[2/3] mmc: core: API for temporarily disabling auto-retuning due to errors
diff mbox series

Message ID 20190517225420.176893-3-dianders@chromium.org
State New
Headers show
Series
  • brcmfmac: sdio: Deal better w/ transmission errors waking from sleep
Related show

Commit Message

Douglas Anderson May 17, 2019, 10:54 p.m. UTC
Normally when the MMC core sees an "-EILSEQ" error returned by a host
controller then it will trigger a retuning of the card.  This is
generally a good idea.

However, if a command is expected to sometimes cause transfer errors
then these transfer errors shouldn't cause a re-tuning.  This
re-tuning will be a needless waste of time.  One example case where a
transfer is expected to cause errors is when transitioning between
sleep and active state on certain Broadcom WiFi cards.  Specifically
if the card was already transitioning between states when the command
was sent it could cause an error on the SDIO bus.

Let's add an API that the SDIO card drivers can call that will
temporarily disable the auto-tuning functionality.  Then we can add a
call to this in the Broadcom WiFi driver and any other driver that
might have similar needs.

Without this change on rk3288-veyron-minnie I periodically see this in
the logs of a machine just sitting there idle:
  dwmmc_rockchip ff0d0000.dwmmc: Successfully tuned phase to XYZ

Fixes: bd11e8bd03ca ("mmc: core: Flag re-tuning is needed on CRC errors")
Signed-off-by: Douglas Anderson <dianders@chromium.org>
---
Note that are are a whole boatload of different ways that we could
provide an API for the Broadcom WiFi SDIO driver.  This patch
illustrates one way but if maintainers feel strongly that this is too
ugly and have a better idea then I can give it a shot too.  From a
purist point of view I kinda felt that the "expect errors" really
belonged as part of the mmc_request structure, but getting it into
there meant changing a whole pile of core SD/MMC APIs.  Simply adding
it to the host seemed to match the current style better and was a less
intrusive change.

 drivers/mmc/core/core.c  | 27 +++++++++++++++++++++++++--
 include/linux/mmc/core.h |  2 ++
 include/linux/mmc/host.h |  1 +
 3 files changed, 28 insertions(+), 2 deletions(-)

Comments

Wolfram Sang May 19, 2019, 9:06 a.m. UTC | #1
> Let's add an API that the SDIO card drivers can call that will
> temporarily disable the auto-tuning functionality.  Then we can add a
> call to this in the Broadcom WiFi driver and any other driver that
> might have similar needs.

Can't you fix the WiFi driver to return something else than -EILSEQ
before calling mmc_request_done() to skip the retuning?
Arend Van Spriel May 20, 2019, 8:46 a.m. UTC | #2
On 5/19/2019 11:06 AM, Wolfram Sang wrote:
> 
>> Let's add an API that the SDIO card drivers can call that will
>> temporarily disable the auto-tuning functionality.  Then we can add a
>> call to this in the Broadcom WiFi driver and any other driver that
>> might have similar needs.
> 
> Can't you fix the WiFi driver to return something else than -EILSEQ
> before calling mmc_request_done() to skip the retuning?

Not really. mmc_request_done() is for the host controller driver so the 
wifi driver is not involved.

Regards,
Arend
Wolfram Sang May 20, 2019, 8:52 a.m. UTC | #3
On Mon, May 20, 2019 at 10:46:19AM +0200, Arend Van Spriel wrote:
> On 5/19/2019 11:06 AM, Wolfram Sang wrote:
> > 
> > > Let's add an API that the SDIO card drivers can call that will
> > > temporarily disable the auto-tuning functionality.  Then we can add a
> > > call to this in the Broadcom WiFi driver and any other driver that
> > > might have similar needs.
> > 
> > Can't you fix the WiFi driver to return something else than -EILSEQ
> > before calling mmc_request_done() to skip the retuning?
> 
> Not really. mmc_request_done() is for the host controller driver so the wifi
> driver is not involved.

Uh, right. Brown paper bag, please...
Arend Van Spriel May 26, 2019, 6:42 p.m. UTC | #4
On 5/18/2019 12:54 AM, Douglas Anderson wrote:
> Normally when the MMC core sees an "-EILSEQ" error returned by a host
> controller then it will trigger a retuning of the card.  This is
> generally a good idea.

Probably a question for Adrian, but how is this retuning scheduled. I 
recall seeing something in mmc_request_done. How about deferring the 
retuning upon a release host or is that too sdio specific.

Regards,
Arend
Adrian Hunter May 28, 2019, 10:04 a.m. UTC | #5
On 26/05/19 9:42 PM, Arend Van Spriel wrote:
> On 5/18/2019 12:54 AM, Douglas Anderson wrote:
>> Normally when the MMC core sees an "-EILSEQ" error returned by a host
>> controller then it will trigger a retuning of the card.  This is
>> generally a good idea.
> 
> Probably a question for Adrian, but how is this retuning scheduled. I recall
> seeing something in mmc_request_done. How about deferring the retuning upon
> a release host or is that too sdio specific.

Below is what I have been carrying the last 4 years.  But according to Douglas'
patch, the release would need to be further down.  See 2nd diff below.
Would that work?


diff --git a/drivers/mmc/core/sdio_io.c b/drivers/mmc/core/sdio_io.c
index 3f67fbbe0d75..7a81a503541b 100644
--- a/drivers/mmc/core/sdio_io.c
+++ b/drivers/mmc/core/sdio_io.c
@@ -16,6 +16,7 @@
 #include <linux/mmc/sdio.h>
 #include <linux/mmc/sdio_func.h>
 
+#include "host.h"
 #include "sdio_ops.h"
 #include "core.h"
 #include "card.h"
@@ -738,3 +739,15 @@ int sdio_set_host_pm_flags(struct sdio_func *func, mmc_pm_flag_t flags)
 	return 0;
 }
 EXPORT_SYMBOL_GPL(sdio_set_host_pm_flags);
+
+void sdio_retune_hold_now(struct sdio_func *func)
+{
+	mmc_retune_hold_now(func->card->host);
+}
+EXPORT_SYMBOL_GPL(sdio_retune_hold_now);
+
+void sdio_retune_release(struct sdio_func *func)
+{
+	mmc_retune_release(func->card->host);
+}
+EXPORT_SYMBOL_GPL(sdio_retune_release);
diff --git a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c
index 22b73da42822..c915c39d519f 100644
--- a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c
+++ b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c
@@ -679,6 +679,11 @@ brcmf_sdio_kso_control(struct brcmf_sdio *bus, bool on)
 	brcmf_dbg(TRACE, "Enter: on=%d\n", on);
 
 	wr_val = (on << SBSDIO_FUNC1_SLEEPCSR_KSO_SHIFT);
+
+	/* Cannot re-tune if device is asleep */
+	if (on)
+		sdio_retune_hold_now(bus->sdiodev->func1);
+
 	/* 1st KSO write goes to AOS wake up core if device is asleep  */
 	brcmf_sdiod_writeb(bus->sdiodev, SBSDIO_FUNC1_SLEEPCSR, wr_val, &err);
 
@@ -691,6 +696,7 @@ brcmf_sdio_kso_control(struct brcmf_sdio *bus, bool on)
 		return err;
 
 	if (on) {
+		sdio_retune_release(bus->sdiodev->func1);
 		/* device WAKEUP through KSO:
 		 * write bit 0 & read back until
 		 * both bits 0 (kso bit) & 1 (dev on status) are set
diff --git a/include/linux/mmc/sdio_func.h b/include/linux/mmc/sdio_func.h
index 5685805533b5..85c24b0694d7 100644
--- a/include/linux/mmc/sdio_func.h
+++ b/include/linux/mmc/sdio_func.h
@@ -171,4 +171,7 @@ extern void sdio_f0_writeb(struct sdio_func *func, unsigned char b,
 extern mmc_pm_flag_t sdio_get_host_pm_caps(struct sdio_func *func);
 extern int sdio_set_host_pm_flags(struct sdio_func *func, mmc_pm_flag_t flags);
 
+extern void sdio_retune_hold_now(struct sdio_func *func);
+extern void sdio_retune_release(struct sdio_func *func);
+
 #endif /* LINUX_MMC_SDIO_FUNC_H */





diff --git a/drivers/mmc/core/sdio_io.c b/drivers/mmc/core/sdio_io.c
index 3f67fbbe0d75..7a81a503541b 100644
--- a/drivers/mmc/core/sdio_io.c
+++ b/drivers/mmc/core/sdio_io.c
@@ -16,6 +16,7 @@
 #include <linux/mmc/sdio.h>
 #include <linux/mmc/sdio_func.h>
 
+#include "host.h"
 #include "sdio_ops.h"
 #include "core.h"
 #include "card.h"
@@ -738,3 +739,15 @@ int sdio_set_host_pm_flags(struct sdio_func *func, mmc_pm_flag_t flags)
 	return 0;
 }
 EXPORT_SYMBOL_GPL(sdio_set_host_pm_flags);
+
+void sdio_retune_hold_now(struct sdio_func *func)
+{
+	mmc_retune_hold_now(func->card->host);
+}
+EXPORT_SYMBOL_GPL(sdio_retune_hold_now);
+
+void sdio_retune_release(struct sdio_func *func)
+{
+	mmc_retune_release(func->card->host);
+}
+EXPORT_SYMBOL_GPL(sdio_retune_release);
diff --git a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c
index 22b73da42822..50c153932683 100644
--- a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c
+++ b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c
@@ -679,6 +679,11 @@ brcmf_sdio_kso_control(struct brcmf_sdio *bus, bool on)
 	brcmf_dbg(TRACE, "Enter: on=%d\n", on);
 
 	wr_val = (on << SBSDIO_FUNC1_SLEEPCSR_KSO_SHIFT);
+
+	/* Cannot re-tune if device is asleep */
+	if (on)
+		sdio_retune_hold_now(bus->sdiodev->func1);
+
 	/* 1st KSO write goes to AOS wake up core if device is asleep  */
 	brcmf_sdiod_writeb(bus->sdiodev, SBSDIO_FUNC1_SLEEPCSR, wr_val, &err);
 
@@ -731,6 +736,9 @@ brcmf_sdio_kso_control(struct brcmf_sdio *bus, bool on)
 
 	} while (try_cnt++ < MAX_KSO_ATTEMPTS);
 
+	if (on)
+		sdio_retune_release(bus->sdiodev->func1);
+
 	if (try_cnt > 2)
 		brcmf_dbg(SDIO, "try_cnt=%d rd_val=0x%x err=%d\n", try_cnt,
 			  rd_val, err);
diff --git a/include/linux/mmc/sdio_func.h b/include/linux/mmc/sdio_func.h
index 5685805533b5..85c24b0694d7 100644
--- a/include/linux/mmc/sdio_func.h
+++ b/include/linux/mmc/sdio_func.h
@@ -171,4 +171,7 @@ extern void sdio_f0_writeb(struct sdio_func *func, unsigned char b,
 extern mmc_pm_flag_t sdio_get_host_pm_caps(struct sdio_func *func);
 extern int sdio_set_host_pm_flags(struct sdio_func *func, mmc_pm_flag_t flags);
 
+extern void sdio_retune_hold_now(struct sdio_func *func);
+extern void sdio_retune_release(struct sdio_func *func);
+
 #endif /* LINUX_MMC_SDIO_FUNC_H */
Arend Van Spriel May 28, 2019, 11:21 a.m. UTC | #6
On 5/28/2019 12:04 PM, Adrian Hunter wrote:
> On 26/05/19 9:42 PM, Arend Van Spriel wrote:
>> On 5/18/2019 12:54 AM, Douglas Anderson wrote:
>>> Normally when the MMC core sees an "-EILSEQ" error returned by a host
>>> controller then it will trigger a retuning of the card.  This is
>>> generally a good idea.
>>
>> Probably a question for Adrian, but how is this retuning scheduled. I recall
>> seeing something in mmc_request_done. How about deferring the retuning upon
>> a release host or is that too sdio specific.
> 
> Below is what I have been carrying the last 4 years.  But according to Douglas'
> patch, the release would need to be further down.  See 2nd diff below.
> Would that work?

That makes sense. The loop is needed because the device can be a bit 
bone headed. So indeed after the loop the device should be awake and 
able to handle CMD19.

Regards,
Arend
Adrian Hunter May 28, 2019, 11:45 a.m. UTC | #7
On 28/05/19 2:21 PM, Arend Van Spriel wrote:
> 
> 
> On 5/28/2019 12:04 PM, Adrian Hunter wrote:
>> On 26/05/19 9:42 PM, Arend Van Spriel wrote:
>>> On 5/18/2019 12:54 AM, Douglas Anderson wrote:
>>>> Normally when the MMC core sees an "-EILSEQ" error returned by a host
>>>> controller then it will trigger a retuning of the card.  This is
>>>> generally a good idea.
>>>
>>> Probably a question for Adrian, but how is this retuning scheduled. I recall
>>> seeing something in mmc_request_done. How about deferring the retuning upon
>>> a release host or is that too sdio specific.
>>
>> Below is what I have been carrying the last 4 years.  But according to
>> Douglas'
>> patch, the release would need to be further down.  See 2nd diff below.
>> Would that work?
> 
> That makes sense. The loop is needed because the device can be a bit bone
> headed. So indeed after the loop the device should be awake and able to
> handle CMD19.

What if tuning is needed to read SBSDIO_FUNC1_SLEEPCSR successfully?
Douglas Anderson May 28, 2019, 3:42 p.m. UTC | #8
Hi,

On Tue, May 28, 2019 at 4:45 AM Adrian Hunter <adrian.hunter@intel.com> wrote:
>
> On 28/05/19 2:21 PM, Arend Van Spriel wrote:
> >
> >
> > On 5/28/2019 12:04 PM, Adrian Hunter wrote:
> >> On 26/05/19 9:42 PM, Arend Van Spriel wrote:
> >>> On 5/18/2019 12:54 AM, Douglas Anderson wrote:
> >>>> Normally when the MMC core sees an "-EILSEQ" error returned by a host
> >>>> controller then it will trigger a retuning of the card.  This is
> >>>> generally a good idea.
> >>>
> >>> Probably a question for Adrian, but how is this retuning scheduled. I recall
> >>> seeing something in mmc_request_done. How about deferring the retuning upon
> >>> a release host or is that too sdio specific.
> >>
> >> Below is what I have been carrying the last 4 years.  But according to
> >> Douglas'
> >> patch, the release would need to be further down.  See 2nd diff below.
> >> Would that work?
> >
> > That makes sense. The loop is needed because the device can be a bit bone
> > headed. So indeed after the loop the device should be awake and able to
> > handle CMD19.

IMO I'd rather not _defer_ the re-tuning.  I believe the correct thing
is to not schedule the re-tuning at all in response to the IO errors.
That's what my patch does.

Specifically the IO errors that come up in this case are not due to
being "out of tune".  They are due to the fact that the other SDIO
card may be in a transitory state and putting garbage on the bus.
Scheduling a retuning for later would be just a waste of time and
needlessly tie up the bus for the retune.

...or am I confused?


> What if tuning is needed to read SBSDIO_FUNC1_SLEEPCSR successfully?

Personally I think this would be pretty unlikely.  If we're at this
point we've already talked to the card quite a bit so we should have a
tuning that's pretty good.  If we're just slightly out of tune then we
should still get through the loop with a few retries and then we can
detect that we're out of tune later, with a more reliable command.

However, if you're worried about this, I can always re-enable the old
behavior if we have already looped a few times.  I suppose I could
increase the loop duration/count slightly too...  Please let me know
one way or the other.


-Doug

Patch
diff mbox series

diff --git a/drivers/mmc/core/core.c b/drivers/mmc/core/core.c
index 6db36dc870b5..ba4f71aa8cd9 100644
--- a/drivers/mmc/core/core.c
+++ b/drivers/mmc/core/core.c
@@ -144,8 +144,9 @@  void mmc_request_done(struct mmc_host *host, struct mmc_request *mrq)
 	int err = cmd->error;
 
 	/* Flag re-tuning needed on CRC errors */
-	if ((cmd->opcode != MMC_SEND_TUNING_BLOCK &&
-	    cmd->opcode != MMC_SEND_TUNING_BLOCK_HS200) &&
+	if (cmd->opcode != MMC_SEND_TUNING_BLOCK &&
+	    cmd->opcode != MMC_SEND_TUNING_BLOCK_HS200 &&
+	    !host->expect_errors &&
 	    (err == -EILSEQ || (mrq->sbc && mrq->sbc->error == -EILSEQ) ||
 	    (mrq->data && mrq->data->error == -EILSEQ) ||
 	    (mrq->stop && mrq->stop->error == -EILSEQ)))
@@ -2163,6 +2164,28 @@  int mmc_sw_reset(struct mmc_host *host)
 }
 EXPORT_SYMBOL(mmc_sw_reset);
 
+void mmc_expect_errors_begin(struct mmc_host *host)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&host->lock, flags);
+	WARN_ON(host->expect_errors);
+	host->expect_errors = true;
+	spin_unlock_irqrestore(&host->lock, flags);
+}
+EXPORT_SYMBOL_GPL(mmc_expect_errors_begin);
+
+void mmc_expect_errors_end(struct mmc_host *host)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&host->lock, flags);
+	WARN_ON(!host->expect_errors);
+	host->expect_errors = false;
+	spin_unlock_irqrestore(&host->lock, flags);
+}
+EXPORT_SYMBOL_GPL(mmc_expect_errors_end);
+
 static int mmc_rescan_try_freq(struct mmc_host *host, unsigned freq)
 {
 	host->f_init = freq;
diff --git a/include/linux/mmc/core.h b/include/linux/mmc/core.h
index 134a6483347a..02a13abf0cda 100644
--- a/include/linux/mmc/core.h
+++ b/include/linux/mmc/core.h
@@ -178,6 +178,8 @@  int mmc_wait_for_cmd(struct mmc_host *host, struct mmc_command *cmd,
 
 int mmc_hw_reset(struct mmc_host *host);
 int mmc_sw_reset(struct mmc_host *host);
+void mmc_expect_errors_begin(struct mmc_host *host);
+void mmc_expect_errors_end(struct mmc_host *host);
 void mmc_set_data_timeout(struct mmc_data *data, const struct mmc_card *card);
 
 #endif /* LINUX_MMC_CORE_H */
diff --git a/include/linux/mmc/host.h b/include/linux/mmc/host.h
index 43d0f0c496f6..8d553fb8c834 100644
--- a/include/linux/mmc/host.h
+++ b/include/linux/mmc/host.h
@@ -398,6 +398,7 @@  struct mmc_host {
 	unsigned int		retune_now:1;	/* do re-tuning at next req */
 	unsigned int		retune_paused:1; /* re-tuning is temporarily disabled */
 	unsigned int		use_blk_mq:1;	/* use blk-mq */
+	unsigned int		expect_errors:1; /* don't trigger retune upon errors */
 
 	int			rescan_disable;	/* disable card detection */
 	int			rescan_entered;	/* used with nonremovable devices */