This series attempts to deal better with the expected transmission
errors related to the idle states (handled by the Always-On-Subsystem
or AOS) on the SDIO-based WiFi on rk3288-veyron-minnie,
rk3288-veyron-speedy, and rk3288-veyron-mickey.
Some details about those errors can be found in
<https://crbug.com/960222>, but to summarize it here: if we try to
send the wakeup command to the WiFi card at the same time it has
decided to wake up itself then it will behave badly on the SDIO bus.
This can cause timeouts or CRC errors.
When I tested on 4.19 and 4.20 these CRC errors can be seen to cause
re-tuning. Since I am currently developing on 4.19 this was the
original problem I attempted to solve.
On mainline it turns out that you don't see the retuning errors but
you see tons of spam about timeouts trying to wakeup from sleep. I
tracked down the commit that was causing that and have partially
reverted it here. I have no real knowledge about Broadcom WiFi, but
the commit that was causing problems sounds (from the descriptioin) to
be a hack commit penalizing all Broadcom WiFi users because of a bug
in a Cypress SD controller. I will let others comment if this is
truly the case and, if so, what the right solution should be.
For v3 of this series I have added 2 patches to the end of the series
to address errors that would show up on systems with these same SDIO
WiFi cards when used on controllers that do periodic retuning. These
systems need an extra fix to prevent the retuning from happening when
the card is asleep.
Changes in v4:
- Moved to SDIO API only (Adrian, Ulf).
- Renamed to make it less generic, now retune_crc_disable (Ulf).
- Function header makes it clear host must be claimed (Ulf).
- No more WARN_ON (Ulf).
- Adjust to API rename (Adrian, Ulf).
- Moved retune hold/release to SDIO API (Adrian).
- Adjust to API rename (Adrian).
Changes in v3:
- Took out the spinlock since I believe this is all in one context.
- Expect errors for all of brcmf_sdio_kso_control() (Adrian).
- ("mmc: core: Export mmc_retune_hold_now() mmc_retune_release()") new for v3.
- ("brcmfmac: sdio: Don't tune while the card is off") new for v3.
Changes in v2:
- A full revert, not just a partial one (Arend). ...with explicit Cc.
- Updated commit message to clarify based on discussion of v1.
Douglas Anderson (5):
Revert "brcmfmac: disable command decode in sdio_aos"
mmc: core: API to temporarily disable retuning for SDIO CRC errors
brcmfmac: sdio: Disable auto-tuning around commands expected to fail
mmc: core: Add sdio_retune_hold_now() and sdio_retune_release()
brcmfmac: sdio: Don't tune while the card is off
drivers/mmc/core/core.c | 5 +-
drivers/mmc/core/sdio_io.c | 76 +++++++++++++++++++
.../broadcom/brcm80211/brcmfmac/sdio.c | 17 +++--
include/linux/mmc/core.h | 2 +
include/linux/mmc/host.h | 1 +
include/linux/mmc/sdio_func.h | 6 ++
6 files changed, 100 insertions(+), 7 deletions(-)
--
2.22.0.rc2.383.gf4fbbf30c2-goog
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
idle (sometimes referred to as "sleep" in Broadcom code) and active
state on certain Broadcom WiFi SDIO 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 function 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.
NOTE: this makes the assumption that the card is already tuned well
enough that it's OK to disable the auto-retuning during one of these
error-prone situations. Presumably the driver code performing the
error-prone transfer knows how to recover / retry from errors. ...and
after we can get back to a state where transfers are no longer
error-prone then we can enable the auto-retuning again. If we truly
find ourselves in a case where the card needs to be retuned sometimes
to handle one of these error-prone transfers then we can always try a
few transfers first without auto-retuning and then re-try with
auto-retuning if the first few fail.
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 <[email protected]>
---
Changes in v4:
- Moved to SDIO API only (Adrian, Ulf).
- Renamed to make it less generic, now retune_crc_disable (Ulf).
- Function header makes it clear host must be claimed (Ulf).
- No more WARN_ON (Ulf).
Changes in v3:
- Took out the spinlock since I believe this is all in one context.
Changes in v2:
- Updated commit message to clarify based on discussion of v1.
drivers/mmc/core/core.c | 5 +++--
drivers/mmc/core/sdio_io.c | 36 +++++++++++++++++++++++++++++++++++
include/linux/mmc/core.h | 2 ++
include/linux/mmc/host.h | 1 +
include/linux/mmc/sdio_func.h | 3 +++
5 files changed, 45 insertions(+), 2 deletions(-)
diff --git a/drivers/mmc/core/core.c b/drivers/mmc/core/core.c
index 6db36dc870b5..9020cb2490f7 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->retune_crc_disable &&
(err == -EILSEQ || (mrq->sbc && mrq->sbc->error == -EILSEQ) ||
(mrq->data && mrq->data->error == -EILSEQ) ||
(mrq->stop && mrq->stop->error == -EILSEQ)))
diff --git a/drivers/mmc/core/sdio_io.c b/drivers/mmc/core/sdio_io.c
index f79f0b0caab8..f822a9630b0e 100644
--- a/drivers/mmc/core/sdio_io.c
+++ b/drivers/mmc/core/sdio_io.c
@@ -734,3 +734,39 @@ 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);
+
+/**
+ * sdio_retune_crc_disable - temporarily disable retuning on CRC errors
+ * @func: SDIO function attached to host
+ *
+ * If the SDIO card is known to be in a state where it might produce
+ * CRC errors on the bus in response to commands (like if we know it is
+ * transitioning between power states), an SDIO function driver can
+ * call this function to temporarily disable the SD/MMC core behavior of
+ * triggering an automatic retuning.
+ *
+ * This function should be called while the host is claimed and the host
+ * should remain claimed until sdio_retune_crc_enable() is called.
+ * Specifically, the expected sequence of calls is:
+ * - sdio_claim_host()
+ * - sdio_retune_crc_disable()
+ * - some number of calls like sdio_writeb() and sdio_readb()
+ * - sdio_release_host()
+ */
+void sdio_retune_crc_disable(struct sdio_func *func)
+{
+ func->card->host->retune_crc_disable = true;
+}
+EXPORT_SYMBOL_GPL(sdio_retune_crc_disable);
+
+/**
+ * sdio_retune_crc_enable - reneable retuning on CRC errors
+ * @func: SDIO function attached to host
+ *
+ * This is the compement to sdio_retune_crc_disable().
+ */
+void sdio_retune_crc_enable(struct sdio_func *func)
+{
+ func->card->host->retune_crc_disable = false;
+}
+EXPORT_SYMBOL_GPL(sdio_retune_crc_enable);
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..ecb7972e2423 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 retune_crc_disable:1; /* don't trigger retune upon crc */
int rescan_disable; /* disable card detection */
int rescan_entered; /* used with nonremovable devices */
diff --git a/include/linux/mmc/sdio_func.h b/include/linux/mmc/sdio_func.h
index e9dfdd501cd1..4820e6d09dac 100644
--- a/include/linux/mmc/sdio_func.h
+++ b/include/linux/mmc/sdio_func.h
@@ -167,4 +167,7 @@ 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_crc_disable(struct sdio_func *func);
+extern void sdio_retune_crc_enable(struct sdio_func *func);
+
#endif /* LINUX_MMC_SDIO_FUNC_H */
--
2.22.0.rc2.383.gf4fbbf30c2-goog
When Broadcom SDIO cards are idled they go to sleep and a whole
separate subsystem takes over their SDIO communication. This is the
Always-On-Subsystem (AOS) and it can't handle tuning requests.
Specifically, as tested on rk3288-veyron-minnie (which reports having
BCM4354/1 in dmesg), if I force a retune in brcmf_sdio_kso_control()
when "on = 1" (aka we're transition from sleep to wake) by whacking:
bus->sdiodev->func1->card->host->need_retune = 1
...then I can often see tuning fail. In this case dw_mmc reports "All
phases bad!"). Note that I don't get 100% failure, presumably because
sometimes the card itself has already transitioned away from the AOS
itself by the time we try to wake it up. If I force retuning when "on
= 0" (AKA force retuning right before sending the command to go to
sleep) then retuning is always OK.
NOTE: we need _both_ this patch and the patch to avoid triggering
tuning due to CRC errors in the sleep/wake transition, AKA ("brcmfmac:
sdio: Disable auto-tuning around commands expected to fail"). Though
both patches handle issues with Broadcom's AOS, the problems are
distinct:
1. We want to defer (but not ignore) asynchronous (like
timer-requested) tuning requests till the card is awake. However,
we want to ignore CRC errors during the transition, we don't want
to queue deferred tuning request.
2. You could imagine that the AOS could implement retuning but we
could still get errors while transitioning in and out of the AOS.
Similarly you could imagine a seamless transition into and out of
the AOS (with no CRC errors) even if the AOS couldn't handle
tuning.
ALSO NOTE: presumably there is never a desperate need to retune in
order to wake up the card, since doing so is impossible. Luckily the
only way the card can get into sleep state is if we had a good enough
tuning to send it a sleep command, so presumably that "good enough"
tuning is enough to wake us up, at least with a few retries.
Signed-off-by: Douglas Anderson <[email protected]>
---
Changes in v4:
- Adjust to API rename (Adrian).
Changes in v3:
- ("brcmfmac: sdio: Don't tune while the card is off") new for v3.
Changes in v2: None
drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c | 7 +++++++
1 file changed, 7 insertions(+)
diff --git a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c
index ee76593259a7..629140b6d7e2 100644
--- a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c
+++ b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c
@@ -669,6 +669,10 @@ brcmf_sdio_kso_control(struct brcmf_sdio *bus, bool on)
sdio_retune_crc_disable(bus->sdiodev->func1);
+ /* Cannot re-tune if device is asleep; defer till we're awake */
+ if (on)
+ sdio_retune_hold_now(bus->sdiodev->func1);
+
wr_val = (on << SBSDIO_FUNC1_SLEEPCSR_KSO_SHIFT);
/* 1st KSO write goes to AOS wake up core if device is asleep */
brcmf_sdiod_writeb(bus->sdiodev, SBSDIO_FUNC1_SLEEPCSR, wr_val, &err);
@@ -729,6 +733,9 @@ brcmf_sdio_kso_control(struct brcmf_sdio *bus, bool on)
if (try_cnt > MAX_KSO_ATTEMPTS)
brcmf_err("max tries: rd_val=0x%x err=%d\n", rd_val, err);
+ if (on)
+ sdio_retune_release(bus->sdiodev->func1);
+
sdio_retune_crc_enable(bus->sdiodev->func1);
return err;
--
2.22.0.rc2.383.gf4fbbf30c2-goog
There are certain cases, notably when transitioning between sleep and
active state, when Broadcom SDIO WiFi cards will produce errors on the
SDIO bus. This is evident from the source code where you can see that
we try commands in a loop until we either get success or we've tried
too many times. The comment in the code reinforces this by saying
"just one write attempt may fail"
Unfortunately these failures sometimes end up causing an "-EILSEQ"
back to the core which triggers a retuning of the SDIO card and that
blocks all traffic to the card until it's done.
Let's disable retuning around the commands we expect might fail.
Fixes: bd11e8bd03ca ("mmc: core: Flag re-tuning is needed on CRC errors")
Signed-off-by: Douglas Anderson <[email protected]>
---
Changes in v4:
- Adjust to API rename (Adrian, Ulf).
Changes in v3:
- Expect errors for all of brcmf_sdio_kso_control() (Adrian).
Changes in v2: None
drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c
index 4a750838d8cd..ee76593259a7 100644
--- a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c
+++ b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c
@@ -667,6 +667,8 @@ brcmf_sdio_kso_control(struct brcmf_sdio *bus, bool on)
brcmf_dbg(TRACE, "Enter: on=%d\n", on);
+ sdio_retune_crc_disable(bus->sdiodev->func1);
+
wr_val = (on << SBSDIO_FUNC1_SLEEPCSR_KSO_SHIFT);
/* 1st KSO write goes to AOS wake up core if device is asleep */
brcmf_sdiod_writeb(bus->sdiodev, SBSDIO_FUNC1_SLEEPCSR, wr_val, &err);
@@ -727,6 +729,8 @@ brcmf_sdio_kso_control(struct brcmf_sdio *bus, bool on)
if (try_cnt > MAX_KSO_ATTEMPTS)
brcmf_err("max tries: rd_val=0x%x err=%d\n", rd_val, err);
+ sdio_retune_crc_enable(bus->sdiodev->func1);
+
return err;
}
--
2.22.0.rc2.383.gf4fbbf30c2-goog
This reverts commit 29f6589140a10ece8c1d73f58043ea5b3473ab3e.
After that patch landed I find that my kernel log on
rk3288-veyron-minnie and rk3288-veyron-speedy is filled with:
brcmfmac: brcmf_sdio_bus_sleep: error while changing bus sleep state -110
This seems to happen every time the Broadcom WiFi transitions out of
sleep mode. Reverting the commit fixes the problem for me, so that's
what this patch does.
Note that, in general, the justification in the original commit seemed
a little weak. It looked like someone was testing on a SD card
controller that would sometimes die if there were CRC errors on the
bus. This used to happen back in early days of dw_mmc (the controller
on my boards), but we fixed it. Disabling a feature on all boards
just because one SD card controller is broken seems bad.
Fixes: 29f6589140a1 ("brcmfmac: disable command decode in sdio_aos")
Cc: Wright Feng <[email protected]>
Cc: Double Lo <[email protected]>
Cc: Madhan Mohan R <[email protected]>
Cc: Chi-Hsien Lin <[email protected]>
Signed-off-by: Douglas Anderson <[email protected]>
---
As far as I know this patch can land anytime.
Changes in v4: None
Changes in v3: None
Changes in v2:
- A full revert, not just a partial one (Arend). ...with explicit Cc.
drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c | 6 +-----
1 file changed, 1 insertion(+), 5 deletions(-)
diff --git a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c
index 4e15ea57d4f5..4a750838d8cd 100644
--- a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c
+++ b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c
@@ -3364,11 +3364,7 @@ static int brcmf_sdio_download_firmware(struct brcmf_sdio *bus,
static bool brcmf_sdio_aos_no_decode(struct brcmf_sdio *bus)
{
- if (bus->ci->chip == CY_CC_43012_CHIP_ID ||
- bus->ci->chip == CY_CC_4373_CHIP_ID ||
- bus->ci->chip == BRCM_CC_4339_CHIP_ID ||
- bus->ci->chip == BRCM_CC_4345_CHIP_ID ||
- bus->ci->chip == BRCM_CC_4354_CHIP_ID)
+ if (bus->ci->chip == CY_CC_43012_CHIP_ID)
return true;
else
return false;
--
2.22.0.rc2.383.gf4fbbf30c2-goog
[...]
> > > --- a/include/linux/mmc/sdio_func.h
> > > +++ b/include/linux/mmc/sdio_func.h
> > > @@ -167,4 +167,7 @@ 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_crc_disable(struct sdio_func *func);
> > > +extern void sdio_retune_crc_enable(struct sdio_func *func);
> > > +
> > > #endif /* LINUX_MMC_SDIO_FUNC_H */
> > > --
> > > 2.22.0.rc2.383.gf4fbbf30c2-goog
> > >
> >
> > Besides the minor comments, this looks good to me.
>
> Thank you for the reviews!
>
> I'll plan to send a v5 on my Monday with the fixes assuming no new
> heated discussion starts up. If it's less work for you, I'm also
> happy if you just want to make the trivial fixes yourself when
> applying.
It really doesn't matter to me, feel free to pick the option you
prefer. At this point I am just awaiting acks for the broadcom patces
and some clarity of what stable releases we should target for
non-tagged patches.
Kind regards
Uffe
On 14/06/19 2:41 AM, Douglas Anderson wrote:
> When Broadcom SDIO cards are idled they go to sleep and a whole
> separate subsystem takes over their SDIO communication. This is the
> Always-On-Subsystem (AOS) and it can't handle tuning requests.
>
> Specifically, as tested on rk3288-veyron-minnie (which reports having
> BCM4354/1 in dmesg), if I force a retune in brcmf_sdio_kso_control()
> when "on = 1" (aka we're transition from sleep to wake) by whacking:
> bus->sdiodev->func1->card->host->need_retune = 1
> ...then I can often see tuning fail. In this case dw_mmc reports "All
> phases bad!"). Note that I don't get 100% failure, presumably because
> sometimes the card itself has already transitioned away from the AOS
> itself by the time we try to wake it up. If I force retuning when "on
> = 0" (AKA force retuning right before sending the command to go to
> sleep) then retuning is always OK.
>
> NOTE: we need _both_ this patch and the patch to avoid triggering
> tuning due to CRC errors in the sleep/wake transition, AKA ("brcmfmac:
> sdio: Disable auto-tuning around commands expected to fail"). Though
> both patches handle issues with Broadcom's AOS, the problems are
> distinct:
> 1. We want to defer (but not ignore) asynchronous (like
> timer-requested) tuning requests till the card is awake. However,
> we want to ignore CRC errors during the transition, we don't want
> to queue deferred tuning request.
> 2. You could imagine that the AOS could implement retuning but we
> could still get errors while transitioning in and out of the AOS.
> Similarly you could imagine a seamless transition into and out of
> the AOS (with no CRC errors) even if the AOS couldn't handle
> tuning.
>
> ALSO NOTE: presumably there is never a desperate need to retune in
> order to wake up the card, since doing so is impossible. Luckily the
> only way the card can get into sleep state is if we had a good enough
> tuning to send it a sleep command, so presumably that "good enough"
> tuning is enough to wake us up, at least with a few retries.
>
> Signed-off-by: Douglas Anderson <[email protected]>
Acked-by: Adrian Hunter <[email protected]>
> ---
>
> Changes in v4:
> - Adjust to API rename (Adrian).
>
> Changes in v3:
> - ("brcmfmac: sdio: Don't tune while the card is off") new for v3.
>
> Changes in v2: None
>
> drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c | 7 +++++++
> 1 file changed, 7 insertions(+)
>
> diff --git a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c
> index ee76593259a7..629140b6d7e2 100644
> --- a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c
> +++ b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c
> @@ -669,6 +669,10 @@ brcmf_sdio_kso_control(struct brcmf_sdio *bus, bool on)
>
> sdio_retune_crc_disable(bus->sdiodev->func1);
>
> + /* Cannot re-tune if device is asleep; defer till we're awake */
> + if (on)
> + sdio_retune_hold_now(bus->sdiodev->func1);
> +
> wr_val = (on << SBSDIO_FUNC1_SLEEPCSR_KSO_SHIFT);
> /* 1st KSO write goes to AOS wake up core if device is asleep */
> brcmf_sdiod_writeb(bus->sdiodev, SBSDIO_FUNC1_SLEEPCSR, wr_val, &err);
> @@ -729,6 +733,9 @@ brcmf_sdio_kso_control(struct brcmf_sdio *bus, bool on)
> if (try_cnt > MAX_KSO_ATTEMPTS)
> brcmf_err("max tries: rd_val=0x%x err=%d\n", rd_val, err);
>
> + if (on)
> + sdio_retune_release(bus->sdiodev->func1);
> +
> sdio_retune_crc_enable(bus->sdiodev->func1);
>
> return err;
>
On 14/06/19 2:41 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.
>
> 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
> idle (sometimes referred to as "sleep" in Broadcom code) and active
> state on certain Broadcom WiFi SDIO 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 function 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.
>
> NOTE: this makes the assumption that the card is already tuned well
> enough that it's OK to disable the auto-retuning during one of these
> error-prone situations. Presumably the driver code performing the
> error-prone transfer knows how to recover / retry from errors. ...and
> after we can get back to a state where transfers are no longer
> error-prone then we can enable the auto-retuning again. If we truly
> find ourselves in a case where the card needs to be retuned sometimes
> to handle one of these error-prone transfers then we can always try a
> few transfers first without auto-retuning and then re-try with
> auto-retuning if the first few fail.
>
> 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 <[email protected]>
Not withstanding Ulf's comments:
Acked-by: Adrian Hunter <[email protected]>
> ---
>
> Changes in v4:
> - Moved to SDIO API only (Adrian, Ulf).
> - Renamed to make it less generic, now retune_crc_disable (Ulf).
> - Function header makes it clear host must be claimed (Ulf).
> - No more WARN_ON (Ulf).
>
> Changes in v3:
> - Took out the spinlock since I believe this is all in one context.
>
> Changes in v2:
> - Updated commit message to clarify based on discussion of v1.
>
> drivers/mmc/core/core.c | 5 +++--
> drivers/mmc/core/sdio_io.c | 36 +++++++++++++++++++++++++++++++++++
> include/linux/mmc/core.h | 2 ++
> include/linux/mmc/host.h | 1 +
> include/linux/mmc/sdio_func.h | 3 +++
> 5 files changed, 45 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/mmc/core/core.c b/drivers/mmc/core/core.c
> index 6db36dc870b5..9020cb2490f7 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->retune_crc_disable &&
> (err == -EILSEQ || (mrq->sbc && mrq->sbc->error == -EILSEQ) ||
> (mrq->data && mrq->data->error == -EILSEQ) ||
> (mrq->stop && mrq->stop->error == -EILSEQ)))
> diff --git a/drivers/mmc/core/sdio_io.c b/drivers/mmc/core/sdio_io.c
> index f79f0b0caab8..f822a9630b0e 100644
> --- a/drivers/mmc/core/sdio_io.c
> +++ b/drivers/mmc/core/sdio_io.c
> @@ -734,3 +734,39 @@ 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);
> +
> +/**
> + * sdio_retune_crc_disable - temporarily disable retuning on CRC errors
> + * @func: SDIO function attached to host
> + *
> + * If the SDIO card is known to be in a state where it might produce
> + * CRC errors on the bus in response to commands (like if we know it is
> + * transitioning between power states), an SDIO function driver can
> + * call this function to temporarily disable the SD/MMC core behavior of
> + * triggering an automatic retuning.
> + *
> + * This function should be called while the host is claimed and the host
> + * should remain claimed until sdio_retune_crc_enable() is called.
> + * Specifically, the expected sequence of calls is:
> + * - sdio_claim_host()
> + * - sdio_retune_crc_disable()
> + * - some number of calls like sdio_writeb() and sdio_readb()
> + * - sdio_release_host()
> + */
> +void sdio_retune_crc_disable(struct sdio_func *func)
> +{
> + func->card->host->retune_crc_disable = true;
> +}
> +EXPORT_SYMBOL_GPL(sdio_retune_crc_disable);
> +
> +/**
> + * sdio_retune_crc_enable - reneable retuning on CRC errors
> + * @func: SDIO function attached to host
> + *
> + * This is the compement to sdio_retune_crc_disable().
> + */
> +void sdio_retune_crc_enable(struct sdio_func *func)
> +{
> + func->card->host->retune_crc_disable = false;
> +}
> +EXPORT_SYMBOL_GPL(sdio_retune_crc_enable);
> 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..ecb7972e2423 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 retune_crc_disable:1; /* don't trigger retune upon crc */
>
> int rescan_disable; /* disable card detection */
> int rescan_entered; /* used with nonremovable devices */
> diff --git a/include/linux/mmc/sdio_func.h b/include/linux/mmc/sdio_func.h
> index e9dfdd501cd1..4820e6d09dac 100644
> --- a/include/linux/mmc/sdio_func.h
> +++ b/include/linux/mmc/sdio_func.h
> @@ -167,4 +167,7 @@ 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_crc_disable(struct sdio_func *func);
> +extern void sdio_retune_crc_enable(struct sdio_func *func);
> +
> #endif /* LINUX_MMC_SDIO_FUNC_H */
>
On 14/06/19 2:41 AM, Douglas Anderson wrote:
> There are certain cases, notably when transitioning between sleep and
> active state, when Broadcom SDIO WiFi cards will produce errors on the
> SDIO bus. This is evident from the source code where you can see that
> we try commands in a loop until we either get success or we've tried
> too many times. The comment in the code reinforces this by saying
> "just one write attempt may fail"
>
> Unfortunately these failures sometimes end up causing an "-EILSEQ"
> back to the core which triggers a retuning of the SDIO card and that
> blocks all traffic to the card until it's done.
>
> Let's disable retuning around the commands we expect might fail.
>
> Fixes: bd11e8bd03ca ("mmc: core: Flag re-tuning is needed on CRC errors")
> Signed-off-by: Douglas Anderson <[email protected]>
Acked-by: Adrian Hunter <[email protected]>
> ---
>
> Changes in v4:
> - Adjust to API rename (Adrian, Ulf).
>
> Changes in v3:
> - Expect errors for all of brcmf_sdio_kso_control() (Adrian).
>
> Changes in v2: None
>
> drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c | 4 ++++
> 1 file changed, 4 insertions(+)
>
> diff --git a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c
> index 4a750838d8cd..ee76593259a7 100644
> --- a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c
> +++ b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/sdio.c
> @@ -667,6 +667,8 @@ brcmf_sdio_kso_control(struct brcmf_sdio *bus, bool on)
>
> brcmf_dbg(TRACE, "Enter: on=%d\n", on);
>
> + sdio_retune_crc_disable(bus->sdiodev->func1);
> +
> wr_val = (on << SBSDIO_FUNC1_SLEEPCSR_KSO_SHIFT);
> /* 1st KSO write goes to AOS wake up core if device is asleep */
> brcmf_sdiod_writeb(bus->sdiodev, SBSDIO_FUNC1_SLEEPCSR, wr_val, &err);
> @@ -727,6 +729,8 @@ brcmf_sdio_kso_control(struct brcmf_sdio *bus, bool on)
> if (try_cnt > MAX_KSO_ATTEMPTS)
> brcmf_err("max tries: rd_val=0x%x err=%d\n", rd_val, err);
>
> + sdio_retune_crc_enable(bus->sdiodev->func1);
> +
> return err;
> }
>
>
On 6/14/2019 1:41 AM, Douglas Anderson wrote:
> When Broadcom SDIO cards are idled they go to sleep and a whole
> separate subsystem takes over their SDIO communication. This is the
> Always-On-Subsystem (AOS) and it can't handle tuning requests.
>
> Specifically, as tested on rk3288-veyron-minnie (which reports having
> BCM4354/1 in dmesg), if I force a retune in brcmf_sdio_kso_control()
> when "on = 1" (aka we're transition from sleep to wake) by whacking:
> bus->sdiodev->func1->card->host->need_retune = 1
> ...then I can often see tuning fail. In this case dw_mmc reports "All
> phases bad!"). Note that I don't get 100% failure, presumably because
> sometimes the card itself has already transitioned away from the AOS
> itself by the time we try to wake it up. If I force retuning when "on
> = 0" (AKA force retuning right before sending the command to go to
> sleep) then retuning is always OK.
>
> NOTE: we need _both_ this patch and the patch to avoid triggering
> tuning due to CRC errors in the sleep/wake transition, AKA ("brcmfmac:
> sdio: Disable auto-tuning around commands expected to fail"). Though
> both patches handle issues with Broadcom's AOS, the problems are
> distinct:
> 1. We want to defer (but not ignore) asynchronous (like
> timer-requested) tuning requests till the card is awake. However,
> we want to ignore CRC errors during the transition, we don't want
> to queue deferred tuning request.
> 2. You could imagine that the AOS could implement retuning but we
> could still get errors while transitioning in and out of the AOS.
> Similarly you could imagine a seamless transition into and out of
> the AOS (with no CRC errors) even if the AOS couldn't handle
> tuning.
>
> ALSO NOTE: presumably there is never a desperate need to retune in
> order to wake up the card, since doing so is impossible. Luckily the
> only way the card can get into sleep state is if we had a good enough
> tuning to send it a sleep command, so presumably that "good enough"
> tuning is enough to wake us up, at least with a few retries.
The term "sleep command" is a bit confusing. There actually is a CMD14
defined in the eSD spec, but that is not what we are using (unless we
program the chip to do so) here. It is simply a specific register access
using CMD52.
Apart from that....
Reviewed-by: Arend van Spriel <[email protected]>
>
> Signed-off-by: Douglas Anderson <[email protected]>
I think the stable version is mostly determined by change in MMC/SDIO so
4.18 as mentioned Adrian seems most sensible. bcm4354 support was
introduced in 3.14 and there were some earlier devices (4335) using same
sleep mechanism.
Regards,
Arend