > Humm, I am a bit worried about reference counts.
> >
> > See
> > https://elixir.bootlin.com/linux/latest/source/drivers/acpi/utils.c#L9
> > 16, it's not clear to me where the get() is done.
> >
> > Adding Andy to make sure this is done right.
>
> Thank you for Cc'ing.
>
> Yes, the above code is problematic. One has to use the respective for_each macro
> (defined nearby the used API).
>
> > > + if (adev)
> > > + dev_num++;
> > > + } while (adev != NULL);
> > > +
> > > + return dev_num;
> > > +}
>
> --
> With Best Regards,
> Andy Shevchenko
>
Hi Andy,
Each invocation of acpi_dev_get_next_match_dev() calls acpi_dev_put() to release the
adev from previous call. And the last call returns NULL. It seems to me the reference count
should be fine. Is my understanding correct?
I saw the macro for_each_acpi_dev_match and re-write the function as follow. Thanks for
suggesting using the macro.
/* helper function to get the number of specific codec */
static int get_num_codecs(const char *hid) {
struct acpi_device *adev;
int dev_num = 0;
for_each_acpi_dev_match(adev, hid, NULL, -1)
dev_num++;
return dev_num;
}
Will test it in next few days.
Regards,
Brent
Powered by blists - more mailing lists