| Message ID | 20171005005336.GA23796@beast (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Thu, 05 Oct 2017 02:53:36 +0200, Kees Cook wrote: > > In preparation for unconditionally passing the struct timer_list pointer to > all timer callbacks, switch to using the new timer_setup() and from_timer() > to pass the timer pointer explicitly. This requires adding a pointer to > hold the timer's target substream, as there won't be a way to pass this in > the future. > > Cc: Jaroslav Kysela <perex@perex.cz> > Cc: Takashi Iwai <tiwai@suse.com> > Cc: Bhumika Goyal <bhumirks@gmail.com> > Cc: Mark Brown <broonie@kernel.org> > Cc: alsa-devel@alsa-project.org > Cc: Thomas Gleixner <tglx@linutronix.de> > Signed-off-by: Kees Cook <keescook@chromium.org> > --- > This requires commit 686fef928bba ("timer: Prepare to change timer > callback argument type") in v4.14-rc3, but should be otherwise > stand-alone. The conversions look straightforward, so I'm going to apply it (and other two you posted). But now I wonder why only three were chosen. Will you keep working on the rest usages in sound/*, or would you expect us doing the rest? thanks, Takashi
On Wed, Oct 4, 2017 at 11:22 PM, Takashi Iwai <tiwai@suse.de> wrote: > On Thu, 05 Oct 2017 02:53:36 +0200, > Kees Cook wrote: >> >> In preparation for unconditionally passing the struct timer_list pointer to >> all timer callbacks, switch to using the new timer_setup() and from_timer() >> to pass the timer pointer explicitly. This requires adding a pointer to >> hold the timer's target substream, as there won't be a way to pass this in >> the future. >> >> Cc: Jaroslav Kysela <perex@perex.cz> >> Cc: Takashi Iwai <tiwai@suse.com> >> Cc: Bhumika Goyal <bhumirks@gmail.com> >> Cc: Mark Brown <broonie@kernel.org> >> Cc: alsa-devel@alsa-project.org >> Cc: Thomas Gleixner <tglx@linutronix.de> >> Signed-off-by: Kees Cook <keescook@chromium.org> >> --- >> This requires commit 686fef928bba ("timer: Prepare to change timer >> callback argument type") in v4.14-rc3, but should be otherwise >> stand-alone. > > The conversions look straightforward, so I'm going to apply it (and > other two you posted). But now I wonder why only three were chosen. > Will you keep working on the rest usages in sound/*, or would you > expect us doing the rest? Hi, thanks for applying them! I'm working my way through the "non-trivial" changes first. I have a coccinelle script that will do a mass-conversion of the common cases. It's possible that those changes cover the other usages in sound/* or that I haven't made my way through the others yet (it's a very long list). I wouldn't object to other people helping with the conversions, of course. :) -Kees
On Thu, 05 Oct 2017 19:17:16 +0200, Kees Cook wrote: > > On Wed, Oct 4, 2017 at 11:22 PM, Takashi Iwai <tiwai@suse.de> wrote: > > On Thu, 05 Oct 2017 02:53:36 +0200, > > Kees Cook wrote: > >> > >> In preparation for unconditionally passing the struct timer_list pointer to > >> all timer callbacks, switch to using the new timer_setup() and from_timer() > >> to pass the timer pointer explicitly. This requires adding a pointer to > >> hold the timer's target substream, as there won't be a way to pass this in > >> the future. > >> > >> Cc: Jaroslav Kysela <perex@perex.cz> > >> Cc: Takashi Iwai <tiwai@suse.com> > >> Cc: Bhumika Goyal <bhumirks@gmail.com> > >> Cc: Mark Brown <broonie@kernel.org> > >> Cc: alsa-devel@alsa-project.org > >> Cc: Thomas Gleixner <tglx@linutronix.de> > >> Signed-off-by: Kees Cook <keescook@chromium.org> > >> --- > >> This requires commit 686fef928bba ("timer: Prepare to change timer > >> callback argument type") in v4.14-rc3, but should be otherwise > >> stand-alone. > > > > The conversions look straightforward, so I'm going to apply it (and > > other two you posted). But now I wonder why only three were chosen. > > Will you keep working on the rest usages in sound/*, or would you > > expect us doing the rest? > > Hi, thanks for applying them! > > I'm working my way through the "non-trivial" changes first. I have a > coccinelle script that will do a mass-conversion of the common cases. > It's possible that those changes cover the other usages in sound/* or > that I haven't made my way through the others yet (it's a very long > list). I wouldn't object to other people helping with the conversions, > of course. :) OK, scripting is a better way, then I'd save my time and wait until you get more systematic coverage, then :) thanks, Takashi
diff --git a/sound/sh/aica.c b/sound/sh/aica.c index fdc680ae8aa0..2b26311405a4 100644 --- a/sound/sh/aica.c +++ b/sound/sh/aica.c @@ -299,14 +299,14 @@ static void run_spu_dma(struct work_struct *work) } } -static void aica_period_elapsed(unsigned long timer_var) +static void aica_period_elapsed(struct timer_list *t) { + struct snd_card_aica *dreamcastcard = from_timer(dreamcastcard, + t, timer); + struct snd_pcm_substream *substream = dreamcastcard->timer_substream; /*timer function - so cannot sleep */ int play_period; struct snd_pcm_runtime *runtime; - struct snd_pcm_substream *substream; - struct snd_card_aica *dreamcastcard; - substream = (struct snd_pcm_substream *) timer_var; runtime = substream->runtime; dreamcastcard = substream->pcm->private_data; /* Have we played out an additional period? */ @@ -336,12 +336,12 @@ static void spu_begin_dma(struct snd_pcm_substream *substream) /*get the queue to do the work */ schedule_work(&(dreamcastcard->spu_dma_work)); /* Timer may already be running */ - if (unlikely(dreamcastcard->timer.data)) { + if (unlikely(dreamcastcard->timer_substream)) { mod_timer(&dreamcastcard->timer, jiffies + 4); return; } - setup_timer(&dreamcastcard->timer, aica_period_elapsed, - (unsigned long) substream); + timer_setup(&dreamcastcard->timer, aica_period_elapsed, 0); + dreamcastcard->timer_substream = substream; mod_timer(&dreamcastcard->timer, jiffies + 4); } @@ -379,7 +379,7 @@ static int snd_aicapcm_pcm_close(struct snd_pcm_substream { struct snd_card_aica *dreamcastcard = substream->pcm->private_data; flush_work(&(dreamcastcard->spu_dma_work)); - if (dreamcastcard->timer.data) + if (dreamcastcard->timer_substream) del_timer(&dreamcastcard->timer); kfree(dreamcastcard->channel); spu_disable(); @@ -600,7 +600,7 @@ static int snd_aica_probe(struct platform_device *devptr) { int err; struct snd_card_aica *dreamcastcard; - dreamcastcard = kmalloc(sizeof(struct snd_card_aica), GFP_KERNEL); + dreamcastcard = kzalloc(sizeof(struct snd_card_aica), GFP_KERNEL); if (unlikely(!dreamcastcard)) return -ENOMEM; err = snd_card_new(&devptr->dev, index, SND_AICA_DRIVER, @@ -619,8 +619,6 @@ static int snd_aica_probe(struct platform_device *devptr) err = snd_aicapcmchip(dreamcastcard, 0); if (unlikely(err < 0)) goto freedreamcast; - dreamcastcard->timer.data = 0; - dreamcastcard->channel = NULL; /* Add basic controls */ err = add_aicamixer_controls(dreamcastcard); if (unlikely(err < 0))
In preparation for unconditionally passing the struct timer_list pointer to all timer callbacks, switch to using the new timer_setup() and from_timer() to pass the timer pointer explicitly. This requires adding a pointer to hold the timer's target substream, as there won't be a way to pass this in the future. Cc: Jaroslav Kysela <perex@perex.cz> Cc: Takashi Iwai <tiwai@suse.com> Cc: Bhumika Goyal <bhumirks@gmail.com> Cc: Mark Brown <broonie@kernel.org> Cc: alsa-devel@alsa-project.org Cc: Thomas Gleixner <tglx@linutronix.de> Signed-off-by: Kees Cook <keescook@chromium.org> --- This requires commit 686fef928bba ("timer: Prepare to change timer callback argument type") in v4.14-rc3, but should be otherwise stand-alone. --- sound/sh/aica.c | 20 +++++++++----------- 1 file changed, 9 insertions(+), 11 deletions(-)