linux-kernel - Re: [PATCH v3 79/79] media: hantro: document the usage of pm_runtime_get

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20210428082742.20d54db1@coco.lan>
Date:   Wed, 28 Apr 2021 08:27:42 +0200
From:   Mauro Carvalho Chehab <mchehab+huawei@...nel.org>
To:     Ezequiel Garcia <ezequiel@...labora.com>
Cc:     Robin Murphy <robin.murphy@....com>, linuxarm@...wei.com,
        mauro.chehab@...wei.com,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        Mauro Carvalho Chehab <mchehab@...nel.org>,
        Philipp Zabel <p.zabel@...gutronix.de>,
        devel@...verdev.osuosl.org, linux-kernel@...r.kernel.org,
        linux-media@...r.kernel.org, linux-rockchip@...ts.infradead.org
Subject: Re: [PATCH v3 79/79] media: hantro: document the usage of
 pm_runtime_get_sync()

Em Tue, 27 Apr 2021 12:18:32 -0300
Ezequiel Garcia <ezequiel@...labora.com> escreveu:

> On Tue, 2021-04-27 at 16:08 +0100, Robin Murphy wrote:
> > On 2021-04-27 11:27, Mauro Carvalho Chehab wrote:  
> > > Despite other *_get()/*_put() functions, where usage count is
> > > incremented only if not errors, the pm_runtime_get_sync() has
> > > a different behavior, incrementing the counter *even* on
> > > errors.
> > > 
> > > That's an error prone behavior, as people often forget to
> > > decrement the usage counter.
> > > 
> > > However, the hantro driver depends on this behavior, as it
> > > will decrement the usage_count unconditionally at the m2m
> > > job finish time, which makes sense.
> > > 
> > > So, intead of using the pm_runtime_resume_and_get() that
> > > would decrement the counter on error, keep the current
> > > API, but add a documentation explaining the rationale for
> > > keep using pm_runtime_get_sync().
> > > 
> > > Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@...nel.org>
> > > ---
> > >   drivers/staging/media/hantro/hantro_drv.c | 7 +++++++
> > >   1 file changed, 7 insertions(+)
> > > 
> > > diff --git a/drivers/staging/media/hantro/hantro_drv.c b/drivers/staging/media/hantro/hantro_drv.c
> > > index 595e82a82728..96f940c1c85c 100644
> > > --- a/drivers/staging/media/hantro/hantro_drv.c
> > > +++ b/drivers/staging/media/hantro/hantro_drv.c
> > > @@ -155,6 +155,13 @@ static void device_run(void *priv)
> > >         ret = clk_bulk_enable(ctx->dev->variant->num_clocks, ctx->dev->clocks);
> > >         if (ret)
> > >                 goto err_cancel_job;  
> > 
> > ..except this can also cause the same pm_runtime_put_autosuspend() call 
> > without even reaching the "matching" get below, so rather than some kind 
> > of cleverness it seems more like it's just broken :/
> >   
> 
> Indeed, I was trying to find time to cook a quick patch, but kept
> getting preempted.
> 
> Feel free to submit a fix for this, otherwise, I'll try to find
> time later this week.

What about doing this instead:

diff --git a/drivers/staging/media/hantro/hantro_drv.c b/drivers/staging/media/hantro/hantro_drv.c
index 595e82a82728..67de6b15236d 100644
--- a/drivers/staging/media/hantro/hantro_drv.c
+++ b/drivers/staging/media/hantro/hantro_drv.c
@@ -56,14 +56,12 @@ dma_addr_t hantro_get_ref(struct hantro_ctx *ctx, u64 ts)
 	return hantro_get_dec_buf_addr(ctx, buf);
 }
 
-static void hantro_job_finish(struct hantro_dev *vpu,
-			      struct hantro_ctx *ctx,
-			      enum vb2_buffer_state result)
+static void hantro_job_finish_no_pm(struct hantro_dev *vpu,
+				    struct hantro_ctx *ctx,
+				    enum vb2_buffer_state result)
 {
 	struct vb2_v4l2_buffer *src, *dst;
 
-	pm_runtime_mark_last_busy(vpu->dev);
-	pm_runtime_put_autosuspend(vpu->dev);
 	clk_bulk_disable(vpu->variant->num_clocks, vpu->clocks);
 
 	src = v4l2_m2m_next_src_buf(ctx->fh.m2m_ctx);
@@ -81,6 +79,16 @@ static void hantro_job_finish(struct hantro_dev *vpu,
 					 result);
 }
 
+static void hantro_job_finish(struct hantro_dev *vpu,
+			      struct hantro_ctx *ctx,
+			      enum vb2_buffer_state result)
+{
+	pm_runtime_mark_last_busy(vpu->dev);
+	pm_runtime_put_autosuspend(vpu->dev);
+
+	hantro_job_finish_no_pm(vpu, ctx, result);
+}
+
 void hantro_irq_done(struct hantro_dev *vpu,
 		     enum vb2_buffer_state result)
 {
@@ -152,12 +160,13 @@ static void device_run(void *priv)
 	src = hantro_get_src_buf(ctx);
 	dst = hantro_get_dst_buf(ctx);
 
+	ret = pm_runtime_resume_and_get(ctx->dev->dev);
+	if (ret < 0)
+		goto err_cancel_job;
+
 	ret = clk_bulk_enable(ctx->dev->variant->num_clocks, ctx->dev->clocks);
 	if (ret)
 		goto err_cancel_job;
-	ret = pm_runtime_get_sync(ctx->dev->dev);
-	if (ret < 0)
-		goto err_cancel_job;
 
 	v4l2_m2m_buf_copy_metadata(src, dst, true);
 
@@ -165,7 +174,7 @@ static void device_run(void *priv)
 	return;
 
 err_cancel_job:
-	hantro_job_finish(ctx->dev, ctx, VB2_BUF_STATE_ERROR);
+	hantro_job_finish_no_pm(ctx->dev, ctx, VB2_BUF_STATE_ERROR);
 }
 
 static struct v4l2_m2m_ops vpu_m2m_ops = {

Thanks,
Mauro