Daniel Drake reported an issue in the libertas sdio client that was triggered by the sdio_single_irq functionality. His SDIO device seems to raise an interrupt even though there are no bits set in the CCCR_INTx register. This behaviour is not supported by the sdio_single_irq feature nor the SDIO spec. The purpose of the sdio_single_irq feature is to avoid the overhead of checking the CCCR_INTx registers, this result in no error handling of the case if there is a pending IRQ with none CCCR_INTx bits set.
This patchset intends to resolve the libertas issue by making sdio_single_irq feature configurable and also report a warning if an SDIO interrupt is raised but none CCCR_INTx bits are set.
Per Forlin (2): sdio: add function to enable and disable sdio_single_irq optimization sdio: report error if pending IRQ but none function bits
drivers/mmc/core/sdio_irq.c | 22 +++++++++++++++++++++- include/linux/mmc/card.h | 1 + 2 files changed, 22 insertions(+), 1 deletions(-)
Make sdio single irq run time configurable, default is disable. This is due to an issue in libertas where the SDIO device seems to raise interrupt even if there are none function bits in CCCR_INTx set. This behaviour is not defined by the SDIO spec.
Signed-off-by: Per Forlin per.forlin@linaro.org --- drivers/mmc/core/sdio_irq.c | 17 ++++++++++++++++- include/linux/mmc/card.h | 1 + 2 files changed, 17 insertions(+), 1 deletions(-)
diff --git a/drivers/mmc/core/sdio_irq.c b/drivers/mmc/core/sdio_irq.c index 03ead02..2f81ddc 100644 --- a/drivers/mmc/core/sdio_irq.c +++ b/drivers/mmc/core/sdio_irq.c @@ -204,7 +204,8 @@ static void sdio_single_irq_set(struct mmc_card *card) int i;
card->sdio_single_irq = NULL; - if ((card->host->caps & MMC_CAP_SDIO_IRQ) && + if (card->sdio_single_irq_en && + (card->host->caps & MMC_CAP_SDIO_IRQ) && card->host->sdio_irqs == 1) for (i = 0; i < card->sdio_funcs; i++) { func = card->sdio_func[i]; @@ -302,3 +303,17 @@ int sdio_release_irq(struct sdio_func *func) } EXPORT_SYMBOL_GPL(sdio_release_irq);
+/** + * sdio_single_irq_enable - enable or disable SDIO single IRQ function + * @card: card to enable SDIO single irq + * @value: true to enable SDIO single irq function, false to disable + * + * If there is only 1 function interrupt registered and SDIO single IRQ + * is enable, the irq handler is called directly without reading + * the CCCR registers + */ +void sdio_single_irq_enable(struct mmc_card *card, bool value) +{ + card->sdio_single_irq_en = value; +} +EXPORT_SYMBOL_GPL(sdio_single_irq_enable); diff --git a/include/linux/mmc/card.h b/include/linux/mmc/card.h index 7b4fd7b..f8b0537 100644 --- a/include/linux/mmc/card.h +++ b/include/linux/mmc/card.h @@ -196,6 +196,7 @@ struct mmc_card { struct sdio_cis cis; /* common tuple info */ struct sdio_func *sdio_func[SDIO_MAX_FUNCS]; /* SDIO functions (devices) */ struct sdio_func *sdio_single_irq; /* SDIO function when only one IRQ active */ + bool sdio_single_irq_en; /* enable single irq */ unsigned num_info; /* number of info strings */ const char **info; /* info strings */ struct sdio_func_tuple *tuples; /* unknown common tuples */
2011/5/31 Per Forlin per.forlin@linaro.org:
+/**
- sdio_single_irq_enable - enable or disable SDIO single IRQ function
- @card: card to enable SDIO single irq
- @value: true to enable SDIO single irq function, false to disable
- If there is only 1 function interrupt registered and SDIO single IRQ
- is enable, the irq handler is called directly without reading
- the CCCR registers
- */
+void sdio_single_irq_enable(struct mmc_card *card, bool value) +{
- card->sdio_single_irq_en = value;
+} +EXPORT_SYMBOL_GPL(sdio_single_irq_enable);
Can we use a quirk for implementing this for the specific problematic card instead?
Daniel, do you have the vendor and device ID for the problematic Libertas card you're working on so this can be quirked explicitly in drivers/mmc/core/quirks.c?
Yours, Linus Walleij
On 1 June 2011 09:30, Linus Walleij linus.walleij@linaro.org wrote:
2011/5/31 Per Forlin per.forlin@linaro.org:
+/**
- sdio_single_irq_enable - enable or disable SDIO single IRQ function
- @card: card to enable SDIO single irq
- @value: true to enable SDIO single irq function, false to disable
- If there is only 1 function interrupt registered and SDIO single IRQ
- is enable, the irq handler is called directly without reading
- the CCCR registers
- */
+void sdio_single_irq_enable(struct mmc_card *card, bool value) +{
- card->sdio_single_irq_en = value;
+} +EXPORT_SYMBOL_GPL(sdio_single_irq_enable);
Can we use a quirk for implementing this for the specific problematic card instead?
Yes, quirks is the thing I should use. I'll remove this function and replace it with a quirk. The default state could then be sdio_single_irq enable and for all none supported hardware (device ID) sdio_single_irq will be disable.
Daniel, do you have the vendor and device ID for the problematic Libertas card you're working on so this can be quirked explicitly in drivers/mmc/core/quirks.c?
Even if Daniel fix this issue in libertas it is still good to have a sdio_single_irq quirk in place for other SDIO devices with the same hardware issue.
Yours, Linus Walleij
Thanks, Per
Return error in case of pending IRQ but none functions bits in CCCR_INTx is set.
Signed-off-by: Per Forlin per.forlin@linaro.org --- drivers/mmc/core/sdio_irq.c | 5 +++++ 1 files changed, 5 insertions(+), 0 deletions(-)
diff --git a/drivers/mmc/core/sdio_irq.c b/drivers/mmc/core/sdio_irq.c index 2f81ddc..8184b6e 100644 --- a/drivers/mmc/core/sdio_irq.c +++ b/drivers/mmc/core/sdio_irq.c @@ -50,6 +50,11 @@ static int process_sdio_pending_irqs(struct mmc_card *card) return ret; }
+ if (!pending) { + printk(KERN_WARNING "%s: pending IRQ but none function bits\n", + mmc_card_id(card)); + ret = -EINVAL; + } count = 0; for (i = 1; i <= 7; i++) { if (pending & (1 << i)) {
On Tue, 31 May 2011, Per Forlin wrote:
Return error in case of pending IRQ but none functions bits in CCCR_INTx is set.
Signed-off-by: Per Forlin per.forlin@linaro.org
drivers/mmc/core/sdio_irq.c | 5 +++++ 1 files changed, 5 insertions(+), 0 deletions(-)
diff --git a/drivers/mmc/core/sdio_irq.c b/drivers/mmc/core/sdio_irq.c index 2f81ddc..8184b6e 100644 --- a/drivers/mmc/core/sdio_irq.c +++ b/drivers/mmc/core/sdio_irq.c @@ -50,6 +50,11 @@ static int process_sdio_pending_irqs(struct mmc_card *card) return ret; }
- if (!pending) {
printk(KERN_WARNING "%s: pending IRQ but none function bits\n",mmc_card_id(card));ret = -EINVAL;- }
Nope, this won't work with the polled interrupt mode.
Nicolas
On 1 June 2011 16:30, Nicolas Pitre nicolas.pitre@linaro.org wrote:
On Tue, 31 May 2011, Per Forlin wrote:
Return error in case of pending IRQ but none functions bits in CCCR_INTx is set.
Signed-off-by: Per Forlin per.forlin@linaro.org
drivers/mmc/core/sdio_irq.c | 5 +++++ 1 files changed, 5 insertions(+), 0 deletions(-)
diff --git a/drivers/mmc/core/sdio_irq.c b/drivers/mmc/core/sdio_irq.c index 2f81ddc..8184b6e 100644 --- a/drivers/mmc/core/sdio_irq.c +++ b/drivers/mmc/core/sdio_irq.c @@ -50,6 +50,11 @@ static int process_sdio_pending_irqs(struct mmc_card *card) return ret; }
- if (!pending) {
- printk(KERN_WARNING "%s: pending IRQ but none function bits\n",
- mmc_card_id(card));
- ret = -EINVAL;
- }
Nope, this won't work with the polled interrupt mode.
I'll drop this patch.
Thanks for your comment, Per
On 31 May 2011 21:33, Per Forlin per.forlin@linaro.org wrote:
Daniel Drake reported an issue in the libertas sdio client that was triggered by the sdio_single_irq functionality. His SDIO device seems to raise an interrupt even though there are no bits set in the CCCR_INTx register. This behaviour is not supported by the sdio_single_irq feature nor the SDIO spec. The purpose of the sdio_single_irq feature is to avoid the overhead of checking the CCCR_INTx registers, this result in no error handling of the case if there is a pending IRQ with none CCCR_INTx bits set.
Thanks a lot for diagnosing this and nice work on figuring out the root cause presumably without even having access to the hardware!
I've looked further, based on your findings, and have found that you are correct. During initialisation, exactly one interrupt is received with CCCR_INTx=0. Previously the mmc stack threw this interrupt away, after the optimization it now calls into libertas before it is ready to handle interrupts, leading to the crash. From that point, all other interrupts that come in are "normal".
This is definitely a weird hardware issue, and it would annoy me for this hardware to cause a second generic mmc layer feature be disabled by default! And actually it is not much work to harden up the libertas driver to be able to accept that spurious IRQ, and during the process of fixing that it actually made the spurious IRQ go away completely.
Patch attached.
So, I vote for that we work around this little hardware issue in the libertas driver, and leave this optimization enabled by default until we find a hardware issue that is more difficult to workaround. I can take on submission of the libertas patch.
Thoughts?
Daniel
On Tue, 31 May 2011, Daniel Drake wrote:
On 31 May 2011 21:33, Per Forlin per.forlin@linaro.org wrote:
Daniel Drake reported an issue in the libertas sdio client that was triggered by the sdio_single_irq functionality. His SDIO device seems to raise an interrupt even though there are no bits set in the CCCR_INTx register. This behaviour is not supported by the sdio_single_irq feature nor the SDIO spec. The purpose of the sdio_single_irq feature is to avoid the overhead of checking the CCCR_INTx registers, this result in no error handling of the case if there is a pending IRQ with none CCCR_INTx bits set.
Thanks a lot for diagnosing this and nice work on figuring out the root cause presumably without even having access to the hardware!
I've looked further, based on your findings, and have found that you are correct. During initialisation, exactly one interrupt is received with CCCR_INTx=0. Previously the mmc stack threw this interrupt away, after the optimization it now calls into libertas before it is ready to handle interrupts, leading to the crash. From that point, all other interrupts that come in are "normal".
This is definitely a weird hardware issue, and it would annoy me for this hardware to cause a second generic mmc layer feature be disabled by default! And actually it is not much work to harden up the libertas driver to be able to accept that spurious IRQ, and during the process of fixing that it actually made the spurious IRQ go away completely.
Patch attached.
So, I vote for that we work around this little hardware issue in the libertas driver, and leave this optimization enabled by default until we find a hardware issue that is more difficult to workaround. I can take on submission of the libertas patch.
Thoughts?
This is definitively the best approach. Thanks for fixing the root cause.
Nicolas
-
Daniel Drake -
Linus Walleij -
Nicolas Pitre -
Per Forlin