From: Peter Ujfalusi <peter.ujfalusi@ti.com>
To: Sascha Hauer <s.hauer@pengutronix.de>, <dmaengine@vger.kernel.org>
Cc: Vinod Koul <vkoul@kernel.org>,
Pengutronix Kernel Team <kernel@pengutronix.de>,
NXP Linux Team <linux-imx@nxp.com>,
Robert Jarzmik <robert.jarzmik@free.fr>
Subject: Re: [PATCH 2/5] dmaengine: virt-dma: Do not call desc_free() under a spin_lock
Date: Mon, 9 Dec 2019 10:33:36 +0200 [thread overview]
Message-ID: <c0839f58-9f85-4b23-59cc-75f7bdb98c18@ti.com> (raw)
In-Reply-To: <65b923ed-4370-089c-1d6c-ce7efac176e6@ti.com>
On 09/12/2019 9.48, Peter Ujfalusi wrote:
> Hi Sascha,
>
>
> On 06/12/2019 15.53, Sascha Hauer wrote:
>> vchan_vdesc_fini() shouldn't be called under a spin_lock. This is done
>> in two places, once in vchan_terminate_vdesc() and once in
>> vchan_synchronize(). Instead of freeing the vdesc right away, collect
>> the aborted vdescs on a separate list and free them along with the other
>> vdescs.
>>
>> Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
>> ---
>> drivers/dma/virt-dma.c | 1 +
>> drivers/dma/virt-dma.h | 17 +++--------------
>> 2 files changed, 4 insertions(+), 14 deletions(-)
>>
>> diff --git a/drivers/dma/virt-dma.c b/drivers/dma/virt-dma.c
>> index ec4adf4260a0..87d5bd53c98b 100644
>> --- a/drivers/dma/virt-dma.c
>> +++ b/drivers/dma/virt-dma.c
>> @@ -135,6 +135,7 @@ void vchan_init(struct virt_dma_chan *vc, struct dma_device *dmadev)
>> INIT_LIST_HEAD(&vc->desc_submitted);
>> INIT_LIST_HEAD(&vc->desc_issued);
>> INIT_LIST_HEAD(&vc->desc_completed);
>> + INIT_LIST_HEAD(&vc->desc_aborted);
>
> Can we keep the terminated term instead of aborted: desc_terminated
>
>>
>> tasklet_init(&vc->task, vchan_complete, (unsigned long)vc);
>>
>> diff --git a/drivers/dma/virt-dma.h b/drivers/dma/virt-dma.h
>> index 41883ee2c29f..6cae93624f0d 100644
>> --- a/drivers/dma/virt-dma.h
>> +++ b/drivers/dma/virt-dma.h
>> @@ -31,9 +31,9 @@ struct virt_dma_chan {
>> struct list_head desc_submitted;
>> struct list_head desc_issued;
>> struct list_head desc_completed;
>> + struct list_head desc_aborted;
>>
>> struct virt_dma_desc *cyclic;
>> - struct virt_dma_desc *vd_terminated;
>> };
>>
>> static inline struct virt_dma_chan *to_virt_chan(struct dma_chan *chan)
>> @@ -146,11 +146,8 @@ static inline void vchan_terminate_vdesc(struct virt_dma_desc *vd)
>> {
>> struct virt_dma_chan *vc = to_virt_chan(vd->tx.chan);
>>
>> - /* free up stuck descriptor */
>> - if (vc->vd_terminated)
>> - vchan_vdesc_fini(vc->vd_terminated);
>> + list_add_tail(&vd->node, &vc->desc_aborted);
>>
>> - vc->vd_terminated = vd;
>> if (vc->cyclic == vd)
>> vc->cyclic = NULL;
>> }
>> @@ -184,6 +181,7 @@ static inline void vchan_get_all_descriptors(struct virt_dma_chan *vc,
>> list_splice_tail_init(&vc->desc_submitted, head);
>> list_splice_tail_init(&vc->desc_issued, head);
>> list_splice_tail_init(&vc->desc_completed, head);
>> + list_splice_tail_init(&vc->desc_aborted, head);
>> }
>>
>> static inline void vchan_free_chan_resources(struct virt_dma_chan *vc)
>> @@ -212,16 +210,7 @@ static inline void vchan_free_chan_resources(struct virt_dma_chan *vc)
>> */
>> static inline void vchan_synchronize(struct virt_dma_chan *vc)
>> {
>> - unsigned long flags;
>> -
>> tasklet_kill(&vc->task);
>> -
>> - spin_lock_irqsave(&vc->lock, flags);
>> - if (vc->vd_terminated) {
>> - vchan_vdesc_fini(vc->vd_terminated);
>> - vc->vd_terminated = NULL;
>> - }
>> - spin_unlock_irqrestore(&vc->lock, flags);
>
> We don't want the terminated descriptors to accumulate until the channel
> is freed up.
Well, most DMA driver will clean it up in their terminate_all, but it is
better to do it in synchronize as well.
>
> spin_lock_irqsave(&vc->lock, flags);
> list_splice_tail_init(&vc->desc_terminated, &head);
> spin_unlock_irqrestore(&vc->lock, flags);
>
> list_for_each_entry_safe(vd, _vd, &head, node) {
> list_del(&vd->node);
> vchan_vdesc_fini(vd);
> }
>
>
>> }
>>
>> #endif
>>
>
> - Péter
>
> Texas Instruments Finland Oy, Porkkalankatu 22, 00180 Helsinki.
> Y-tunnus/Business ID: 0615521-4. Kotipaikka/Domicile: Helsinki
>
- Péter
Texas Instruments Finland Oy, Porkkalankatu 22, 00180 Helsinki.
Y-tunnus/Business ID: 0615521-4. Kotipaikka/Domicile: Helsinki
next prev parent reply other threads:[~2019-12-09 8:33 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-12-06 13:53 [PATCH 0/5] virt-dma and i.MX SDMA fixes Sascha Hauer
2019-12-06 13:53 ` [PATCH 1/5] dmaengine: virt-dma: Add missing locking around list operations Sascha Hauer
2019-12-09 7:38 ` Peter Ujfalusi
2019-12-10 11:56 ` Sascha Hauer
2019-12-06 13:53 ` [PATCH 2/5] dmaengine: virt-dma: Do not call desc_free() under a spin_lock Sascha Hauer
2019-12-09 7:48 ` Peter Ujfalusi
2019-12-09 8:33 ` Peter Ujfalusi [this message]
2019-12-10 11:58 ` Sascha Hauer
2019-12-06 13:53 ` [PATCH 3/5] dmaengine: imx-sdma: rename function Sascha Hauer
2019-12-06 13:53 ` [PATCH 4/5] dmaengine: imx-sdma: find desc first in sdma_tx_status Sascha Hauer
2019-12-06 13:53 ` [PATCH 5/5] dmaengine: imx-sdma: Fix memory leak Sascha Hauer
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=c0839f58-9f85-4b23-59cc-75f7bdb98c18@ti.com \
--to=peter.ujfalusi@ti.com \
--cc=dmaengine@vger.kernel.org \
--cc=kernel@pengutronix.de \
--cc=linux-imx@nxp.com \
--cc=robert.jarzmik@free.fr \
--cc=s.hauer@pengutronix.de \
--cc=vkoul@kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).