From: Vinod Koul <vkoul@kernel.org>
To: Peter Ujfalusi <peter.ujfalusi@ti.com>
Cc: dan.j.williams@intel.com, dmaengine@vger.kernel.org,
linux-arm-kernel@lists.infradead.org, linux-omap@vger.kernel.org
Subject: Re: [PATCH] dmaengine: ti: omap-dma: Improved memcpy polling support
Date: Fri, 5 Jul 2019 11:53:34 +0530 [thread overview]
Message-ID: <20190705062334.GV2911@vkoul-mobl> (raw)
In-Reply-To: <20190618132416.26874-1-peter.ujfalusi@ti.com>
On 18-06-19, 16:24, Peter Ujfalusi wrote:
> When a DMA client driver does not set the DMA_PREP_INTERRUPT because it
> does not want to use interrupts for DMA completion or because it can not
> rely on DMA interrupts due to executing the memcpy when interrupts are
> disabled it will poll the status of the transfer.
>
> If the interrupts are enabled then the cookie will be set completed in the
> interrupt handler so only check in HW completion when the polling is really
> needed.
>
> Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
> ---
> Hi,
>
> This patch fine-tunes the omap-dma polled memcpy support to be inline with how
> the EDMA driver is handling it.
>
> The polled completion can be tested by applying:
> https://patchwork.kernel.org/patch/10966499/
>
> and run the dmatest with polled = 1 on boards where sDMA is used.
>
> Or boot up any dra7 family device with display enabled. The workaround for DMM
> errata i878 uses polled DMA memcpy.
>
> Regards,
> Peter
>
> drivers/dma/ti/omap-dma.c | 37 ++++++++++++++++++++++++-------------
> 1 file changed, 24 insertions(+), 13 deletions(-)
>
> diff --git a/drivers/dma/ti/omap-dma.c b/drivers/dma/ti/omap-dma.c
> index 5ba7d8485026..75d8f7e13c8d 100644
> --- a/drivers/dma/ti/omap-dma.c
> +++ b/drivers/dma/ti/omap-dma.c
> @@ -94,6 +94,7 @@ struct omap_desc {
> bool using_ll;
> enum dma_transfer_direction dir;
> dma_addr_t dev_addr;
> + bool polled;
>
> int32_t fi; /* for OMAP_DMA_SYNC_PACKET / double indexing */
> int16_t ei; /* for double indexing */
> @@ -834,20 +835,10 @@ static enum dma_status omap_dma_tx_status(struct dma_chan *chan,
>
> ret = dma_cookie_status(chan, cookie, txstate);
>
> - if (!c->paused && c->running) {
> - uint32_t ccr = omap_dma_chan_read(c, CCR);
> - /*
> - * The channel is no longer active, set the return value
> - * accordingly
> - */
> - if (!(ccr & CCR_ENABLE))
> - ret = DMA_COMPLETE;
> - }
> -
> + spin_lock_irqsave(&c->vc.lock, flags);
> if (ret == DMA_COMPLETE || !txstate)
> - return ret;
> + goto out;
>
> - spin_lock_irqsave(&c->vc.lock, flags);
> vd = vchan_find_desc(&c->vc, cookie);
> if (vd) {
> txstate->residue = omap_dma_desc_size(to_omap_dma_desc(&vd->tx));
> @@ -868,6 +859,23 @@ static enum dma_status omap_dma_tx_status(struct dma_chan *chan,
> }
> if (ret == DMA_IN_PROGRESS && c->paused)
> ret = DMA_PAUSED;
> +
> +out:
> + if (ret == DMA_IN_PROGRESS && c->running && c->desc &&
> + c->desc->polled && c->desc->vd.tx.cookie == cookie) {
heh, that makes quite a read!
checking DMA_IN_PROGRESS should not make sense, as we should bail out at
the start if it is completed
I think other can be optimzed to make it a better read!
> + uint32_t ccr = omap_dma_chan_read(c, CCR);
> + /*
> + * The channel is no longer active, set the return value
> + * accordingly
> + */
> + if (!(ccr & CCR_ENABLE)) {
> + struct omap_desc *d = c->desc;
> + ret = DMA_COMPLETE;
> + omap_dma_start_desc(c);
> + vchan_cookie_complete(&d->vd);
> + }
> + }
> +
> spin_unlock_irqrestore(&c->vc.lock, flags);
>
> return ret;
> @@ -1233,7 +1241,10 @@ static struct dma_async_tx_descriptor *omap_dma_prep_dma_memcpy(
> d->ccr = c->ccr;
> d->ccr |= CCR_DST_AMODE_POSTINC | CCR_SRC_AMODE_POSTINC;
>
> - d->cicr = CICR_DROP_IE | CICR_FRAME_IE;
> + if (tx_flags & DMA_PREP_INTERRUPT)
> + d->cicr |= CICR_FRAME_IE;
> + else
> + d->polled = true;
>
> d->csdp = data_type;
>
> --
> Peter
>
> Texas Instruments Finland Oy, Porkkalankatu 22, 00180 Helsinki.
> Y-tunnus/Business ID: 0615521-4. Kotipaikka/Domicile: Helsinki
--
~Vinod
next prev parent reply other threads:[~2019-07-05 6:26 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-06-18 13:24 [PATCH] dmaengine: ti: omap-dma: Improved memcpy polling support Peter Ujfalusi
2019-07-05 6:23 ` Vinod Koul [this message]
2019-07-12 21:10 ` Peter Ujfalusi
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=20190705062334.GV2911@vkoul-mobl \
--to=vkoul@kernel.org \
--cc=dan.j.williams@intel.com \
--cc=dmaengine@vger.kernel.org \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=linux-omap@vger.kernel.org \
--cc=peter.ujfalusi@ti.com \
/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).