Re: [PATCH] spi: add spi controller master driver for Blackfin 6xx processor
On Tue, Jun 4, 2013 at 3:13 AM, Scott Jiang wrote: > diff --git a/drivers/spi/spi-bfin6xx.c b/drivers/spi/spi-bfin6xx.c > +#include Why do you need this? -- Get 100% visibility into Java/.NET code with AppDynamics Lite It's a free troubleshooting tool designed for production Get down to code-level detail for bottlenecks, with <2% overhead. Download for free and get started troubleshooting in minutes. http://p.sf.net/sfu/appdyn_d2d_ap2 ___ spi-devel-general mailing list spi-devel-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/spi-devel-general
Re: [PATCH 1/2] spi:pl022: Disable/Enable functional clock from suspend/resume
On Wed, Sep 26, 2012 at 5:49 PM, Mark Brown wrote: > On Wed, Sep 26, 2012 at 02:17:36PM +0200, Linus Walleij wrote: >> On Wed, Sep 26, 2012 at 1:24 PM, Vipul Kumar Samar > >> > SPI functional clock must be disalble/enable in non RTPM suspend/resume >> > hooks. Currently it is only done for RTPM cases. > >> > This patch add support to disable/enbale clock for conventional >> > suspend/resume calls. > >> Cross dependency between runtime suspend/resume and >> common suspend/resume. Oh the horror ... > > This should be fine, we runtime resume before we suspend. I believe Vipul sent this patch for the cases where RTPM in not enabled in the configs. -- viresh -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ spi-devel-general mailing list spi-devel-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/spi-devel-general
Re: [PATCH] spi: pl022: Add clk_{un}prepare() support in runtime PM
On Tue, Sep 18, 2012 at 5:20 PM, Linus Walleij wrote:
> On Tue, Sep 18, 2012 at 6:09 AM, viresh kumar wrote:
>
>> Yes, we don't need to call prepare() again atleast for SPEAr. You are
>> correct.
>> I saw the driver after a long time :)
>
> I'm asking because it's actually OK to do this, I was more asking whether it
> was really needed by any platforms...
Yes, I got that. Patch from Vipul is correct and should be there for
any platforms
which do anything in prepare/unprepare. But Atleast for SPEAr we don't need it.
But i would still insist in keeping it for completeness. :)
> We clk_disable() at runtime_suspend() and clk_enable() at runtime resume,
> and the driver gives hints to the runtime PM layer to autosuspend the
> driver whenever it's unused. Check the pm_runtime_* calls.
Ahh.. How could i miss it.
>> The amba layer is taking care of interface clock only and not
>> functional clock. So
>> i believe that's not the magic code. :)
>
> This clock is the one for the external bus. In some designs these two
> clocks are one and the same, and these won't currently get into any clock
> disabled states, sadly. (We need to fix that some day.)
I went through the code and found following in amba/bus.c:
static int amba_pm_runtime_suspend(struct device *dev)
{
struct amba_device *pcdev = to_amba_device(dev);
int ret = pm_generic_runtime_suspend(dev);
if (ret == 0 && dev->driver)
clk_disable(pcdev->pclk);
return ret;
}
static int amba_pm_runtime_resume(struct device *dev)
{
struct amba_device *pcdev = to_amba_device(dev);
int ret;
if (dev->driver) {
ret = clk_enable(pcdev->pclk);
/* Failure is probably fatal to the system, but... */
if (ret)
return ret;
}
return pm_generic_runtime_resume(dev);
}
If i am not wrong, these routines also get called with runtiime suspend/resume
of pl022? If that is the case, the even the interface clock of pl022 is getting
disabled when not in used.
And so for Architectures like SPEAr (where functional and interface
clock are controlled
by a single bit), we don't need anything else for power saving, with
respect to clocks.
Isn't it so?
@Vipul/Vinit: Can you please confirm this behavior?
--
viresh
--
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
___
spi-devel-general mailing list
spi-devel-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/spi-devel-general
Re: [PATCH] spi: pl022: Add clk_{un}prepare() support in runtime PM
On Mon, Sep 17, 2012 at 7:09 PM, Linus Walleij wrote: > This driver does clk_prepare/unprepare at probe > and removed, so I guess what you're trying to say is that > on your platform the clk_unprepare() process context call > is needed to save power? > > Please elaborate... Hi Linus, Yes, we don't need to call prepare() again atleast for SPEAr. You are correct. I saw the driver after a long time :) Can you please elaborate, why can't i see any clk_disable/enable calls anywhere else from probe. If i remember correctly, earlier we used to enable/disable clk after transfers and also during suspend/resume. The amba layer is taking care of interface clock only and not functional clock. So i believe that's not the magic code. :) -- viresh -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ spi-devel-general mailing list spi-devel-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/spi-devel-general
Re: [PATCH] spi: pl022: Add clk_{un}prepare() support in runtime PM
On Mon, Sep 17, 2012 at 4:07 PM, Vipul Kumar Samar
wrote:
> clk_{un}prepare is mandatory for platforms using common clock framework. Add
> clk_{un}prepare() support for spi-pl022 runtime PM.
You are not calling these routines in actualy patch.. Fix commit log and add my
Reviewed-by.
> Signed-off-by: Vipul Kumar Samar
> ---
> drivers/spi/spi-pl022.c |9 ++---
> 1 files changed, 6 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/spi/spi-pl022.c b/drivers/spi/spi-pl022.c
> index f2a80ff..500e75e 100644
> --- a/drivers/spi/spi-pl022.c
> +++ b/drivers/spi/spi-pl022.c
> @@ -2334,7 +2334,7 @@ static int pl022_runtime_suspend(struct device *dev)
> {
> struct pl022 *pl022 = dev_get_drvdata(dev);
>
> - clk_disable(pl022->clk);
> + clk_disable_unprepare(pl022->clk);
>
> return 0;
> }
> @@ -2342,10 +2342,13 @@ static int pl022_runtime_suspend(struct device *dev)
> static int pl022_runtime_resume(struct device *dev)
> {
> struct pl022 *pl022 = dev_get_drvdata(dev);
> + int ret = 0;
>
> - clk_enable(pl022->clk);
> + ret = clk_prepare_enable(pl022->clk);
> + if (ret)
> + dev_err(dev, "could not enable SSP/SPI bus clock\n");
>
> - return 0;
> + return ret;
> }
> #endif
>
> --
> 1.7.2.2
>
>
> --
> Live Security Virtual Conference
> Exclusive live event will cover all the ways today's security and
> threat landscape has changed and how IT managers can respond. Discussions
> will include endpoint security, mobile security and the latest in malware
> threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
> ___
> spi-devel-general mailing list
> spi-devel-general@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/spi-devel-general
--
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
___
spi-devel-general mailing list
spi-devel-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/spi-devel-general
Re: [PATCH] spi/pl022: Fill master->dev.of_node to get spi devices registered via DT
On 19 August 2012 03:56, Linus Walleij wrote: > On Sat, Aug 18, 2012 at 4:25 AM, Viresh Kumar > wrote: > > > spi_register_master() calls of_register_spi_devices() to register spi > devices. > > This routine expects master->dev.of_node to be a valid pointer. This is > > responsibility of master driver to fill this field, which wasn't done > for pl022. > > Fix it to get devices added to pl022. > > Isn't this one of those things that *have* to be #ifdef CONFIG_OF? > > Next iteration, remember to add Mark Brown on To: because je's taking > care of SPI patches for the moment. > Ahh!! Forgot to reply :( Dropping this patch as it is already fixed in Roland's patches viresh -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ spi-devel-general mailing list spi-devel-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/spi-devel-general
[PATCH] spi/pl022: Fill master->dev.of_node to get spi devices registered via DT
spi_register_master() calls of_register_spi_devices() to register spi devices. This routine expects master->dev.of_node to be a valid pointer. This is responsibility of master driver to fill this field, which wasn't done for pl022. Fix it to get devices added to pl022. Signed-off-by: Viresh Kumar --- drivers/spi/spi-pl022.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/spi/spi-pl022.c b/drivers/spi/spi-pl022.c index aab518e..3a46848 100644 --- a/drivers/spi/spi-pl022.c +++ b/drivers/spi/spi-pl022.c @@ -2029,6 +2029,7 @@ pl022_probe(struct amba_device *adev, const struct amba_id *id) master->transfer_one_message = pl022_transfer_one_message; master->unprepare_transfer_hardware = pl022_unprepare_transfer_hardware; master->rt = platform_info->rt; + master->dev.of_node = dev->of_node; /* * Supports mode 0-3, loopback, and active low CS. Transfers are -- 1.7.12.rc2.18.g61b472e -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ spi-devel-general mailing list spi-devel-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/spi-devel-general
Re: [PATCH] spi: pl022: Fix calculate_effective_freq()
On 4/19/2012 2:44 PM, Viresh KUMAR wrote: > calculate_effective_freq() was still not optimized and there were cases when > it > returned without error and with values of cpsr and scr as zero. > > Also, the variable named found is not used well. > > This patch targets to optimize and correct this routine. Tested for SPEAr. > > Signed-off-by: Viresh Kumar > Tested-by: Vinit Kamalaksha Shenoy You forgot to review this :) -- viresh -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ spi-devel-general mailing list spi-devel-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/spi-devel-general
[PATCH] spi: pl022: Fix calculate_effective_freq()
calculate_effective_freq() was still not optimized and there were cases when it
returned without error and with values of cpsr and scr as zero.
Also, the variable named found is not used well.
This patch targets to optimize and correct this routine. Tested for SPEAr.
Signed-off-by: Viresh Kumar
Tested-by: Vinit Kamalaksha Shenoy
---
drivers/spi/spi-pl022.c | 23 +--
1 files changed, 17 insertions(+), 6 deletions(-)
diff --git a/drivers/spi/spi-pl022.c b/drivers/spi/spi-pl022.c
index 99d5f6d..490be03 100644
--- a/drivers/spi/spi-pl022.c
+++ b/drivers/spi/spi-pl022.c
@@ -1687,26 +1687,37 @@ static int calculate_effective_freq(struct pl022
*pl022, int freq, struct
while (scr <= SCR_MAX) {
tmp = spi_rate(rate, cpsdvsr, scr);
- if (tmp > freq)
+ if (tmp > freq) {
+ /* we need lower freq */
scr++;
+ continue;
+ }
+
/*
-* If found exact value, update and break.
-* If found more closer value, update and continue.
+* If found exact value, mark found and break.
+* If found more closer value, update and break.
*/
- else if ((tmp == freq) || (tmp > best_freq)) {
+ if (tmp > best_freq) {
best_freq = tmp;
best_cpsdvsr = cpsdvsr;
best_scr = scr;
if (tmp == freq)
- break;
+ found = 1;
}
- scr++;
+ /*
+* increased scr will give lower rates, which are not
+* required
+*/
+ break;
}
cpsdvsr += 2;
scr = SCR_MIN;
}
+ WARN(!best_freq, "pl022: Matching cpsdvsr and scr not found for %d Hz
rate \n",
+ freq);
+
clk_freq->cpsdvsr = (u8) (best_cpsdvsr & 0xFF);
clk_freq->scr = (u8) (best_scr & 0xFF);
dev_dbg(&pl022->adev->dev,
--
1.7.9
--
For Developers, A Lot Can Happen In A Second.
Boundary is the first to Know...and Tell You.
Monitor Your Applications in Ultra-Fine Resolution. Try it FREE!
http://p.sf.net/sfu/Boundary-d2dvs2
___
spi-devel-general mailing list
spi-devel-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/spi-devel-general
[PATCH] spi: pl022: Allow request for higher frequency than maximum possible
Currently, if we request for frequency greater than maximum possible, spi driver
returns error.
For example, if the spi block src frequency is 333/4 MHz, i.e. 83.33.. MHz,
maximum frequency programmable would be src/2. Which would come around 41.6...
It is difficult to pass frequency in these figures. We normally try to program
in round figures, like 42 MHz and it should get programmed to <=
requested_frequency, i.e. 41.6...
For this to happen, we must not return error even if requested freq is higher
than max possible. But should program it to max possible.
Reported-by: Vinit Kamalaksha Shenoy
Signed-off-by: Viresh Kumar
---
drivers/spi/spi-pl022.c | 10 --
1 files changed, 8 insertions(+), 2 deletions(-)
diff --git a/drivers/spi/spi-pl022.c b/drivers/spi/spi-pl022.c
index 09c925a..99d5f6d 100644
--- a/drivers/spi/spi-pl022.c
+++ b/drivers/spi/spi-pl022.c
@@ -1667,9 +1667,15 @@ static int calculate_effective_freq(struct pl022 *pl022,
int freq, struct
/* cpsdvsr = 254 & scr = 255 */
min_tclk = spi_rate(rate, CPSDVR_MAX, SCR_MAX);
- if (!((freq <= max_tclk) && (freq >= min_tclk))) {
+ if (freq > max_tclk)
+ dev_warn(&pl022->adev->dev,
+ "Max speed that can be programmed is %d Hz, you
requested %d\n",
+ max_tclk, freq);
+
+ if (freq < min_tclk) {
dev_err(&pl022->adev->dev,
- "controller data is incorrect: out of range frequency");
+ "Requested frequency: %d Hz is less than minimum
possible %d Hz\n",
+ freq, min_tclk);
return -EINVAL;
}
--
1.7.9
--
For Developers, A Lot Can Happen In A Second.
Boundary is the first to Know...and Tell You.
Monitor Your Applications in Ultra-Fine Resolution. Try it FREE!
http://p.sf.net/sfu/Boundary-d2dvs2
___
spi-devel-general mailing list
spi-devel-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/spi-devel-general
Re: [PATCH] spi/pl022: Fix error print while setting bits/word on pl022
On 4/17/2012 11:28 AM, Vinit Kamalaksha SHENOY wrote: > pl022 ssp controller supports word lengths from 4 to 16 bits. > The kernel error message says: > "a standard pl022 can only handle 1 <= n <= 16 bit words". > > Fixed above range to 4 <= n <= 16. > > Signed-off-by: Vinit Shenoy > --- > drivers/spi/spi-pl022.c |2 +- > 1 files changed, 1 insertions(+), 1 deletions(-) > > diff --git a/drivers/spi/spi-pl022.c b/drivers/spi/spi-pl022.c > index 09c925a..905752f 100644 > --- a/drivers/spi/spi-pl022.c > +++ b/drivers/spi/spi-pl022.c > @@ -1848,7 +1848,7 @@ static int pl022_setup(struct spi_device *spi) > "illegal data size for this controller!\n"); > dev_err(&spi->dev, > "a standard pl022 can only handle " > - "1 <= n <= 16 bit words\n"); > + "4 <= n <= 16 bit words\n"); > status = -ENOTSUPP; > goto err_config_params; > } Reviewed-by: Viresh Kumar -- viresh -- Better than sec? Nothing is better than sec when it comes to monitoring Big Data applications. Try Boundary one-second resolution app monitoring today. Free. http://p.sf.net/sfu/Boundary-dev2dev ___ spi-devel-general mailing list spi-devel-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/spi-devel-general
[PATCH] SPI/Pl022: Include types.h to remove compilation warnings
linux/pl022.h uses definitions like, u8, u16, etc, which have dependency of types.h file, which isn't included in it. So, we get compilation warnings. This patch includes types.h there to fix these warnings. Signed-off-by: Viresh Kumar --- include/linux/amba/pl022.h |2 ++ 1 files changed, 2 insertions(+), 0 deletions(-) diff --git a/include/linux/amba/pl022.h b/include/linux/amba/pl022.h index b8c5112..76dd1b1 100644 --- a/include/linux/amba/pl022.h +++ b/include/linux/amba/pl022.h @@ -25,6 +25,8 @@ #ifndef _SSP_PL022_H #define _SSP_PL022_H +#include + /** * whether SSP is in loopback mode or not */ -- 1.7.8.110.g4cb5d -- This SF email is sponsosred by: Try Windows Azure free for 90 days Click Here http://p.sf.net/sfu/sfd2d-msazure ___ spi-devel-general mailing list spi-devel-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/spi-devel-general
Re: [PATCH] spi/pl022: Add high priority message pump support
IO - 1 };
> +
> INIT_LIST_HEAD(&pl022->queue);
> spin_lock_init(&pl022->queue_lock);
>
> @@ -1581,11 +1586,29 @@ static int __init init_queue(struct pl022 *pl022)
> tasklet_init(&pl022->pump_transfers, pump_transfers,
> (unsigned long)pl022);
>
> - INIT_WORK(&pl022->pump_messages, pump_messages);
> - pl022->workqueue = create_singlethread_workqueue(
> + init_kthread_worker(&pl022->kworker);
> + pl022->kworker_task = kthread_run(kthread_worker_fn,
> + &pl022->kworker,
> dev_name(pl022->master->dev.parent));
> - if (pl022->workqueue == NULL)
> - return -EBUSY;
> + if (IS_ERR(pl022->kworker_task)) {
> + dev_err(&pl022->adev->dev,
> + "failed to create message pump task\n");
> + return -ENOMEM;
> + }
> + init_kthread_work(&pl022->pump_messages, pump_messages);
> +
> + /*
> + * Board config will indicate if this controller should run the
> + * message pump with high (realtime) priority to reduce the transfer
> + * latency on the bus by minimising the delay between a transfer
> + * request and the scheduling of the message pump thread. Without this
> + * setting the message pump thread will remain at default priority.
> + */
> + if (pl022->master_info->rt) {
> + dev_info(&pl022->adev->dev,
> + "will run message pump with realtime priority\n");
> + sched_setscheduler(pl022->kworker_task, SCHED_FIFO, ¶m);
> + }
>
> return 0;
> }
> @@ -1608,7 +1631,7 @@ static int start_queue(struct pl022 *pl022)
> pl022->next_msg_cs_active = false;
> spin_unlock_irqrestore(&pl022->queue_lock, flags);
>
> - queue_work(pl022->workqueue, &pl022->pump_messages);
> + queue_kthread_work(&pl022->kworker, &pl022->pump_messages);
>
> return 0;
> }
> @@ -1646,16 +1669,20 @@ static int destroy_queue(struct pl022 *pl022)
> int status;
>
> status = stop_queue(pl022);
> - /* we are unloading the module or failing to load (only two calls
> +
> + /*
> + * We are unloading the module or failing to load (only two calls
>* to this routine), and neither call can handle a return value.
> - * However, destroy_workqueue calls flush_workqueue, and that will
> - * block until all work is done. If the reason that stop_queue
> - * timed out is that the work will never finish, then it does no
> - * good to call destroy_workqueue, so return anyway. */
> + * However, flush_kthread_worker will block until all work is done.
> + * If the reason that stop_queue timed out is that the work will never
> + * finish, then it does no good to call flush/stop thread, so
> + * return anyway.
> + */
> if (status != 0)
> return status;
>
> - destroy_workqueue(pl022->workqueue);
> + flush_kthread_worker(&pl022->kworker);
> + kthread_stop(pl022->kworker_task);
>
> return 0;
> }
> @@ -1802,7 +1829,7 @@ static int pl022_transfer(struct spi_device *spi,
> struct spi_message *msg)
>
> list_add_tail(&msg->queue, &pl022->queue);
> if (pl022->running && !pl022->busy)
> - queue_work(pl022->workqueue, &pl022->pump_messages);
> + queue_kthread_work(&pl022->kworker, &pl022->pump_messages);
>
> spin_unlock_irqrestore(&pl022->queue_lock, flags);
> return 0;
> diff --git a/include/linux/amba/pl022.h b/include/linux/amba/pl022.h
> index 572f637..3672f40 100644
> --- a/include/linux/amba/pl022.h
> +++ b/include/linux/amba/pl022.h
> @@ -241,6 +241,8 @@ struct dma_chan;
> * @autosuspend_delay: delay in ms following transfer completion before the
> * runtime power management system suspends the device. A setting of 0
> * indicates no delay and the device will be suspended immediately.
> + * @rt: indicates the controller should run the message pump with realtime
> + * priority to minimise the transfer latency on the bus.
> */
> struct pl022_ssp_controller {
> u16 bus_id;
> @@ -250,6 +252,7 @@ struct pl022_ssp_controller {
> void *dma_rx_param;
> void *dma_tx_param;
> int autosuspend_delay;
> + bool rt;
> };
>
> /**
Looks fine.
Acked-by: Viresh Kumar
--
viresh
--
Keep Your Developer Skills Current with LearnDevNow!
The most comprehensive online learning library for Microsoft developers
is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3,
Metro Style Apps, more. Free future releases when you subscribe now!
http://p.sf.net/sfu/learndevnow-d2d
___
spi-devel-general mailing list
spi-devel-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/spi-devel-general
Re: [PATCH 7/7] spi/pl022: make the chip deselect handling thread safe
* enabled if there is message in the queue and it is
> + * for the same spi device.
>*
>* We cannot postpone this until pump_messages, because
>* after calling msg->complete (below) the driver that
> @@ -501,14 +487,26 @@ static void giveback(struct pl022 *pl022)
> struct spi_message, queue);
> spin_unlock_irqrestore(&pl022->queue_lock, flags);
>
> - /* see if the next and current messages point
> - * to the same chip
> + /*
> + * see if the next and current messages point
> + * to the same spi device.
>*/
> - if (next_msg && next_msg->spi != msg->spi)
> + if (next_msg && next_msg->spi != pl022->cur_msg->spi)
> next_msg = NULL;
> - if (!next_msg || msg->state == STATE_ERROR)
> - curr_cs_control(SSP_CHIP_DESELECT);
> + if (!next_msg || pl022->cur_msg->state == STATE_ERROR)
> + pl022->cur_chip->cs_control(SSP_CHIP_DESELECT);
> + else
> + pl022->next_msg_cs_active = true;
> }
> +
> + spin_lock_irqsave(&pl022->queue_lock, flags);
> + msg = pl022->cur_msg;
> + pl022->cur_msg = NULL;
> + pl022->cur_transfer = NULL;
> + pl022->cur_chip = NULL;
> + queue_work(pl022->workqueue, &pl022->pump_messages);
> + spin_unlock_irqrestore(&pl022->queue_lock, flags);
> +
> msg->state = NULL;
> if (msg->complete)
> msg->complete(msg->context);
> @@ -1350,7 +1348,7 @@ static void pump_transfers(unsigned long data)
>*/
> udelay(previous->delay_usecs);
>
> - /* Drop chip select only if cs_change is requested */
> + /* Reselect chip select only if cs_change was requested */
> if (previous->cs_change)
> pl022->cur_chip->cs_control(SSP_CHIP_SELECT);
> } else {
> @@ -1389,8 +1387,10 @@ static void do_interrupt_dma_transfer(struct pl022
> *pl022)
>*/
> u32 irqflags = ENABLE_ALL_INTERRUPTS & ~SSP_IMSC_MASK_RXIM;
>
> - /* Enable target chip */
> - pl022->cur_chip->cs_control(SSP_CHIP_SELECT);
> + /* Enable target chip, if not already active */
> + if (!pl022->next_msg_cs_active)
> + pl022->cur_chip->cs_control(SSP_CHIP_SELECT);
> +
> if (set_up_next_transfer(pl022, pl022->cur_transfer)) {
> /* Error path */
> pl022->cur_msg->state = STATE_ERROR;
> @@ -1445,7 +1445,8 @@ static void do_polling_transfer(struct pl022 *pl022)
> } else {
> /* STATE_START */
> message->state = STATE_RUNNING;
> - pl022->cur_chip->cs_control(SSP_CHIP_SELECT);
> + if (!pl022->next_msg_cs_active)
> + pl022->cur_chip->cs_control(SSP_CHIP_SELECT);
> }
>
> /* Configuration Changing Per Transfer */
> @@ -1604,6 +1605,7 @@ static int start_queue(struct pl022 *pl022)
> pl022->cur_msg = NULL;
> pl022->cur_transfer = NULL;
> pl022->cur_chip = NULL;
> + pl022->next_msg_cs_active = false;
> spin_unlock_irqrestore(&pl022->queue_lock, flags);
>
> queue_work(pl022->workqueue, &pl022->pump_messages);
Reviewed-by: Viresh Kumar
--
viresh
--
All the data continuously generated in your IT infrastructure
contains a definitive record of customers, application performance,
security threats, fraudulent activity, and more. Splunk takes this
data and makes sense of it. IT sense. And common sense.
http://p.sf.net/sfu/splunk-novd2d
___
spi-devel-general mailing list
spi-devel-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/spi-devel-general
Re: [PATCH 4/7] spi/pl022: move device disable to workqueue thread
On 11/22/2011 1:55 PM, Linus WALLEIJ wrote:
> if (list_empty(&pl022->queue) || !pl022->running) {
> + if (pl022->busy) {
> + pm_runtime_put(&pl022->adev->dev);
> + }
We used to get warnings from checkpatch, for single line code inside {}.
Don't we get them anymore?
Reviewed-by: Viresh Kumar
--
viresh
--
All the data continuously generated in your IT infrastructure
contains a definitive record of customers, application performance,
security threats, fraudulent activity, and more. Splunk takes this
data and makes sense of it. IT sense. And common sense.
http://p.sf.net/sfu/splunk-novd2d
___
spi-devel-general mailing list
spi-devel-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/spi-devel-general
Re: [PATCH 3/7] spi/pl022: skip default configuration before suspending
On 11/22/2011 1:54 PM, Linus WALLEIJ wrote: > From: Virupax Sadashivpetimath > > The loading of the default configuration before suspending has > been in the driver since its inception, but it is not really > needed. Especially so since we take to all the trouble of > enabling and disabling power and clock just to do this. Let's > scrap this now. > > Signed-off-by: Virupax Sadashivpetimath > > Signed-off-by: Linus Walleij > --- > ChangeLog v1->v2: > - Use local dev pointer instead of copy/paste bug &adev->dev > --- > drivers/spi/spi-pl022.c |5 - > 1 files changed, 0 insertions(+), 5 deletions(-) > > diff --git a/drivers/spi/spi-pl022.c b/drivers/spi/spi-pl022.c > index 0d0b165..b4038f9 100644 > --- a/drivers/spi/spi-pl022.c > +++ b/drivers/spi/spi-pl022.c > @@ -2310,11 +2310,6 @@ static int pl022_suspend(struct device *dev) > return status; > } > > - amba_vcore_enable(pl022->adev); > - amba_pclk_enable(pl022->adev); > - load_ssp_default_config(pl022); > - amba_pclk_disable(pl022->adev); > - amba_vcore_disable(pl022->adev); > dev_dbg(dev, "suspended\n"); > return 0; > } Reviewed-by: Viresh Kumar -- viresh -- All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity, and more. Splunk takes this data and makes sense of it. IT sense. And common sense. http://p.sf.net/sfu/splunk-novd2d ___ spi-devel-general mailing list spi-devel-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/spi-devel-general
Re: [PATCH 6/6] spi/pl022: add support for pm_runtime autosuspend
On 11/9/2011 4:09 PM, Linus WALLEIJ wrote:
> From: Chris Blair
>
> Adds support for configuring the spi bus to use autosuspend for
> runtime power management. This can reduce the latency in starting an
> spi transfer by not suspending the device immediately following
> completion of a transfer. If another transfer then takes place before
> the autosuspend timeout, the call to resume the device can return
> immediately rather than needing to risk sleeping in order to resume
> the device.
>
> Signed-off-by: Chris Blair
> Signed-off-by: Linus Walleij
> ---
> drivers/spi/spi-pl022.c| 20 ++--
> include/linux/amba/pl022.h |4
> 2 files changed, 22 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/spi/spi-pl022.c b/drivers/spi/spi-pl022.c
> index 2e3522d..d3d6521 100644
> --- a/drivers/spi/spi-pl022.c
> +++ b/drivers/spi/spi-pl022.c
> @@ -1515,7 +1515,13 @@ static void pump_messages(struct work_struct *work)
> /* nothing more to do - disable spi/ssp and power off */
> writew((readw(SSP_CR1(pl022->virtbase)) &
> (~SSP_CR1_MASK_SSE)), SSP_CR1(pl022->virtbase));
> - pm_runtime_put(&pl022->adev->dev);
> +
> + if (pl022->master_info->autosuspend_delay > 0) {
> + pm_runtime_mark_last_busy(&pl022->adev->dev);
> + pm_runtime_put_autosuspend(&pl022->adev->dev);
> + } else {
> + pm_runtime_put(&pl022->adev->dev);
> + }
> }
> pl022->busy = false;
> spin_unlock_irqrestore(&pl022->queue_lock, flags);
> @@ -2245,7 +2251,17 @@ pl022_probe(struct amba_device *adev, const struct
> amba_id *id)
> dev_dbg(dev, "probe succeeded\n");
>
> /* let runtime pm put suspend */
> - pm_runtime_put(dev);
> + if (platform_info->autosuspend_delay > 0) {
> + dev_info(&adev->dev,
> + "will use autosuspend for runtime pm, delay %dms\n",
> + platform_info->autosuspend_delay);
> + pm_runtime_set_autosuspend_delay(dev,
> + platform_info->autosuspend_delay);
> + pm_runtime_use_autosuspend(dev);
> + pm_runtime_put_autosuspend(dev);
> + } else {
> + pm_runtime_put(dev);
> + }
> return 0;
>
> err_spi_register:
> diff --git a/include/linux/amba/pl022.h b/include/linux/amba/pl022.h
> index 4ce98f5..572f637 100644
> --- a/include/linux/amba/pl022.h
> +++ b/include/linux/amba/pl022.h
> @@ -238,6 +238,9 @@ struct dma_chan;
> * @enable_dma: if true enables DMA driven transfers.
> * @dma_rx_param: parameter to locate an RX DMA channel.
> * @dma_tx_param: parameter to locate a TX DMA channel.
> + * @autosuspend_delay: delay in ms following transfer completion before the
> + * runtime power management system suspends the device. A setting of 0
> + * indicates no delay and the device will be suspended immediately.
> */
> struct pl022_ssp_controller {
> u16 bus_id;
> @@ -246,6 +249,7 @@ struct pl022_ssp_controller {
> bool (*dma_filter)(struct dma_chan *chan, void *filter_param);
> void *dma_rx_param;
> void *dma_tx_param;
> + int autosuspend_delay;
> };
>
> /**
Reviewed-by: Viresh Kumar
--
viresh
--
RSA(R) Conference 2012
Save $700 by Nov 18
Register now
http://p.sf.net/sfu/rsa-sfdev2dev1
___
spi-devel-general mailing list
spi-devel-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/spi-devel-general
Re: [PATCH 5/6] spi/pl022: move device disable to workqueue thread
On 11/9/2011 4:09 PM, Linus WALLEIJ wrote:
> From: Chris Blair
> diff --git a/drivers/spi/spi-pl022.c b/drivers/spi/spi-pl022.c
> index bffad2a..2e3522d 100644
> --- a/drivers/spi/spi-pl022.c
> +++ b/drivers/spi/spi-pl022.c
> @@ -512,13 +512,6 @@ static void giveback(struct pl022 *pl022)
> msg->state = NULL;
> if (msg->complete)
> msg->complete(msg->context);
> -
> - /* disable the SPI/SSP operation */
> - writew((readw(SSP_CR1(pl022->virtbase)) &
> - (~SSP_CR1_MASK_SSE)), SSP_CR1(pl022->virtbase));
> -
We added this in an earlier patch in the same patchset. It would be better
if we can shift that patch after this one.
> - /* This message is completed, so let's turn off the clocks & power */
> - pm_runtime_put(&pl022->adev->dev);
> }
>
> /**
> @@ -1513,10 +1506,17 @@ static void pump_messages(struct work_struct *work)
> struct pl022 *pl022 =
> container_of(work, struct pl022, pump_messages);
> unsigned long flags;
> + bool was_busy = false;
>
> /* Lock queue and check for queue work */
> spin_lock_irqsave(&pl022->queue_lock, flags);
> if (list_empty(&pl022->queue) || !pl022->running) {
> + if (pl022->busy) {
> + /* nothing more to do - disable spi/ssp and power off */
> + writew((readw(SSP_CR1(pl022->virtbase)) &
> + (~SSP_CR1_MASK_SSE)), SSP_CR1(pl022->virtbase));
> + pm_runtime_put(&pl022->adev->dev);
> + }
Probably it will look better if we add a blank line here.
--
viresh
--
RSA(R) Conference 2012
Save $700 by Nov 18
Register now
http://p.sf.net/sfu/rsa-sfdev2dev1
___
spi-devel-general mailing list
spi-devel-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/spi-devel-general
Re: [PATCH 4/6] spi/pl022: skip default configuration before suspending
On 11/9/2011 4:09 PM, Linus WALLEIJ wrote: > From: Virupax Sadashivpetimath > > The loading of the default configuration before suspending has > been in the driver since its inception, but it is not really > needed. Especially so since we take to all the trouble of > enabling and disabling power and clock just to do this. Let's > scrap this now. > > Signed-off-by: Virupax Sadashivpetimath > > Signed-off-by: Linus Walleij > --- > drivers/spi/spi-pl022.c |7 +-- > 1 files changed, 1 insertions(+), 6 deletions(-) > > diff --git a/drivers/spi/spi-pl022.c b/drivers/spi/spi-pl022.c > index fa3eaae..bffad2a 100644 > --- a/drivers/spi/spi-pl022.c > +++ b/drivers/spi/spi-pl022.c > @@ -2314,12 +2314,7 @@ static int pl022_suspend(struct device *dev) > return status; > } > > - amba_vcore_enable(pl022->adev); > - amba_pclk_enable(pl022->adev); > - load_ssp_default_config(pl022); > - amba_pclk_disable(pl022->adev); > - amba_vcore_disable(pl022->adev); > - dev_dbg(dev, "suspended\n"); > + dev_dbg(&adev->dev, "suspended\n"); why adev->dev instead of dev? > return 0; > } > -- viresh -- RSA(R) Conference 2012 Save $700 by Nov 18 Register now http://p.sf.net/sfu/rsa-sfdev2dev1 ___ spi-devel-general mailing list spi-devel-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/spi-devel-general
Re: [PATCH 3/6] spi/pl022: disable the PL022 block when unused
On 11/9/2011 4:09 PM, Linus WALLEIJ wrote: > From: Virupax Sadashivpetimath > > Make sure we clear the enable bit when the block is not used. > This will save some energy in certain hardware versions. > > Signed-off-by: Virupax Sadashivpetimath > > Signed-off-by: Linus Walleij > --- > drivers/spi/spi-pl022.c |5 + > 1 files changed, 5 insertions(+), 0 deletions(-) > > diff --git a/drivers/spi/spi-pl022.c b/drivers/spi/spi-pl022.c > index 305f2ba..fa3eaae 100644 > --- a/drivers/spi/spi-pl022.c > +++ b/drivers/spi/spi-pl022.c > @@ -512,6 +512,11 @@ static void giveback(struct pl022 *pl022) > msg->state = NULL; > if (msg->complete) > msg->complete(msg->context); > + > + /* disable the SPI/SSP operation */ > + writew((readw(SSP_CR1(pl022->virtbase)) & > + (~SSP_CR1_MASK_SSE)), SSP_CR1(pl022->virtbase)); > + > /* This message is completed, so let's turn off the clocks & power */ > pm_runtime_put(&pl022->adev->dev); > } Reviewed-by: Viresh Kumar -- viresh -- RSA(R) Conference 2012 Save $700 by Nov 18 Register now http://p.sf.net/sfu/rsa-sfdev2dev1 ___ spi-devel-general mailing list spi-devel-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/spi-devel-general
Re: [PATCH 2/6] spi/pl022: fix build warnings
On 11/9/2011 4:08 PM, Linus WALLEIJ wrote:
> From: Jonas Aaberg
>
> The driver build complains with newer compilers unless you
> initialize this struct properly.
>
> Signed-off-by: Jonas Aaberg
> Signed-off-by: Linus Walleij
> ---
> drivers/spi/spi-pl022.c |2 +-
> 1 files changed, 1 insertions(+), 1 deletions(-)
>
> diff --git a/drivers/spi/spi-pl022.c b/drivers/spi/spi-pl022.c
> index b251926..305f2ba 100644
> --- a/drivers/spi/spi-pl022.c
> +++ b/drivers/spi/spi-pl022.c
> @@ -1885,7 +1885,7 @@ static int pl022_setup(struct spi_device *spi)
> {
> struct pl022_config_chip const *chip_info;
> struct chip_data *chip;
> - struct ssp_clock_params clk_freq = {0, };
> + struct ssp_clock_params clk_freq = { .cpsdvsr = 0, .scr = 0};
> int status = 0;
> struct pl022 *pl022 = spi_master_get_devdata(spi->master);
> unsigned int bits = spi->bits_per_word;
Reviewed-by: Viresh Kumar
--
viresh
--
RSA(R) Conference 2012
Save $700 by Nov 18
Register now
http://p.sf.net/sfu/rsa-sfdev2dev1
___
spi-devel-general mailing list
spi-devel-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/spi-devel-general
Re: [PATCH 1/6] spi/pl022: only enable RX interrupts when TX is complete
On 11/9/2011 4:08 PM, Linus WALLEIJ wrote: > + /* default is to enable all interrupts except RX - > + * this will be enabled once TX is complete > + */ Other than fixing this multiline comment, everything other looks fine. Reviewed-by: Viresh Kumar -- viresh -- RSA(R) Conference 2012 Save $700 by Nov 18 Register now http://p.sf.net/sfu/rsa-sfdev2dev1 ___ spi-devel-general mailing list spi-devel-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/spi-devel-general
Re: [PATCH 3/3] spi/pl022: skip default configuration before suspending
On 10/17/2011 6:42 PM, Linus WALLEIJ wrote: > From: Virupax Sadashivpetimath > > The loading of the default configuration before suspending has > been in the driver since its inception, but it is not really > needed. Especially so since we take to all the trouble of > enabling and disabling power and clock just to do this. Let's > scrap this now. > > Signed-off-by: Virupax Sadashivpetimath > > Signed-off-by: Linus Walleij > --- > drivers/spi/spi-pl022.c |5 - > 1 files changed, 0 insertions(+), 5 deletions(-) > > diff --git a/drivers/spi/spi-pl022.c b/drivers/spi/spi-pl022.c > index 0a1d8ed..29dc70f 100644 > --- a/drivers/spi/spi-pl022.c > +++ b/drivers/spi/spi-pl022.c > @@ -2342,11 +2342,6 @@ static int pl022_suspend(struct amba_device *adev, > pm_message_t state) > return status; > } > > - amba_vcore_enable(adev); > - amba_pclk_enable(adev); > - load_ssp_default_config(pl022); > - amba_pclk_disable(adev); > - amba_vcore_disable(adev); > dev_dbg(&adev->dev, "suspended\n"); > return 0; > } Reviewed-by: Viresh Kumar -- viresh -- All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity and more. Splunk takes this data and makes sense of it. Business sense. IT sense. Common sense. http://p.sf.net/sfu/splunk-d2d-oct ___ spi-devel-general mailing list spi-devel-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/spi-devel-general
Re: [PATCH 2/3] spi/pl022: disable the PL022 block when unused
On 10/17/2011 6:42 PM, Linus WALLEIJ wrote: > From: Virupax Sadashivpetimath > > Make sure we clear the enable bit when the block is not used. > This will save some energy in certain hardware versions. > > Signed-off-by: Virupax Sadashivpetimath > > Signed-off-by: Linus Walleij > --- > drivers/spi/spi-pl022.c |7 ++- > 1 files changed, 6 insertions(+), 1 deletions(-) > > diff --git a/drivers/spi/spi-pl022.c b/drivers/spi/spi-pl022.c > index 91700bb..0a1d8ed 100644 > --- a/drivers/spi/spi-pl022.c > +++ b/drivers/spi/spi-pl022.c > @@ -513,7 +513,12 @@ static void giveback(struct pl022 *pl022) > msg->state = NULL; > if (msg->complete) > msg->complete(msg->context); > - /* This message is completed, so let's turn off the clocks & power */ > + > + /* disable the SPI/SSP operation */ > + writew((readw(SSP_CR1(pl022->virtbase)) & > + (~SSP_CR1_MASK_SSE)), SSP_CR1(pl022->virtbase)); > + > + /* This message is completed, so let's turn off the clock! */ > clk_disable(pl022->clk); > amba_pclk_disable(pl022->adev); > amba_vcore_disable(pl022->adev); Reviewed-by: Viresh Kumar -- viresh -- All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity and more. Splunk takes this data and makes sense of it. Business sense. IT sense. Common sense. http://p.sf.net/sfu/splunk-d2d-oct ___ spi-devel-general mailing list spi-devel-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/spi-devel-general
Re: [PATCH 1/3] spi/pl022: fix build warnings
On 10/17/2011 6:42 PM, Linus WALLEIJ wrote:
> From: Jonas Aaberg
>
> The driver build complains with newer compilers unless you
> initialize this struct properly.
>
> Signed-off-by: Jonas Aaberg
> Signed-off-by: Linus Walleij
> ---
> drivers/spi/spi-pl022.c |2 +-
> 1 files changed, 1 insertions(+), 1 deletions(-)
>
> diff --git a/drivers/spi/spi-pl022.c b/drivers/spi/spi-pl022.c
> index debdaeb..91700bb 100644
> --- a/drivers/spi/spi-pl022.c
> +++ b/drivers/spi/spi-pl022.c
> @@ -1930,7 +1930,7 @@ static int pl022_setup(struct spi_device *spi)
> {
> struct pl022_config_chip const *chip_info;
> struct chip_data *chip;
> - struct ssp_clock_params clk_freq = {0, };
> + struct ssp_clock_params clk_freq = { .cpsdvsr = 0, .scr = 0};
> int status = 0;
> struct pl022 *pl022 = spi_master_get_devdata(spi->master);
> unsigned int bits = spi->bits_per_word;
Reviewed-by: Viresh Kumar
--
viresh
--
All the data continuously generated in your IT infrastructure contains a
definitive record of customers, application performance, security
threats, fraudulent activity and more. Splunk takes this data and makes
sense of it. Business sense. IT sense. Common sense.
http://p.sf.net/sfu/splunk-d2d-oct
___
spi-devel-general mailing list
spi-devel-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/spi-devel-general
Re: [PATCH V2 0/6] spi/spi-pl022 fixes
On 9/1/2011 4:26 PM, Linus Walleij wrote: > I agree. Acked-by: Linus Walleij for > all except 6/6, it's probably just as possible to keep the channel handle > taken in the driver if we modify the affected DMA driver instead. > This would be better for performance I think. Hi Grant, Any updates on this? -- viresh -- All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity and more. Splunk takes this data and makes sense of it. Business sense. IT sense. Common sense. http://p.sf.net/sfu/splunk-d2dcopy1 ___ spi-devel-general mailing list spi-devel-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/spi-devel-general
Re: [PATCH V2 0/6] spi/spi-pl022 fixes
On 8/10/2011 2:20 PM, Viresh KUMAR wrote: > Hi Grant, > > I have added Tested-by: Linus Walleij on all > patches. > > This patchset mainly covers following fixes: > - formatting related issues > - Passing GFP_ATOMIC for sg allocation from tasklet > - Fixing calculate_effective_freq() routine > - Allocate/free DMA channels as and when required. > Grant, Did you already pushed this patchset? There are few concerns on PATCH 6/6 and i am not sure if there are any decisions on it. @Linus: What do you say? So, probably you can push first 5 patches for now. -- viresh -- Special Offer -- Download ArcSight Logger for FREE! Finally, a world-class log management solution at an even better price-free! And you'll get a free "Love Thy Logs" t-shirt when you download Logger. Secure your free ArcSight Logger TODAY! http://p.sf.net/sfu/arcsisghtdev2dev ___ spi-devel-general mailing list spi-devel-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/spi-devel-general
Re: [PATCH V2 3/6] spi/spi-pl022: Don't allocate more sg than required.
On 08/10/2011 05:12 PM, Sergei Shtylyov wrote: >> > - pages = (pl022->cur_transfer->len>> PAGE_SHIFT) + 1; >> > + pages = ((pl022->cur_transfer->len + (1 << PAGE_SHIFT) - 1) >> > + >> PAGE_SHIFT); > No need for parens around the rvalue. Oops, that was a mistake. Anyway i have send V3 for the same and this issue is not present anymore. -- viresh -- uberSVN's rich system and user administration capabilities and model configuration take the hassle out of deploying and managing Subversion and the tools developers use with it. Learn more about uberSVN and get a free download at: http://p.sf.net/sfu/wandisco-dev2dev ___ spi-devel-general mailing list spi-devel-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/spi-devel-general
[PATCH V3 3/6] spi/spi-pl022: Don't allocate more sg than required.
In routine configure_dma(), if transfer->len = PAGE_SIZE, then pages is one more than required. While leads to one more sg getting allocated. This is wrong. Correct this to allocate correct number of sg. Signed-off-by: Viresh Kumar Tested-by: Linus Walleij --- drivers/spi/spi-pl022.c |2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/drivers/spi/spi-pl022.c b/drivers/spi/spi-pl022.c index 80116be..832361e3a 100644 --- a/drivers/spi/spi-pl022.c +++ b/drivers/spi/spi-pl022.c @@ -1016,7 +1016,7 @@ static int configure_dma(struct pl022 *pl022) dmaengine_slave_config(txchan, &tx_conf); /* Create sglists for the transfers */ - pages = (pl022->cur_transfer->len >> PAGE_SHIFT) + 1; + pages = DIV_ROUND_UP(pl022->cur_transfer->len, PAGE_SIZE); dev_dbg(&pl022->adev->dev, "using %d pages for transfer\n", pages); ret = sg_alloc_table(&pl022->sgt_rx, pages, GFP_ATOMIC); -- 1.7.2.2 -- uberSVN's rich system and user administration capabilities and model configuration take the hassle out of deploying and managing Subversion and the tools developers use with it. Learn more about uberSVN and get a free download at: http://p.sf.net/sfu/wandisco-dev2dev ___ spi-devel-general mailing list spi-devel-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/spi-devel-general
Re: [PATCH V2 6/6] spi/spi-pl022: Request/free DMA channels as and when required.
On 08/10/2011 03:31 PM, Koul, Vinod wrote: > And on your patch, are you able to dynamically assign the channels for > platform? What is the intended usage? (as Russell articulated it is bad > to dynamically assign channel for something like uart) Are you talking about channels or DMA request lines? For channels yes, we can always allocate channels as they are independent of peripherals. About request lines, they are muxed in our case between several peripherals, but support for that has to be added in dw_dmac. SPI is not as much heavily used as uart. So, it might be fine there to allocate channels dynamically. -- viresh -- uberSVN's rich system and user administration capabilities and model configuration take the hassle out of deploying and managing Subversion and the tools developers use with it. Learn more about uberSVN and get a free download at: http://p.sf.net/sfu/wandisco-dev2dev ___ spi-devel-general mailing list spi-devel-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/spi-devel-general
[PATCH V2 6/6] spi/spi-pl022: Request/free DMA channels as and when required.
Currently we request DMA channels at probe time and free them at remove. They
are always occupied, irrespective of their usage.
They must be allocated when they are required and must be freed after we are
done with transfers. So that they can be used by other users.
Signed-off-by: Viresh Kumar
Tested-by: Linus Walleij
---
drivers/spi/spi-pl022.c | 98 +-
1 files changed, 70 insertions(+), 28 deletions(-)
diff --git a/drivers/spi/spi-pl022.c b/drivers/spi/spi-pl022.c
index 01e84e3..a596b96 100644
--- a/drivers/spi/spi-pl022.c
+++ b/drivers/spi/spi-pl022.c
@@ -389,6 +389,7 @@ struct pl022 {
struct sg_table sgt_rx;
struct sg_table sgt_tx;
char*dummypage;
+ dma_cap_mask_t mask;
#endif
};
@@ -1093,16 +1094,33 @@ err_alloc_rx_sg:
static int __init pl022_dma_probe(struct pl022 *pl022)
{
- dma_cap_mask_t mask;
+ /* set dma mask */
+ dma_cap_zero(pl022->mask);
+ dma_cap_set(DMA_SLAVE, pl022->mask);
- /* Try to acquire a generic DMA engine slave channel */
- dma_cap_zero(mask);
- dma_cap_set(DMA_SLAVE, mask);
+ pl022->dummypage = kmalloc(PAGE_SIZE, GFP_KERNEL);
+ if (!pl022->dummypage) {
+ dev_err(&pl022->adev->dev,
+ "Failed to work in dma mode, work without dma!\n");
+ return -ENODEV;
+ }
+
+ return 0;
+}
+
+static int pl022_alloc_dmachan(struct pl022 *pl022)
+{
/*
-* We need both RX and TX channels to do DMA, else do none
-* of them.
+* Both channels should be allocated or not allocated. It is wrong if
+* only one allocated
*/
- pl022->dma_rx_channel = dma_request_channel(mask,
+ if (pl022->dma_rx_channel && pl022->dma_tx_channel)
+ return 0;
+
+ BUG_ON(pl022->dma_rx_channel || pl022->dma_tx_channel);
+
+ /* We need both RX and TX channels to do DMA, else do none of them. */
+ pl022->dma_rx_channel = dma_request_channel(pl022->mask,
pl022->master_info->dma_filter,
pl022->master_info->dma_rx_param);
if (!pl022->dma_rx_channel) {
@@ -1110,7 +1128,7 @@ static int __init pl022_dma_probe(struct pl022 *pl022)
goto err_no_rxchan;
}
- pl022->dma_tx_channel = dma_request_channel(mask,
+ pl022->dma_tx_channel = dma_request_channel(pl022->mask,
pl022->master_info->dma_filter,
pl022->master_info->dma_tx_param);
if (!pl022->dma_tx_channel) {
@@ -1118,20 +1136,12 @@ static int __init pl022_dma_probe(struct pl022 *pl022)
goto err_no_txchan;
}
- pl022->dummypage = kmalloc(PAGE_SIZE, GFP_KERNEL);
- if (!pl022->dummypage) {
- dev_dbg(&pl022->adev->dev, "no DMA dummypage!\n");
- goto err_no_dummypage;
- }
-
- dev_info(&pl022->adev->dev, "setup for DMA on RX %s, TX %s\n",
+ dev_dbg(&pl022->adev->dev, "setup for DMA on RX %s, TX %s\n",
dma_chan_name(pl022->dma_rx_channel),
dma_chan_name(pl022->dma_tx_channel));
return 0;
-err_no_dummypage:
- dma_release_channel(pl022->dma_tx_channel);
err_no_txchan:
dma_release_channel(pl022->dma_rx_channel);
pl022->dma_rx_channel = NULL;
@@ -1143,22 +1153,30 @@ err_no_rxchan:
static void terminate_dma(struct pl022 *pl022)
{
- struct dma_chan *rxchan = pl022->dma_rx_channel;
- struct dma_chan *txchan = pl022->dma_tx_channel;
-
- dmaengine_terminate_all(rxchan);
- dmaengine_terminate_all(txchan);
- unmap_free_dma_scatter(pl022);
+ if (pl022->dma_rx_channel)
+ dmaengine_terminate_all(pl022->dma_rx_channel);
+ if (pl022->dma_tx_channel)
+ dmaengine_terminate_all(pl022->dma_tx_channel);
}
-static void pl022_dma_remove(struct pl022 *pl022)
+static void pl022_free_dmachan(struct pl022 *pl022)
{
- if (pl022->busy)
- terminate_dma(pl022);
if (pl022->dma_tx_channel)
dma_release_channel(pl022->dma_tx_channel);
if (pl022->dma_rx_channel)
dma_release_channel(pl022->dma_rx_channel);
+
+ pl022->dma_rx_channel = pl022->dma_tx_channel = NULL;
+}
+
+static void pl022_dma_remove(struct pl022 *pl022)
+{
+ if (pl022->busy) {
+ terminate_dma(pl022);
+ unmap_free_dma_scatter(pl022);
+ }
+
+ pl022_free_dmachan(pl022);
kfree(pl022->dummypage);
}
Re: [PATCH V2 3/6] spi/spi-pl022: Don't allocate more sg than required.
On 08/10/2011 02:24 PM, Russell King - ARM Linux wrote: > It would be far better for this to be: > > pages = DIV_ROUND_UP(pl022->cur_transfer->len, PAGE_SIZE); > > The compiler will probably optimize it to the same code anyway. I thought of it too, but missed to update code. Thanks. -- viresh -- uberSVN's rich system and user administration capabilities and model configuration take the hassle out of deploying and managing Subversion and the tools developers use with it. Learn more about uberSVN and get a free download at: http://p.sf.net/sfu/wandisco-dev2dev ___ spi-devel-general mailing list spi-devel-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/spi-devel-general
Re: [PATCH V2 6/6] spi/spi-pl022: Request/free DMA channels as and when required.
On 08/10/2011 02:30 PM, Russell King - ARM Linux wrote: >> > They must be allocated when they are required and must be freed after we >> > are >> > done with transfers. So that they can be used by other users. > Which DMA engine driver requires this? > dw_dmac.c > Normally, when we have DMA engine drivers with multiple request signals, > the slave peripheral side publishes several virtual channels which are > claimed by the peripheral drivers. This (amongst other things) allows > the peripheral drivers to hold claim to one of the virtual channels > all the time that it's required. If users of DMA expect DMA engine drivers to work this way, then we should have this mentioned clearly in DMA slave documentation. @Dan/Vinod: What do you say? -- viresh -- uberSVN's rich system and user administration capabilities and model configuration take the hassle out of deploying and managing Subversion and the tools developers use with it. Learn more about uberSVN and get a free download at: http://p.sf.net/sfu/wandisco-dev2dev ___ spi-devel-general mailing list spi-devel-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/spi-devel-general
[PATCH V2 5/6] spi/spi-pl022: Call pl022_dma_remove(pl022) only if enable_dma is true
pl022_dma_remove() should be called only if enable_dma is true. There is no point calling it when pl022_dma_probe() is not called, which again depends on enable_dma. Signed-off-by: Viresh Kumar Tested-by: Linus Walleij --- drivers/spi/spi-pl022.c |8 ++-- 1 files changed, 6 insertions(+), 2 deletions(-) diff --git a/drivers/spi/spi-pl022.c b/drivers/spi/spi-pl022.c index d1bcc79..01e84e3 100644 --- a/drivers/spi/spi-pl022.c +++ b/drivers/spi/spi-pl022.c @@ -2238,7 +2238,9 @@ err_spi_register: err_start_queue: err_init_queue: destroy_queue(pl022); - pl022_dma_remove(pl022); + if (platform_info->enable_dma) + pl022_dma_remove(pl022); + free_irq(adev->irq[0], pl022); pm_runtime_disable(&adev->dev); err_no_irq: @@ -2266,7 +2268,9 @@ pl022_remove(struct amba_device *adev) if (destroy_queue(pl022) != 0) dev_err(&adev->dev, "queue remove failed\n"); load_ssp_default_config(pl022); - pl022_dma_remove(pl022); + if (pl022->master_info->enable_dma) + pl022_dma_remove(pl022); + free_irq(adev->irq[0], pl022); clk_disable(pl022->clk); clk_put(pl022->clk); -- 1.7.2.2 -- uberSVN's rich system and user administration capabilities and model configuration take the hassle out of deploying and managing Subversion and the tools developers use with it. Learn more about uberSVN and get a free download at: http://p.sf.net/sfu/wandisco-dev2dev ___ spi-devel-general mailing list spi-devel-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/spi-devel-general
[PATCH V2 4/6] spi/spi-pl022: calculate_effective_freq() must set rate <= requested rate
There were few issues with calculate_effective_freq() routine:
- It was returning first rate found >= requested rate. Now, if system have spi's
rate as 83 MHz, with possible prescaled rates as 83, 41.5, 20.75, 13.83 (as we
can prescale with multiples of 2). If user has given rate to be programmed as
22 MHz, then driver programmes it to 41.5 MHz. This looks to be incorrect, as
user might have given the upper limit of the device, and we are programming it
above it.
- Driver finds the first satisfying rate and programmes it, but with other
values of scr & cpsdvsr, it is possible to get more closer rate.
This patch fixes these two issues, with some reformatting inside the code. This
also creates a inline routine to calculate prescaled rate based on spi's rate,
cpsdvsr and scr.
Signed-off-by: Viresh Kumar
Tested-by: Linus Walleij
---
drivers/spi/spi-pl022.c | 102 +++---
1 files changed, 51 insertions(+), 51 deletions(-)
diff --git a/drivers/spi/spi-pl022.c b/drivers/spi/spi-pl022.c
index 1c8b9ec..d1bcc79 100644
--- a/drivers/spi/spi-pl022.c
+++ b/drivers/spi/spi-pl022.c
@@ -1791,67 +1791,67 @@ static int pl022_transfer(struct spi_device *spi,
struct spi_message *msg)
return 0;
}
-static int calculate_effective_freq(struct pl022 *pl022,
- int freq,
- struct ssp_clock_params *clk_freq)
+static inline u32 spi_rate(u32 rate, u16 cpsdvsr, u16 scr)
+{
+ return rate / (cpsdvsr * (1 + scr));
+}
+
+static int calculate_effective_freq(struct pl022 *pl022, int freq, struct
+ ssp_clock_params * clk_freq)
{
/* Lets calculate the frequency parameters */
- u16 cpsdvsr = 2;
- u16 scr = 0;
- bool freq_found = false;
- u32 rate;
- u32 max_tclk;
- u32 min_tclk;
+ u16 cpsdvsr = CPSDVR_MIN, scr = SCR_MIN;
+ u32 rate, max_tclk, min_tclk, best_freq = 0, best_cpsdvsr = 0,
+ best_scr = 0, tmp, found = 0;
rate = clk_get_rate(pl022->clk);
/* cpsdvscr = 2 & scr 0 */
- max_tclk = (rate / (CPSDVR_MIN * (1 + SCR_MIN)));
+ max_tclk = spi_rate(rate, CPSDVR_MIN, SCR_MIN);
/* cpsdvsr = 254 & scr = 255 */
- min_tclk = (rate / (CPSDVR_MAX * (1 + SCR_MAX)));
-
- if ((freq <= max_tclk) && (freq >= min_tclk)) {
- while (cpsdvsr <= CPSDVR_MAX && !freq_found) {
- while (scr <= SCR_MAX && !freq_found) {
- if ((rate /
-(cpsdvsr * (1 + scr))) > freq)
- scr += 1;
- else {
- /*
-* This bool is made true when
-* effective frequency >=
-* target frequency is found
-*/
- freq_found = true;
- if ((rate /
-(cpsdvsr * (1 + scr))) != freq) {
- if (scr == SCR_MIN) {
- cpsdvsr -= 2;
- scr = SCR_MAX;
- } else
- scr -= 1;
- }
- }
- }
- if (!freq_found) {
- cpsdvsr += 2;
- scr = SCR_MIN;
- }
- }
- if (cpsdvsr != 0) {
- dev_dbg(&pl022->adev->dev,
- "SSP Effective Frequency is %u\n",
- (rate / (cpsdvsr * (1 + scr;
- clk_freq->cpsdvsr = (u8) (cpsdvsr & 0xFF);
- clk_freq->scr = (u8) (scr & 0xFF);
- dev_dbg(&pl022->adev->dev,
- "SSP cpsdvsr = %d, scr = %d\n",
- clk_freq->cpsdvsr, clk_freq->scr);
- }
- } else {
+ min_tclk = spi_rate(rate, CPSDVR_MAX, SCR_MAX);
+
+ if (!((freq <= max_tclk) && (freq >= min_tclk))) {
dev_err(&pl022->adev->dev,
"controller data is incorrect: out of range frequency");
return -EINVAL;
}
+
+ /*
+* best_freq will give closest possible available rate (<= requested
+* freq) for all values of scr &
[PATCH V2 3/6] spi/spi-pl022: Don't allocate more sg than required.
In routine configure_dma(), if transfer->len = PAGE_SIZE, then pages is one more than required. While leads to one more sg getting allocated. This is wrong. Correct this to allocate correct number of sg. Signed-off-by: Viresh Kumar Tested-by: Linus Walleij --- drivers/spi/spi-pl022.c |3 ++- 1 files changed, 2 insertions(+), 1 deletions(-) diff --git a/drivers/spi/spi-pl022.c b/drivers/spi/spi-pl022.c index 80116be..1c8b9ec 100644 --- a/drivers/spi/spi-pl022.c +++ b/drivers/spi/spi-pl022.c @@ -1016,7 +1016,8 @@ static int configure_dma(struct pl022 *pl022) dmaengine_slave_config(txchan, &tx_conf); /* Create sglists for the transfers */ - pages = (pl022->cur_transfer->len >> PAGE_SHIFT) + 1; + pages = ((pl022->cur_transfer->len + (1 << PAGE_SHIFT) - 1) + >> PAGE_SHIFT); dev_dbg(&pl022->adev->dev, "using %d pages for transfer\n", pages); ret = sg_alloc_table(&pl022->sgt_rx, pages, GFP_ATOMIC); -- 1.7.2.2 -- uberSVN's rich system and user administration capabilities and model configuration take the hassle out of deploying and managing Subversion and the tools developers use with it. Learn more about uberSVN and get a free download at: http://p.sf.net/sfu/wandisco-dev2dev ___ spi-devel-general mailing list spi-devel-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/spi-devel-general
[PATCH V2 1/6] spi/spi-pl022: Resolve formatting issues
There were few formatting related issues in code. This patch fixes them.
Fixes include:
- Remove extra blank lines
- align code to 80 cols
- combine several lines to one line
- Replace multiple spaces with tabs
- Remove spaces before labels
Signed-off-by: Viresh Kumar
Tested-by: Linus Walleij
---
drivers/spi/spi-pl022.c | 54 +-
1 files changed, 20 insertions(+), 34 deletions(-)
diff --git a/drivers/spi/spi-pl022.c b/drivers/spi/spi-pl022.c
index 730b4a3..f600d00 100644
--- a/drivers/spi/spi-pl022.c
+++ b/drivers/spi/spi-pl022.c
@@ -113,7 +113,6 @@
#define SSP_CR0_MASK_CSS_ST(0x1FUL << 16)
#define SSP_CR0_MASK_FRF_ST(0x3UL << 21)
-
/*
* SSP Control Register 0 - SSP_CR1
*/
@@ -283,7 +282,6 @@
#define SPI_POLLING_TIMEOUT 1000
-
/*
* The type of reading going on on this chip
*/
@@ -752,7 +750,6 @@ static void readwriter(struct pl022 *pl022)
*/
}
-
/**
* next_transfer - Move to the Next transfer in the current spi message
* @pl022: SSP driver private data structure
@@ -1534,8 +1531,7 @@ static void pump_messages(struct work_struct *work)
/* Initial message state */
pl022->cur_msg->state = STATE_START;
pl022->cur_transfer = list_entry(pl022->cur_msg->transfers.next,
- struct spi_transfer,
- transfer_list);
+ struct spi_transfer, transfer_list);
/* Setup the SPI using the per chip configuration */
pl022->cur_chip = spi_get_ctldata(pl022->cur_msg->spi);
@@ -1557,7 +1553,6 @@ static void pump_messages(struct work_struct *work)
do_interrupt_dma_transfer(pl022);
}
-
static int __init init_queue(struct pl022 *pl022)
{
INIT_LIST_HEAD(&pl022->queue);
@@ -1566,8 +1561,8 @@ static int __init init_queue(struct pl022 *pl022)
pl022->running = false;
pl022->busy = false;
- tasklet_init(&pl022->pump_transfers,
- pump_transfers, (unsigned long)pl022);
+ tasklet_init(&pl022->pump_transfers, pump_transfers,
+ (unsigned long)pl022);
INIT_WORK(&pl022->pump_messages, pump_messages);
pl022->workqueue = create_singlethread_workqueue(
@@ -1578,7 +1573,6 @@ static int __init init_queue(struct pl022 *pl022)
return 0;
}
-
static int start_queue(struct pl022 *pl022)
{
unsigned long flags;
@@ -1601,7 +1595,6 @@ static int start_queue(struct pl022 *pl022)
return 0;
}
-
static int stop_queue(struct pl022 *pl022)
{
unsigned long flags;
@@ -1861,7 +1854,6 @@ static int calculate_effective_freq(struct pl022 *pl022,
return 0;
}
-
/*
* A piece of default chip info unless the platform
* supplies it.
@@ -1879,7 +1871,6 @@ static const struct pl022_config_chip
pl022_default_chip_info = {
.cs_control = null_cs_control,
};
-
/**
* pl022_setup - setup function registered to SPI master framework
* @spi: spi device which is requesting setup
@@ -1956,7 +1947,6 @@ static int pl022_setup(struct spi_device *spi)
goto err_config_params;
}
-
status = verify_controller_parameters(pl022, chip_info);
if (status) {
dev_err(&spi->dev, "controller data is incorrect");
@@ -2096,12 +2086,13 @@ static int pl022_setup(struct spi_device *spi)
}
SSP_WRITE_BITS(chip->cr1, SSP_DISABLED, SSP_CR1_MASK_SSE, 1);
SSP_WRITE_BITS(chip->cr1, chip_info->hierarchy, SSP_CR1_MASK_MS, 2);
- SSP_WRITE_BITS(chip->cr1, chip_info->slave_tx_disable,
SSP_CR1_MASK_SOD, 3);
+ SSP_WRITE_BITS(chip->cr1, chip_info->slave_tx_disable, SSP_CR1_MASK_SOD,
+ 3);
/* Save controller_state */
spi_set_ctldata(spi, chip);
return status;
- err_config_params:
+err_config_params:
spi_set_ctldata(spi, NULL);
kfree(chip);
return status;
@@ -2122,7 +2113,6 @@ static void pl022_cleanup(struct spi_device *spi)
kfree(chip);
}
-
static int __devinit
pl022_probe(struct amba_device *adev, const struct amba_id *id)
{
@@ -2243,23 +2233,23 @@ pl022_probe(struct amba_device *adev, const struct
amba_id *id)
amba_vcore_disable(adev);
return 0;
- err_spi_register:
- err_start_queue:
- err_init_queue:
+err_spi_register:
+err_start_queue:
+err_init_queue:
destroy_queue(pl022);
pl022_dma_remove(pl022);
free_irq(adev->irq[0], pl022);
pm_runtime_disable(&adev->dev);
- err_no_irq:
+err_no_irq:
clk_put(pl022->clk);
- err_no_clk:
+err_no_clk:
iounmap(pl022->virtbase);
- err_no_ioremap:
+err_no_ioremap:
amba_release_regions(adev);
- err_no_ioregion:
+err_no_ioregion:
spi_master_put(master);
- err_no_mas
[PATCH V2 0/6] spi/spi-pl022 fixes
Hi Grant, I have added Tested-by: Linus Walleij on all patches. This patchset mainly covers following fixes: - formatting related issues - Passing GFP_ATOMIC for sg allocation from tasklet - Fixing calculate_effective_freq() routine - Allocate/free DMA channels as and when required. Changes since V1: - Replace GFP_NOWAIT with GFP_ATOMIC - Wrote inline routine spi_rate instead of macro Viresh Kumar (6): spi/spi-pl022: Resolve formatting issues spi/spi-pl022: Use GFP_ATOMIC for allocation from tasklet spi/spi-pl022: Don't allocate more sg than required. spi/spi-pl022: calculate_effective_freq() must set rate <= requested rate spi/spi-pl022: Call pl022_dma_remove(pl022) only if enable_dma is true spi/spi-pl022: Request/free DMA channels as and when required. drivers/spi/spi-pl022.c | 269 ++- 1 files changed, 151 insertions(+), 118 deletions(-) -- 1.7.2.2 -- uberSVN's rich system and user administration capabilities and model configuration take the hassle out of deploying and managing Subversion and the tools developers use with it. Learn more about uberSVN and get a free download at: http://p.sf.net/sfu/wandisco-dev2dev ___ spi-devel-general mailing list spi-devel-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/spi-devel-general
[PATCH V2 2/6] spi/spi-pl022: Use GFP_ATOMIC for allocation from tasklet
tasklets don't allow invocation to sleeping routines. In configure_dma() routine, sg_alloc_table() was called with GFP_KERNEL flag and so this causes crash when called from tasklet. Replace GFP_KERNEL with GFP_ATOMIC to get this fixed. Signed-off-by: Viresh Kumar Tested-by: Linus Walleij --- drivers/spi/spi-pl022.c |4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/spi/spi-pl022.c b/drivers/spi/spi-pl022.c index f600d00..80116be 100644 --- a/drivers/spi/spi-pl022.c +++ b/drivers/spi/spi-pl022.c @@ -1019,11 +1019,11 @@ static int configure_dma(struct pl022 *pl022) pages = (pl022->cur_transfer->len >> PAGE_SHIFT) + 1; dev_dbg(&pl022->adev->dev, "using %d pages for transfer\n", pages); - ret = sg_alloc_table(&pl022->sgt_rx, pages, GFP_KERNEL); + ret = sg_alloc_table(&pl022->sgt_rx, pages, GFP_ATOMIC); if (ret) goto err_alloc_rx_sg; - ret = sg_alloc_table(&pl022->sgt_tx, pages, GFP_KERNEL); + ret = sg_alloc_table(&pl022->sgt_tx, pages, GFP_ATOMIC); if (ret) goto err_alloc_tx_sg; -- 1.7.2.2 -- uberSVN's rich system and user administration capabilities and model configuration take the hassle out of deploying and managing Subversion and the tools developers use with it. Learn more about uberSVN and get a free download at: http://p.sf.net/sfu/wandisco-dev2dev ___ spi-devel-general mailing list spi-devel-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/spi-devel-general
Re: [PATCH 0/6] spi/spi-pl022 fixes
On 08/10/2011 01:40 AM, Linus Walleij wrote: > On Tue, Aug 9, 2011 at 1:05 PM, Viresh Kumar wrote: > >> This patchset mainly covers following fixes: >> - formatting related issues >> - Passing GFP_NOWAIT for sg allocation from tasklet >> - Fixing calculate_effective_freq() routine >> - Allocate/free DMA channels as and when required. > > Excellent work Viresh, apart from the small comment on 4/6 > they all look good. > > I also tested them all on the U300, works perfectly. > > Please fixup 4/6 and respin with my: > Tested-by: Linus Walleij Thanks for testing. > > So that Grant can pick them up. > >> I have rebased them on linux-next/master over following patch: >> >> commit 9be355da3bae9feb09cdaf80c3ab560f1f0172cb >> Author: Stephen Rothwell >> Date: Tue Aug 9 13:30:19 2011 +1000 >> >>Add linux-next specific files for 20110809 > > They all work perfectly well on top of v3.1-rc1, please use > mainline kernel tags as patch base if you can. > Sure, can take care of this. But i have a doubt here, What if a patch on the same driver is applied after kernel tag. Shouldn't i rebase my patches on the latest commit available. -- viresh -- uberSVN's rich system and user administration capabilities and model configuration take the hassle out of deploying and managing Subversion and the tools developers use with it. Learn more about uberSVN and get a free download at: http://p.sf.net/sfu/wandisco-dev2dev ___ spi-devel-general mailing list spi-devel-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/spi-devel-general
Re: [PATCH 4/6] spi/spi-pl022: calculate_effective_freq() must set rate <= requested rate
[Probably you missed reply-all by mistake, so adding them again] On 08/10/2011 01:34 AM, Linus Walleij wrote: > On Tue, Aug 9, 2011 at 1:05 PM, Viresh Kumar wrote: > >> There were few issues with calculate_effective_freq() routine: >> - It was returning first rate found >= requested rate. Now, if system have >> spi's >> rate as 83 MHz, with possible prescaled rates as 83, 41.5, 20.75, 13.83 (as >> we >> can prescale with multiples of 2). If user has given rate to be programmed >> as >> 22 MHz, then driver programmes it to 41.5 MHz. This looks to be incorrect, >> as >> user might have given the upper limit of the device, and we are programming >> it >> above it. >> - Driver finds the first satisfying rate and programmes it, but with other >> values of scr & cpsdvsr, it is possible to get more closer rate. > > Good that you found this bug! > >> +#define SPI_RATE(rate, cpsdvsr, scr) (rate / (cpsdvsr * (1 + scr))) > > Can you use a static inline instead of a macro? It is often preferred. > If you do, this is Acked-by. > Not a problem. Will surely do that. -- viresh -- uberSVN's rich system and user administration capabilities and model configuration take the hassle out of deploying and managing Subversion and the tools developers use with it. Learn more about uberSVN and get a free download at: http://p.sf.net/sfu/wandisco-dev2dev ___ spi-devel-general mailing list spi-devel-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/spi-devel-general
Re: [PATCH 2/6] spi/spi-pl022: Use GFP_NOWAIT for allocation from tasklet
On 08/09/2011 04:56 PM, Jassi Brar wrote: > Since this could be called from when we actually need the transfer start, > maybe we could try harder using GFP_ATOMIC instead ? Yes. Will do it. -- viresh -- uberSVN's rich system and user administration capabilities and model configuration take the hassle out of deploying and managing Subversion and the tools developers use with it. Learn more about uberSVN and get a free download at: http://p.sf.net/sfu/wandisco-dev2dev ___ spi-devel-general mailing list spi-devel-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/spi-devel-general
[PATCH 2/6] spi/spi-pl022: Use GFP_NOWAIT for allocation from tasklet
tasklets don't allow invocation to sleeping routines. In configure_dma() routine, sg_alloc_table() was called with GFP_KERNEL flag and so this causes crash when called from tasklet. Replace GFP_KERNEL with GFP_NOWAIT to get this fixed. Signed-off-by: Viresh Kumar --- drivers/spi/spi-pl022.c |4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/spi/spi-pl022.c b/drivers/spi/spi-pl022.c index f600d00..cbd9afe 100644 --- a/drivers/spi/spi-pl022.c +++ b/drivers/spi/spi-pl022.c @@ -1019,11 +1019,11 @@ static int configure_dma(struct pl022 *pl022) pages = (pl022->cur_transfer->len >> PAGE_SHIFT) + 1; dev_dbg(&pl022->adev->dev, "using %d pages for transfer\n", pages); - ret = sg_alloc_table(&pl022->sgt_rx, pages, GFP_KERNEL); + ret = sg_alloc_table(&pl022->sgt_rx, pages, GFP_NOWAIT); if (ret) goto err_alloc_rx_sg; - ret = sg_alloc_table(&pl022->sgt_tx, pages, GFP_KERNEL); + ret = sg_alloc_table(&pl022->sgt_tx, pages, GFP_NOWAIT); if (ret) goto err_alloc_tx_sg; -- 1.7.2.2 -- uberSVN's rich system and user administration capabilities and model configuration take the hassle out of deploying and managing Subversion and the tools developers use with it. Learn more about uberSVN and get a free download at: http://p.sf.net/sfu/wandisco-dev2dev ___ spi-devel-general mailing list spi-devel-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/spi-devel-general
[PATCH 6/6] spi/spi-pl022: Request/free DMA channels as and when required.
Currently we request DMA channels at probe time and free them at remove. They
are always occupied, irrespective of their usage.
They must be allocated when they are required and must be freed after we are
done with transfers. So that they can be used by other users.
Signed-off-by: Viresh Kumar
---
drivers/spi/spi-pl022.c | 98 +-
1 files changed, 70 insertions(+), 28 deletions(-)
diff --git a/drivers/spi/spi-pl022.c b/drivers/spi/spi-pl022.c
index e2333e7..3653f00 100644
--- a/drivers/spi/spi-pl022.c
+++ b/drivers/spi/spi-pl022.c
@@ -389,6 +389,7 @@ struct pl022 {
struct sg_table sgt_rx;
struct sg_table sgt_tx;
char*dummypage;
+ dma_cap_mask_t mask;
#endif
};
@@ -1093,16 +1094,33 @@ err_alloc_rx_sg:
static int __init pl022_dma_probe(struct pl022 *pl022)
{
- dma_cap_mask_t mask;
+ /* set dma mask */
+ dma_cap_zero(pl022->mask);
+ dma_cap_set(DMA_SLAVE, pl022->mask);
- /* Try to acquire a generic DMA engine slave channel */
- dma_cap_zero(mask);
- dma_cap_set(DMA_SLAVE, mask);
+ pl022->dummypage = kmalloc(PAGE_SIZE, GFP_KERNEL);
+ if (!pl022->dummypage) {
+ dev_err(&pl022->adev->dev,
+ "Failed to work in dma mode, work without dma!\n");
+ return -ENODEV;
+ }
+
+ return 0;
+}
+
+static int pl022_alloc_dmachan(struct pl022 *pl022)
+{
/*
-* We need both RX and TX channels to do DMA, else do none
-* of them.
+* Both channels should be allocated or not allocated. It is wrong if
+* only one allocated
*/
- pl022->dma_rx_channel = dma_request_channel(mask,
+ if (pl022->dma_rx_channel && pl022->dma_tx_channel)
+ return 0;
+
+ BUG_ON(pl022->dma_rx_channel || pl022->dma_tx_channel);
+
+ /* We need both RX and TX channels to do DMA, else do none of them. */
+ pl022->dma_rx_channel = dma_request_channel(pl022->mask,
pl022->master_info->dma_filter,
pl022->master_info->dma_rx_param);
if (!pl022->dma_rx_channel) {
@@ -1110,7 +1128,7 @@ static int __init pl022_dma_probe(struct pl022 *pl022)
goto err_no_rxchan;
}
- pl022->dma_tx_channel = dma_request_channel(mask,
+ pl022->dma_tx_channel = dma_request_channel(pl022->mask,
pl022->master_info->dma_filter,
pl022->master_info->dma_tx_param);
if (!pl022->dma_tx_channel) {
@@ -1118,20 +1136,12 @@ static int __init pl022_dma_probe(struct pl022 *pl022)
goto err_no_txchan;
}
- pl022->dummypage = kmalloc(PAGE_SIZE, GFP_KERNEL);
- if (!pl022->dummypage) {
- dev_dbg(&pl022->adev->dev, "no DMA dummypage!\n");
- goto err_no_dummypage;
- }
-
- dev_info(&pl022->adev->dev, "setup for DMA on RX %s, TX %s\n",
+ dev_dbg(&pl022->adev->dev, "setup for DMA on RX %s, TX %s\n",
dma_chan_name(pl022->dma_rx_channel),
dma_chan_name(pl022->dma_tx_channel));
return 0;
-err_no_dummypage:
- dma_release_channel(pl022->dma_tx_channel);
err_no_txchan:
dma_release_channel(pl022->dma_rx_channel);
pl022->dma_rx_channel = NULL;
@@ -1143,22 +1153,30 @@ err_no_rxchan:
static void terminate_dma(struct pl022 *pl022)
{
- struct dma_chan *rxchan = pl022->dma_rx_channel;
- struct dma_chan *txchan = pl022->dma_tx_channel;
-
- dmaengine_terminate_all(rxchan);
- dmaengine_terminate_all(txchan);
- unmap_free_dma_scatter(pl022);
+ if (pl022->dma_rx_channel)
+ dmaengine_terminate_all(pl022->dma_rx_channel);
+ if (pl022->dma_tx_channel)
+ dmaengine_terminate_all(pl022->dma_tx_channel);
}
-static void pl022_dma_remove(struct pl022 *pl022)
+static void pl022_free_dmachan(struct pl022 *pl022)
{
- if (pl022->busy)
- terminate_dma(pl022);
if (pl022->dma_tx_channel)
dma_release_channel(pl022->dma_tx_channel);
if (pl022->dma_rx_channel)
dma_release_channel(pl022->dma_rx_channel);
+
+ pl022->dma_rx_channel = pl022->dma_tx_channel = NULL;
+}
+
+static void pl022_dma_remove(struct pl022 *pl022)
+{
+ if (pl022->busy) {
+ terminate_dma(pl022);
+ unmap_free_dma_scatter(pl022);
+ }
+
+ pl022_free_dmachan(pl022);
kfree(pl022->dummypage);
}
@@ -1176,6 +1194,19 @@
[PATCH 4/6] spi/spi-pl022: calculate_effective_freq() must set rate <= requested rate
There were few issues with calculate_effective_freq() routine:
- It was returning first rate found >= requested rate. Now, if system have spi's
rate as 83 MHz, with possible prescaled rates as 83, 41.5, 20.75, 13.83 (as we
can prescale with multiples of 2). If user has given rate to be programmed as
22 MHz, then driver programmes it to 41.5 MHz. This looks to be incorrect, as
user might have given the upper limit of the device, and we are programming it
above it.
- Driver finds the first satisfying rate and programmes it, but with other
values of scr & cpsdvsr, it is possible to get more closer rate.
This patch fixes these two issues, with some reformatting inside the code.
This also creates a macro to calculate prescaled rate based on spi's rate,
cpsdvsr and scr.
Signed-off-by: Viresh Kumar
---
drivers/spi/spi-pl022.c | 98 ++
1 files changed, 47 insertions(+), 51 deletions(-)
diff --git a/drivers/spi/spi-pl022.c b/drivers/spi/spi-pl022.c
index 452952b..e0cfb8c 100644
--- a/drivers/spi/spi-pl022.c
+++ b/drivers/spi/spi-pl022.c
@@ -1791,67 +1791,63 @@ static int pl022_transfer(struct spi_device *spi,
struct spi_message *msg)
return 0;
}
-static int calculate_effective_freq(struct pl022 *pl022,
- int freq,
- struct ssp_clock_params *clk_freq)
+#define SPI_RATE(rate, cpsdvsr, scr) (rate / (cpsdvsr * (1 + scr)))
+static int calculate_effective_freq(struct pl022 *pl022, int freq, struct
+ ssp_clock_params * clk_freq)
{
/* Lets calculate the frequency parameters */
- u16 cpsdvsr = 2;
- u16 scr = 0;
- bool freq_found = false;
- u32 rate;
- u32 max_tclk;
- u32 min_tclk;
+ u16 cpsdvsr = CPSDVR_MIN, scr = SCR_MIN;
+ u32 rate, max_tclk, min_tclk, best_freq = 0, best_cpsdvsr = 0,
+ best_scr = 0, tmp, found = 0;
rate = clk_get_rate(pl022->clk);
/* cpsdvscr = 2 & scr 0 */
- max_tclk = (rate / (CPSDVR_MIN * (1 + SCR_MIN)));
+ max_tclk = SPI_RATE(rate, CPSDVR_MIN, SCR_MIN);
/* cpsdvsr = 254 & scr = 255 */
- min_tclk = (rate / (CPSDVR_MAX * (1 + SCR_MAX)));
-
- if ((freq <= max_tclk) && (freq >= min_tclk)) {
- while (cpsdvsr <= CPSDVR_MAX && !freq_found) {
- while (scr <= SCR_MAX && !freq_found) {
- if ((rate /
-(cpsdvsr * (1 + scr))) > freq)
- scr += 1;
- else {
- /*
-* This bool is made true when
-* effective frequency >=
-* target frequency is found
-*/
- freq_found = true;
- if ((rate /
-(cpsdvsr * (1 + scr))) != freq) {
- if (scr == SCR_MIN) {
- cpsdvsr -= 2;
- scr = SCR_MAX;
- } else
- scr -= 1;
- }
- }
- }
- if (!freq_found) {
- cpsdvsr += 2;
- scr = SCR_MIN;
- }
- }
- if (cpsdvsr != 0) {
- dev_dbg(&pl022->adev->dev,
- "SSP Effective Frequency is %u\n",
- (rate / (cpsdvsr * (1 + scr;
- clk_freq->cpsdvsr = (u8) (cpsdvsr & 0xFF);
- clk_freq->scr = (u8) (scr & 0xFF);
- dev_dbg(&pl022->adev->dev,
- "SSP cpsdvsr = %d, scr = %d\n",
- clk_freq->cpsdvsr, clk_freq->scr);
- }
- } else {
+ min_tclk = SPI_RATE(rate, CPSDVR_MAX, SCR_MAX);
+
+ if (!((freq <= max_tclk) && (freq >= min_tclk))) {
dev_err(&pl022->adev->dev,
"controller data is incorrect: out of range frequency");
return -EINVAL;
}
+
+ /*
+* best_freq will give closest possible available rate (<= requested
+* freq) for all values of scr & cpsdvsr.
+*/
+ while ((cpsdvsr <= CPSDVR_MAX) && !found
[PATCH 3/6] spi/spi-pl022: Don't allocate more sg than required.
In routine configure_dma(), if transfer->len = PAGE_SIZE, then pages is one more than required. While leads to one more sg getting allocated. This is wrong. Correct this to allocate correct number of sg. Signed-off-by: Viresh Kumar --- drivers/spi/spi-pl022.c |3 ++- 1 files changed, 2 insertions(+), 1 deletions(-) diff --git a/drivers/spi/spi-pl022.c b/drivers/spi/spi-pl022.c index cbd9afe..452952b 100644 --- a/drivers/spi/spi-pl022.c +++ b/drivers/spi/spi-pl022.c @@ -1016,7 +1016,8 @@ static int configure_dma(struct pl022 *pl022) dmaengine_slave_config(txchan, &tx_conf); /* Create sglists for the transfers */ - pages = (pl022->cur_transfer->len >> PAGE_SHIFT) + 1; + pages = ((pl022->cur_transfer->len + (1 << PAGE_SHIFT) - 1) + >> PAGE_SHIFT); dev_dbg(&pl022->adev->dev, "using %d pages for transfer\n", pages); ret = sg_alloc_table(&pl022->sgt_rx, pages, GFP_NOWAIT); -- 1.7.2.2 -- uberSVN's rich system and user administration capabilities and model configuration take the hassle out of deploying and managing Subversion and the tools developers use with it. Learn more about uberSVN and get a free download at: http://p.sf.net/sfu/wandisco-dev2dev ___ spi-devel-general mailing list spi-devel-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/spi-devel-general
[PATCH 5/6] spi/spi-pl022: Call pl022_dma_remove(pl022) only if enable_dma is true
pl022_dma_remove() should be called only if enable_dma is true. There is no point calling it when pl022_dma_probe() is not called, which again depends on enable_dma. Signed-off-by: Viresh Kumar --- drivers/spi/spi-pl022.c |8 ++-- 1 files changed, 6 insertions(+), 2 deletions(-) diff --git a/drivers/spi/spi-pl022.c b/drivers/spi/spi-pl022.c index e0cfb8c..e2333e7 100644 --- a/drivers/spi/spi-pl022.c +++ b/drivers/spi/spi-pl022.c @@ -2234,7 +2234,9 @@ err_spi_register: err_start_queue: err_init_queue: destroy_queue(pl022); - pl022_dma_remove(pl022); + if (platform_info->enable_dma) + pl022_dma_remove(pl022); + free_irq(adev->irq[0], pl022); pm_runtime_disable(&adev->dev); err_no_irq: @@ -2262,7 +2264,9 @@ pl022_remove(struct amba_device *adev) if (destroy_queue(pl022) != 0) dev_err(&adev->dev, "queue remove failed\n"); load_ssp_default_config(pl022); - pl022_dma_remove(pl022); + if (pl022->master_info->enable_dma) + pl022_dma_remove(pl022); + free_irq(adev->irq[0], pl022); clk_disable(pl022->clk); clk_put(pl022->clk); -- 1.7.2.2 -- uberSVN's rich system and user administration capabilities and model configuration take the hassle out of deploying and managing Subversion and the tools developers use with it. Learn more about uberSVN and get a free download at: http://p.sf.net/sfu/wandisco-dev2dev ___ spi-devel-general mailing list spi-devel-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/spi-devel-general
[PATCH 0/6] spi/spi-pl022 fixes
Hi Linus, This patchset mainly covers following fixes: - formatting related issues - Passing GFP_NOWAIT for sg allocation from tasklet - Fixing calculate_effective_freq() routine - Allocate/free DMA channels as and when required. I have rebased them on linux-next/master over following patch: commit 9be355da3bae9feb09cdaf80c3ab560f1f0172cb Author: Stephen Rothwell Date: Tue Aug 9 13:30:19 2011 +1000 Add linux-next specific files for 20110809 Signed-off-by: Stephen Rothwell Looking forward to see your view on the patchset. Viresh Kumar (6): spi/spi-pl022: Resolve formatting issues spi/spi-pl022: Use GFP_NOWAIT for allocation from tasklet spi/spi-pl022: Don't allocate more sg than required. spi/spi-pl022: calculate_effective_freq() must set rate <= requested rate spi/spi-pl022: Call pl022_dma_remove(pl022) only if enable_dma is true spi/spi-pl022: Request/free DMA channels as and when required. drivers/spi/spi-pl022.c | 265 ++- 1 files changed, 147 insertions(+), 118 deletions(-) -- 1.7.2.2 -- uberSVN's rich system and user administration capabilities and model configuration take the hassle out of deploying and managing Subversion and the tools developers use with it. Learn more about uberSVN and get a free download at: http://p.sf.net/sfu/wandisco-dev2dev ___ spi-devel-general mailing list spi-devel-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/spi-devel-general
[PATCH 1/6] spi/spi-pl022: Resolve formatting issues
There were few formatting related issues in code. This patch fixes them.
Fixes include:
- Remove extra blank lines
- align code to 80 cols
- combine several lines to one line
- Replace multiple spaces with tabs
- Remove spaces before labels
Signed-off-by: Viresh Kumar
---
drivers/spi/spi-pl022.c | 54 +-
1 files changed, 20 insertions(+), 34 deletions(-)
diff --git a/drivers/spi/spi-pl022.c b/drivers/spi/spi-pl022.c
index 730b4a3..f600d00 100644
--- a/drivers/spi/spi-pl022.c
+++ b/drivers/spi/spi-pl022.c
@@ -113,7 +113,6 @@
#define SSP_CR0_MASK_CSS_ST(0x1FUL << 16)
#define SSP_CR0_MASK_FRF_ST(0x3UL << 21)
-
/*
* SSP Control Register 0 - SSP_CR1
*/
@@ -283,7 +282,6 @@
#define SPI_POLLING_TIMEOUT 1000
-
/*
* The type of reading going on on this chip
*/
@@ -752,7 +750,6 @@ static void readwriter(struct pl022 *pl022)
*/
}
-
/**
* next_transfer - Move to the Next transfer in the current spi message
* @pl022: SSP driver private data structure
@@ -1534,8 +1531,7 @@ static void pump_messages(struct work_struct *work)
/* Initial message state */
pl022->cur_msg->state = STATE_START;
pl022->cur_transfer = list_entry(pl022->cur_msg->transfers.next,
- struct spi_transfer,
- transfer_list);
+ struct spi_transfer, transfer_list);
/* Setup the SPI using the per chip configuration */
pl022->cur_chip = spi_get_ctldata(pl022->cur_msg->spi);
@@ -1557,7 +1553,6 @@ static void pump_messages(struct work_struct *work)
do_interrupt_dma_transfer(pl022);
}
-
static int __init init_queue(struct pl022 *pl022)
{
INIT_LIST_HEAD(&pl022->queue);
@@ -1566,8 +1561,8 @@ static int __init init_queue(struct pl022 *pl022)
pl022->running = false;
pl022->busy = false;
- tasklet_init(&pl022->pump_transfers,
- pump_transfers, (unsigned long)pl022);
+ tasklet_init(&pl022->pump_transfers, pump_transfers,
+ (unsigned long)pl022);
INIT_WORK(&pl022->pump_messages, pump_messages);
pl022->workqueue = create_singlethread_workqueue(
@@ -1578,7 +1573,6 @@ static int __init init_queue(struct pl022 *pl022)
return 0;
}
-
static int start_queue(struct pl022 *pl022)
{
unsigned long flags;
@@ -1601,7 +1595,6 @@ static int start_queue(struct pl022 *pl022)
return 0;
}
-
static int stop_queue(struct pl022 *pl022)
{
unsigned long flags;
@@ -1861,7 +1854,6 @@ static int calculate_effective_freq(struct pl022 *pl022,
return 0;
}
-
/*
* A piece of default chip info unless the platform
* supplies it.
@@ -1879,7 +1871,6 @@ static const struct pl022_config_chip
pl022_default_chip_info = {
.cs_control = null_cs_control,
};
-
/**
* pl022_setup - setup function registered to SPI master framework
* @spi: spi device which is requesting setup
@@ -1956,7 +1947,6 @@ static int pl022_setup(struct spi_device *spi)
goto err_config_params;
}
-
status = verify_controller_parameters(pl022, chip_info);
if (status) {
dev_err(&spi->dev, "controller data is incorrect");
@@ -2096,12 +2086,13 @@ static int pl022_setup(struct spi_device *spi)
}
SSP_WRITE_BITS(chip->cr1, SSP_DISABLED, SSP_CR1_MASK_SSE, 1);
SSP_WRITE_BITS(chip->cr1, chip_info->hierarchy, SSP_CR1_MASK_MS, 2);
- SSP_WRITE_BITS(chip->cr1, chip_info->slave_tx_disable,
SSP_CR1_MASK_SOD, 3);
+ SSP_WRITE_BITS(chip->cr1, chip_info->slave_tx_disable, SSP_CR1_MASK_SOD,
+ 3);
/* Save controller_state */
spi_set_ctldata(spi, chip);
return status;
- err_config_params:
+err_config_params:
spi_set_ctldata(spi, NULL);
kfree(chip);
return status;
@@ -2122,7 +2113,6 @@ static void pl022_cleanup(struct spi_device *spi)
kfree(chip);
}
-
static int __devinit
pl022_probe(struct amba_device *adev, const struct amba_id *id)
{
@@ -2243,23 +2233,23 @@ pl022_probe(struct amba_device *adev, const struct
amba_id *id)
amba_vcore_disable(adev);
return 0;
- err_spi_register:
- err_start_queue:
- err_init_queue:
+err_spi_register:
+err_start_queue:
+err_init_queue:
destroy_queue(pl022);
pl022_dma_remove(pl022);
free_irq(adev->irq[0], pl022);
pm_runtime_disable(&adev->dev);
- err_no_irq:
+err_no_irq:
clk_put(pl022->clk);
- err_no_clk:
+err_no_clk:
iounmap(pl022->virtbase);
- err_no_ioremap:
+err_no_ioremap:
amba_release_regions(adev);
- err_no_ioregion:
+err_no_ioregion:
spi_master_put(master);
- err_no_master:
- err_
Re: spi/spi-pl022: Query on working of DMA mode
On 08/04/2011 03:46 PM, Linus Walleij wrote: > No maybe not, my test cases are for single messages, and on production > systems we mainly use IRQ-driven traffic. I need to have these > tests added I believe :-/ > > If you fix this up it will be much appreciated! Ya I am fixing this up. I have resolved this yesterday only, with GFP_NOWAIT. But it looks there are few more issues. Let me fix everything first and then will send patchset with updates. -- viresh -- BlackBerry® DevCon Americas, Oct. 18-20, San Francisco, CA The must-attend event for mobile developers. Connect with experts. Get tools for creating Super Apps. See the latest technologies. Sessions, hands-on labs, demos & much more. Register early & save! http://p.sf.net/sfu/rim-blackberry-1 ___ spi-devel-general mailing list spi-devel-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/spi-devel-general
Re: spi/spi-pl022: Query on working of DMA mode
On 08/03/2011 08:28 PM, Linus Walleij wrote: >> > Here, allocation is requested with GFP_KERNEL flag, from a tasklet. >> > Which gives following crash logs: > What happens if you replace it with GFP_NOWAIT? > > That should work for everybody. > > Or you could remove the tasklet allover and replace it with a workqueue. > Ok. I will fix this and include in my patchset. >> > Is DMA mode working in your case? > Yes, magically :-/ But how? Don't you have any message with more than one transfer in it? -- viresh -- BlackBerry® DevCon Americas, Oct. 18-20, San Francisco, CA The must-attend event for mobile developers. Connect with experts. Get tools for creating Super Apps. See the latest technologies. Sessions, hands-on labs, demos & much more. Register early & save! http://p.sf.net/sfu/rim-blackberry-1 ___ spi-devel-general mailing list spi-devel-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/spi-devel-general
spi/spi-pl022: Query on working of DMA mode
Hi Linus, I am preparing few patches for spi-pl022 for "Allocating DMA channels as and when required". And i found something strange. When we have many transfers in a single message, following routines are called in specified order: tasklet_schedule(pump_transfers) pump_transfers(unsigned long data) configure_dma(pl022) sg_alloc_table(&pl022->sgt_rx, pages, GFP_KERNEL); Here, allocation is requested with GFP_KERNEL flag, from a tasklet. Which gives following crash logs: BUG: sleeping function called from invalid context at /data/csd_sw/spear/drives_os/vireshk/spear/kernel/linux-2.6/mm/slub.c:7 93 in_atomic(): 1, irqs_disabled(): 0, pid: 11, name: kworker/u:1 Backtrace: [<8003dabc>] (dump_backtrace+0x0/0x10c) from [<803aa4ec>] (dump_stack+0x18/0x1c) r6:0001 r5:00d0 r4:bf402200 r3:6113 [<803aa4d4>] (dump_stack+0x0/0x1c) from [<8004f330>] (__might_sleep+0xec/0x10c) [<8004f244>] (__might_sleep+0x0/0x10c) from [<800bbd2c>] (__kmalloc+0x74/0x110) [<800bbcb8>] (__kmalloc+0x0/0x110) from [<8019f69c>] (sg_kmalloc+0x2c/0x30) r8:80541b40 r7:0001 r6:0001 r5:0100 r4:bf47f1e8 r3: [<8019f670>] (sg_kmalloc+0x0/0x30) from [<8019f710>] (__sg_alloc_table+0x70/0x118) [<8019f6a0>] (__sg_alloc_table+0x0/0x118) from [<8019f7e0>] (sg_alloc_table+0x28/0x54) [<8019f7b8>] (sg_alloc_table+0x0/0x54) from [<802649a0>] (configure_dma+0x230/0x3fc) r5:bf47def8 r4:bf47f150 [<80264770>] (configure_dma+0x0/0x3fc) from [<8026509c>] (pump_transfers+0xcc/0x100) [<80264fd0>] (pump_transfers+0x0/0x100) from [<8005ecb0>] (tasklet_action+0xb0/0x160) r7:805203c4 r6:bf488000 r5:bf47f1a0 r4:bf47f19c [<8005ec00>] (tasklet_action+0x0/0x160) from [<8005f104>] (__do_softirq+0xa4/0x13c) [<8005f060>] (__do_softirq+0x0/0x13c) from [<8005f590>] (irq_exit+0x4c/0x54) [<8005f544>] (irq_exit+0x0/0x54) from [<80034094>] (asm_do_IRQ+0x94/0xd0) [<80034000>] (asm_do_IRQ+0x0/0xd0) from [<800396f4>] (__irq_svc+0x34/0xc0) Exception stack(0xbf489e90 to 0xbf489ed8) Now, i didn't get them earlier due to a mistake of mine. I have updated pl022_ssp_controller.enable_dma = 1, but didn't do pl022_config_chip.com_mode = DMA_TRANSFER So, finally DMA channels are allocated but never used, as both above must have been set. Today only I came to know of this mistake, and found it is actually not working, with above crash occurring. Is DMA mode working in your case? -- viresh -- BlackBerry® DevCon Americas, Oct. 18-20, San Francisco, CA The must-attend event for mobile developers. Connect with experts. Get tools for creating Super Apps. See the latest technologies. Sessions, hands-on labs, demos & much more. Register early & save! http://p.sf.net/sfu/rim-blackberry-1 ___ spi-devel-general mailing list spi-devel-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/spi-devel-general
Re: [PATCH resend] spi/amba-pl022: work in polling or interrupt mode if pl022_dma_probe fails
On 05/16/2011 09:40 AM, Viresh KUMAR wrote:
> If pl022_dma_probe fails, we can try to transfer data in polling or interrupt
> mode. Also, set platform_info->enable_dma to 0, so that no other code tries to
> use dma.
>
> Signed-off-by: Viresh Kumar
> Acked-by: Linus Walleij
> ---
> drivers/spi/amba-pl022.c | 11 ++-
> 1 files changed, 6 insertions(+), 5 deletions(-)
>
David,
Sorry for missing you earlier in "to" list. Can you please push this
through spi-devel tree.
--
viresh
> diff --git a/drivers/spi/amba-pl022.c b/drivers/spi/amba-pl022.c
> index dae886e..77f90de 100644
> --- a/drivers/spi/amba-pl022.c
> +++ b/drivers/spi/amba-pl022.c
> @@ -1068,7 +1068,7 @@ static int __init pl022_dma_probe(struct pl022 *pl022)
> pl022->master_info->dma_filter,
> pl022->master_info->dma_rx_param);
> if (!pl022->dma_rx_channel) {
> - dev_err(&pl022->adev->dev, "no RX DMA channel!\n");
> + dev_dbg(&pl022->adev->dev, "no RX DMA channel!\n");
> goto err_no_rxchan;
> }
>
> @@ -1076,13 +1076,13 @@ static int __init pl022_dma_probe(struct pl022 *pl022)
> pl022->master_info->dma_filter,
> pl022->master_info->dma_tx_param);
> if (!pl022->dma_tx_channel) {
> - dev_err(&pl022->adev->dev, "no TX DMA channel!\n");
> + dev_dbg(&pl022->adev->dev, "no TX DMA channel!\n");
> goto err_no_txchan;
> }
>
> pl022->dummypage = kmalloc(PAGE_SIZE, GFP_KERNEL);
> if (!pl022->dummypage) {
> - dev_err(&pl022->adev->dev, "no DMA dummypage!\n");
> + dev_dbg(&pl022->adev->dev, "no DMA dummypage!\n");
> goto err_no_dummypage;
> }
>
> @@ -1098,6 +1098,8 @@ err_no_txchan:
> dma_release_channel(pl022->dma_rx_channel);
> pl022->dma_rx_channel = NULL;
> err_no_rxchan:
> + dev_err(&pl022->adev->dev,
> + "Failed to work in dma mode, work without dma!\n");
> return -ENODEV;
> }
>
> @@ -2111,7 +2113,7 @@ pl022_probe(struct amba_device *adev, const struct
> amba_id *id)
> if (platform_info->enable_dma) {
> status = pl022_dma_probe(pl022);
> if (status != 0)
> - goto err_no_dma;
> + platform_info->enable_dma = 0;
> }
>
> /* Initialize and start queue */
> @@ -2143,7 +2145,6 @@ pl022_probe(struct amba_device *adev, const struct
> amba_id *id)
> err_init_queue:
> destroy_queue(pl022);
> pl022_dma_remove(pl022);
> - err_no_dma:
> free_irq(adev->irq[0], pl022);
> err_no_irq:
> clk_put(pl022->clk);
--
viresh
--
Achieve unprecedented app performance and reliability
What every C/C++ and Fortran developer should know.
Learn how Intel has extended the reach of its next-generation tools
to help boost performance applications - inlcuding clusters.
http://p.sf.net/sfu/intel-dev2devmay
___
spi-devel-general mailing list
spi-devel-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/spi-devel-general
Re: [PATCH resend] spi/amba-pl022: work in polling or interrupt mode if pl022_dma_probe fails
On 05/16/2011 04:34 PM, Russell King - ARM Linux wrote: > On Mon, May 16, 2011 at 09:40:10AM +0530, Viresh Kumar wrote: >> > If pl022_dma_probe fails, we can try to transfer data in polling or >> > interrupt >> > mode. Also, set platform_info->enable_dma to 0, so that no other code >> > tries to >> > use dma. >> > >> > Signed-off-by: Viresh Kumar >> > Acked-by: Linus Walleij > Shouldn't this go via the SPI people rather than my tree? This can, but i have seen amba drivers patches going through your drivers branch, so sent to tracker. Should i get them through spi-devel?? -- viresh -- Achieve unprecedented app performance and reliability What every C/C++ and Fortran developer should know. Learn how Intel has extended the reach of its next-generation tools to help boost performance applications - inlcuding clusters. http://p.sf.net/sfu/intel-dev2devmay ___ spi-devel-general mailing list spi-devel-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/spi-devel-general
Re: [QUERY] amba/pl022: DMA channel allocation always fail.
On 05/13/2011 08:01 PM, Linus Walleij wrote: > 2011/5/13 viresh kumar : > >> amba/pl022 and my dma driver dw_dmac.c, both have registered init() routines >> with subsys_initcall(). Now at bootup, spi boots up before DMA and so DMA >> channels are never available at spi probe. >> >> What should be done to solve this issue? > > If you check drivers/dma/ste_dma40.c you can see that our solution was to > simply move the DMA engine to arch_initcall(). > > Which sort of makes sense for a DMA engine. > > Does this work for dw_dmac.c? Yes, it worked here too.. -- viresh -- Achieve unprecedented app performance and reliability What every C/C++ and Fortran developer should know. Learn how Intel has extended the reach of its next-generation tools to help boost performance applications - inlcuding clusters. http://p.sf.net/sfu/intel-dev2devmay ___ spi-devel-general mailing list spi-devel-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/spi-devel-general
[PATCH resend] spi/amba-pl022: work in polling or interrupt mode if pl022_dma_probe fails
If pl022_dma_probe fails, we can try to transfer data in polling or interrupt
mode. Also, set platform_info->enable_dma to 0, so that no other code tries to
use dma.
Signed-off-by: Viresh Kumar
Acked-by: Linus Walleij
---
drivers/spi/amba-pl022.c | 11 ++-
1 files changed, 6 insertions(+), 5 deletions(-)
diff --git a/drivers/spi/amba-pl022.c b/drivers/spi/amba-pl022.c
index dae886e..77f90de 100644
--- a/drivers/spi/amba-pl022.c
+++ b/drivers/spi/amba-pl022.c
@@ -1068,7 +1068,7 @@ static int __init pl022_dma_probe(struct pl022 *pl022)
pl022->master_info->dma_filter,
pl022->master_info->dma_rx_param);
if (!pl022->dma_rx_channel) {
- dev_err(&pl022->adev->dev, "no RX DMA channel!\n");
+ dev_dbg(&pl022->adev->dev, "no RX DMA channel!\n");
goto err_no_rxchan;
}
@@ -1076,13 +1076,13 @@ static int __init pl022_dma_probe(struct pl022 *pl022)
pl022->master_info->dma_filter,
pl022->master_info->dma_tx_param);
if (!pl022->dma_tx_channel) {
- dev_err(&pl022->adev->dev, "no TX DMA channel!\n");
+ dev_dbg(&pl022->adev->dev, "no TX DMA channel!\n");
goto err_no_txchan;
}
pl022->dummypage = kmalloc(PAGE_SIZE, GFP_KERNEL);
if (!pl022->dummypage) {
- dev_err(&pl022->adev->dev, "no DMA dummypage!\n");
+ dev_dbg(&pl022->adev->dev, "no DMA dummypage!\n");
goto err_no_dummypage;
}
@@ -1098,6 +1098,8 @@ err_no_txchan:
dma_release_channel(pl022->dma_rx_channel);
pl022->dma_rx_channel = NULL;
err_no_rxchan:
+ dev_err(&pl022->adev->dev,
+ "Failed to work in dma mode, work without dma!\n");
return -ENODEV;
}
@@ -2111,7 +2113,7 @@ pl022_probe(struct amba_device *adev, const struct
amba_id *id)
if (platform_info->enable_dma) {
status = pl022_dma_probe(pl022);
if (status != 0)
- goto err_no_dma;
+ platform_info->enable_dma = 0;
}
/* Initialize and start queue */
@@ -2143,7 +2145,6 @@ pl022_probe(struct amba_device *adev, const struct
amba_id *id)
err_init_queue:
destroy_queue(pl022);
pl022_dma_remove(pl022);
- err_no_dma:
free_irq(adev->irq[0], pl022);
err_no_irq:
clk_put(pl022->clk);
--
1.7.2.2
--
Achieve unprecedented app performance and reliability
What every C/C++ and Fortran developer should know.
Learn how Intel has extended the reach of its next-generation tools
to help boost performance applications - inlcuding clusters.
http://p.sf.net/sfu/intel-dev2devmay
___
spi-devel-general mailing list
spi-devel-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/spi-devel-general
Re: [PATCH] spi/amba-pl022: work in polling or interrupt mode if pl022_dma_probe fails
On 5/13/11, Viresh Kumar wrote:
> diff --git a/drivers/spi/amba-pl022.c b/drivers/spi/amba-pl022.c
> index 08de58e..82b98b8 100644
> --- a/drivers/spi/amba-pl022.c
> +++ b/drivers/spi/amba-pl022.c
> @@ -1063,7 +1063,7 @@ static int __init pl022_dma_probe(struct pl022 *pl022)
> pl022->master_info->dma_filter,
> pl022->master_info->dma_rx_param);
> if (!pl022->dma_rx_channel) {
> - dev_err(&pl022->adev->dev, "no RX DMA channel!\n");
> + dev_warn(&pl022->adev->dev, "no RX DMA channel!\n");
Sorry, this must be dev_dbg
> goto err_no_rxchan;
> }
>
> @@ -1071,13 +1071,13 @@ static int __init pl022_dma_probe(struct pl022
> *pl022)
> pl022->master_info->dma_filter,
> pl022->master_info->dma_tx_param);
> if (!pl022->dma_tx_channel) {
> - dev_err(&pl022->adev->dev, "no TX DMA channel!\n");
> + dev_warn(&pl022->adev->dev, "no TX DMA channel!\n");
dev_dbg
> goto err_no_txchan;
> }
>
> pl022->dummypage = kmalloc(PAGE_SIZE, GFP_KERNEL);
> if (!pl022->dummypage) {
> - dev_err(&pl022->adev->dev, "no DMA dummypage!\n");
> + dev_warn(&pl022->adev->dev, "no DMA dummypage!\n");
dev_dbg
> goto err_no_dummypage;
> }
>
> @@ -1093,6 +1093,8 @@ err_no_txchan:
> dma_release_channel(pl022->dma_rx_channel);
> pl022->dma_rx_channel = NULL;
> err_no_rxchan:
> + dev_warn(&pl022->adev->dev,
dev_err
> + "Failed to work in dma mode, work without dma!\n");
> return -ENODEV;
> }
>
Will resend patch on Monday. Till that time, please see if this patch
with mentioned changes is acceptable or not.
--
viresh
--
Achieve unprecedented app performance and reliability
What every C/C++ and Fortran developer should know.
Learn how Intel has extended the reach of its next-generation tools
to help boost performance applications - inlcuding clusters.
http://p.sf.net/sfu/intel-dev2devmay
___
spi-devel-general mailing list
spi-devel-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/spi-devel-general
[QUERY] amba/pl022: DMA channel allocation always fail.
Linus, amba/pl022 and my dma driver dw_dmac.c, both have registered init() routines with subsys_initcall(). Now at bootup, spi boots up before DMA and so DMA channels are never available at spi probe. What should be done to solve this issue? One idea is allocate DMA channel as and when required, instead of allocating them at probe. But in that case too i am not sure, that this issue will be solved. As, if spi slave device is added from board file, then it will also try to transfer data as soon as spi driver is up. So that too may be called before dma_probe(). -- viresh -- Achieve unprecedented app performance and reliability What every C/C++ and Fortran developer should know. Learn how Intel has extended the reach of its next-generation tools to help boost performance applications - inlcuding clusters. http://p.sf.net/sfu/intel-dev2devmay ___ spi-devel-general mailing list spi-devel-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/spi-devel-general
[PATCH] spi/amba-pl022: work in polling or interrupt mode if pl022_dma_probe fails
If pl022_dma_probe fails, we can try to transfer data in polling or interrupt
mode. Also, set platform_info->enable_dma to 0, so that no other code tries to
use dma.
Signed-off-by: Viresh Kumar
---
drivers/spi/amba-pl022.c | 11 ++-
1 files changed, 6 insertions(+), 5 deletions(-)
diff --git a/drivers/spi/amba-pl022.c b/drivers/spi/amba-pl022.c
index 08de58e..82b98b8 100644
--- a/drivers/spi/amba-pl022.c
+++ b/drivers/spi/amba-pl022.c
@@ -1063,7 +1063,7 @@ static int __init pl022_dma_probe(struct pl022 *pl022)
pl022->master_info->dma_filter,
pl022->master_info->dma_rx_param);
if (!pl022->dma_rx_channel) {
- dev_err(&pl022->adev->dev, "no RX DMA channel!\n");
+ dev_warn(&pl022->adev->dev, "no RX DMA channel!\n");
goto err_no_rxchan;
}
@@ -1071,13 +1071,13 @@ static int __init pl022_dma_probe(struct pl022 *pl022)
pl022->master_info->dma_filter,
pl022->master_info->dma_tx_param);
if (!pl022->dma_tx_channel) {
- dev_err(&pl022->adev->dev, "no TX DMA channel!\n");
+ dev_warn(&pl022->adev->dev, "no TX DMA channel!\n");
goto err_no_txchan;
}
pl022->dummypage = kmalloc(PAGE_SIZE, GFP_KERNEL);
if (!pl022->dummypage) {
- dev_err(&pl022->adev->dev, "no DMA dummypage!\n");
+ dev_warn(&pl022->adev->dev, "no DMA dummypage!\n");
goto err_no_dummypage;
}
@@ -1093,6 +1093,8 @@ err_no_txchan:
dma_release_channel(pl022->dma_rx_channel);
pl022->dma_rx_channel = NULL;
err_no_rxchan:
+ dev_warn(&pl022->adev->dev,
+ "Failed to work in dma mode, work without dma!\n");
return -ENODEV;
}
@@ -2107,7 +2109,7 @@ pl022_probe(struct amba_device *adev, const struct
amba_id *id)
if (platform_info->enable_dma) {
status = pl022_dma_probe(pl022);
if (status != 0)
- goto err_no_dma;
+ platform_info->enable_dma = 0;
}
/* Initialize and start queue */
@@ -2143,7 +2145,6 @@ pl022_probe(struct amba_device *adev, const struct
amba_id *id)
err_init_queue:
destroy_queue(pl022);
pl022_dma_remove(pl022);
- err_no_dma:
free_irq(adev->irq[0], pl022);
err_no_irq:
clk_put(pl022->clk);
--
1.7.2.2
--
Achieve unprecedented app performance and reliability
What every C/C++ and Fortran developer should know.
Learn how Intel has extended the reach of its next-generation tools
to help boost performance applications - inlcuding clusters.
http://p.sf.net/sfu/intel-dev2devmay
___
spi-devel-general mailing list
spi-devel-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/spi-devel-general
Re: [QUERY] Behavior of spi slave memories w.r.t chip select signal.
On 05/13/2011 12:24 PM, Linus Walleij wrote: > 2011/5/13 viresh kumar : >> Linus, Jamie, >> >> Have you ever seen this kind of issue? Which spi slave memories did you used >> for testing? >> I am using standard pl0022 and m25p80 driver. Tried in all modes: polling, >> interrupt, dma. > > Not really. I'll throw in a few people on CC and see if they have some hints. > Thanks guys. Issue is solved now, data must be latched on Rising Edge and i have passed phase as falling edge. -- viresh -- Achieve unprecedented app performance and reliability What every C/C++ and Fortran developer should know. Learn how Intel has extended the reach of its next-generation tools to help boost performance applications - inlcuding clusters. http://p.sf.net/sfu/intel-dev2devmay ___ spi-devel-general mailing list spi-devel-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/spi-devel-general
Re: [QUERY] Behavior of spi slave memories w.r.t chip select signal.
On 05/11/2011 09:37 AM, viresh kumar wrote: > > Hello, > > Following is what i understood after reading m25p80 driver and spi master > drivers in drivers/spi folder. > > "chip_select signal controls start and end of transfer. For ex: if we have to > read > status reg of spi memory, then we use write_and_then_read() routine. which > writes > 0x9F in one spi transfer and writes dummy bytes and reads rx reg in other > transfer. > And these two transfers are part of single spi_message. > > Now, it is controllable to handle cs, and if we send cs_change == 0, then > chip select > is activated at start of message and deactivated at end of message, instead > at end > of every transfer. > > Which means, even if there is a delay between command and dummy bytes > received at > spi memory, current transfer will not be terminated by memory as cs is low." > > Is this correct?? > > Actually i am seeing a different behavior by some of the spi memories, like > m25p10. > If there is a delay between read_sts_reg command and dummy bytes, then > 0xFF is > returned in response. If there is no delay then transfer always passes. > Linus, Jamie, Have you ever seen this kind of issue? Which spi slave memories did you used for testing? I am using standard pl0022 and m25p80 driver. Tried in all modes: polling, interrupt, dma. -- viresh -- Achieve unprecedented app performance and reliability What every C/C++ and Fortran developer should know. Learn how Intel has extended the reach of its next-generation tools to help boost performance applications - inlcuding clusters. http://p.sf.net/sfu/intel-dev2devmay ___ spi-devel-general mailing list spi-devel-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/spi-devel-general
Re: [QUERY] Behavior of spi slave memories w.r.t chip select signal.
On 05/11/2011 12:47 PM, Jamie Iles wrote: > What SPI controller are you using? I've seen a similar issue with the > Synopsys DesignWare SPI controller where the controller manages the chip > select itself, and once the FIFO is emptied the CS goes away, but the > flash chip requires it to stay high for the whole transaction. In this > case the workaround is to use a GPIO as a chip select and there are > other controllers that do the same thing for the same reason (I think > that pxa2xx_spi is one of them). I am using amba pl022 controller and using external gpio's to control chip select signal. cs is kept low for entire message, but still i am facing this issue. -- viresh -- Achieve unprecedented app performance and reliability What every C/C++ and Fortran developer should know. Learn how Intel has extended the reach of its next-generation tools to help boost performance applications - inlcuding clusters. http://p.sf.net/sfu/intel-dev2devmay ___ spi-devel-general mailing list spi-devel-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/spi-devel-general
[QUERY] Behavior of spi slave memories w.r.t chip select signal.
Hello, Following is what i understood after reading m25p80 driver and spi master drivers in drivers/spi folder. "chip_select signal controls start and end of transfer. For ex: if we have to read status reg of spi memory, then we use write_and_then_read() routine. which writes 0x9F in one spi transfer and writes dummy bytes and reads rx reg in other transfer. And these two transfers are part of single spi_message. Now, it is controllable to handle cs, and if we send cs_change == 0, then chip select is activated at start of message and deactivated at end of message, instead at end of every transfer. Which means, even if there is a delay between command and dummy bytes received at spi memory, current transfer will not be terminated by memory as cs is low." Is this correct?? Actually i am seeing a different behavior by some of the spi memories, like m25p10. If there is a delay between read_sts_reg command and dummy bytes, then 0xFF is returned in response. If there is no delay then transfer always passes. -- viresh -- Achieve unprecedented app performance and reliability What every C/C++ and Fortran developer should know. Learn how Intel has extended the reach of its next-generation tools to help boost performance applications - inlcuding clusters. http://p.sf.net/sfu/intel-dev2devmay ___ spi-devel-general mailing list spi-devel-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/spi-devel-general
[PATCH] amba-pl022: fixing compilation warning.
clk_freq is used uninitialized in pl022_setup routine. This patch
fix compilation warning for using uninitialized variable
Signed-off-by: Viresh Kumar
---
drivers/spi/amba-pl022.c |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/drivers/spi/amba-pl022.c b/drivers/spi/amba-pl022.c
index fb3d1b3..3903ab9 100644
--- a/drivers/spi/amba-pl022.c
+++ b/drivers/spi/amba-pl022.c
@@ -1799,7 +1799,7 @@ static int pl022_setup(struct spi_device *spi)
{
struct pl022_config_chip const *chip_info;
struct chip_data *chip;
- struct ssp_clock_params clk_freq;
+ struct ssp_clock_params clk_freq = {0, };
int status = 0;
struct pl022 *pl022 = spi_master_get_devdata(spi->master);
unsigned int bits = spi->bits_per_word;
--
1.7.2.2
--
Protect Your Site and Customers from Malware Attacks
Learn about various malware tactics and how to avoid them. Understand
malware threats, the impact they can have on your business, and how you
can protect your company and customers by using code signing.
http://p.sf.net/sfu/oracle-sfdevnl
___
spi-devel-general mailing list
spi-devel-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/spi-devel-general
