LKML Archive on lore.kernel.org
help / color / mirror / Atom feed
* [PATCH v2 1/2] sdhci: tegra: Implement Tegra specific set_timeout callback
@ 2020-03-11 15:47 Sowjanya Komatineni
  2020-03-11 15:47 ` [PATCH v2 2/2] sdhci: tegra: Enable MMC_CAP_WAIT_WHILE_BUSY host capability Sowjanya Komatineni
  0 siblings, 1 reply; 16+ messages in thread
From: Sowjanya Komatineni @ 2020-03-11 15:47 UTC (permalink / raw)
  To: adrian.hunter, ulf.hansson, baolin.wang, kstewart, tglx,
	bradleybolen, gregkh, thierry.reding, jonathanh, skomatineni
  Cc: anrao, linux-tegra, linux-kernel, linux-mmc

Tegra host supports HW busy detection and timeouts based on the
count programmed in SDHCI_TIMEOUT_CONTROL register and max busy
timeout it supports is 11s in finite busy wait mode.

Some operations like SLEEP_AWAKE, ERASE and flush cache through
SWITCH commands take longer than 11s and Tegra host supports
infinite HW busy wait mode where HW waits forever till the card
is busy without HW timeout.

This patch implements Tegra specific set_timeout sdhci_ops to allow
switching between finite and infinite HW busy detection wait modes
based on the device command expected operation time.

Signed-off-by: Sowjanya Komatineni <skomatineni@nvidia.com>
---
 drivers/mmc/host/sdhci-tegra.c | 31 +++++++++++++++++++++++++++++++
 1 file changed, 31 insertions(+)

diff --git a/drivers/mmc/host/sdhci-tegra.c b/drivers/mmc/host/sdhci-tegra.c
index a25c3a4..fa8f6a4 100644
--- a/drivers/mmc/host/sdhci-tegra.c
+++ b/drivers/mmc/host/sdhci-tegra.c
@@ -45,6 +45,7 @@
 #define SDHCI_TEGRA_CAP_OVERRIDES_DQS_TRIM_SHIFT	8
 
 #define SDHCI_TEGRA_VENDOR_MISC_CTRL			0x120
+#define SDHCI_MISC_CTRL_ERASE_TIMEOUT_LIMIT		BIT(0)
 #define SDHCI_MISC_CTRL_ENABLE_SDR104			0x8
 #define SDHCI_MISC_CTRL_ENABLE_SDR50			0x10
 #define SDHCI_MISC_CTRL_ENABLE_SDHCI_SPEC_300		0x20
@@ -1227,6 +1228,34 @@ static u32 sdhci_tegra_cqhci_irq(struct sdhci_host *host, u32 intmask)
 	return 0;
 }
 
+static void tegra_sdhci_set_timeout(struct sdhci_host *host,
+				    struct mmc_command *cmd)
+{
+	u32 val;
+
+	/*
+	 * HW busy detection timeout is based on programmed data timeout
+	 * counter and maximum supported timeout is 11s which may not be
+	 * enough for long operations like cache flush, sleep awake, erase.
+	 *
+	 * ERASE_TIMEOUT_LIMIT bit of VENDOR_MISC_CTRL register allows
+	 * host controller to wait for busy state until the card is busy
+	 * without HW timeout.
+	 *
+	 * So, use infinite busy wait mode for operations that may take
+	 * more than maximum HW busy timeout of 11s otherwise use finite
+	 * busy wait mode.
+	 */
+	val = sdhci_readl(host, SDHCI_TEGRA_VENDOR_MISC_CTRL);
+	if (cmd && cmd->busy_timeout >= 11 * HZ)
+		val |= SDHCI_MISC_CTRL_ERASE_TIMEOUT_LIMIT;
+	else
+		val &= ~SDHCI_MISC_CTRL_ERASE_TIMEOUT_LIMIT;
+	sdhci_writel(host, val, SDHCI_TEGRA_VENDOR_MISC_CTRL);
+
+	__sdhci_set_timeout(host, cmd);
+}
+
 static const struct cqhci_host_ops sdhci_tegra_cqhci_ops = {
 	.write_l    = tegra_cqhci_writel,
 	.enable	= sdhci_tegra_cqe_enable,
@@ -1366,6 +1395,7 @@ static const struct sdhci_ops tegra210_sdhci_ops = {
 	.set_uhs_signaling = tegra_sdhci_set_uhs_signaling,
 	.voltage_switch = tegra_sdhci_voltage_switch,
 	.get_max_clock = tegra_sdhci_get_max_clock,
+	.set_timeout = tegra_sdhci_set_timeout,
 };
 
 static const struct sdhci_pltfm_data sdhci_tegra210_pdata = {
@@ -1403,6 +1433,7 @@ static const struct sdhci_ops tegra186_sdhci_ops = {
 	.voltage_switch = tegra_sdhci_voltage_switch,
 	.get_max_clock = tegra_sdhci_get_max_clock,
 	.irq = sdhci_tegra_cqhci_irq,
+	.set_timeout = tegra_sdhci_set_timeout,
 };
 
 static const struct sdhci_pltfm_data sdhci_tegra186_pdata = {
-- 
2.7.4


^ permalink raw reply	[flat|nested] 16+ messages in thread

* [PATCH v2 2/2] sdhci: tegra: Enable MMC_CAP_WAIT_WHILE_BUSY host capability
  2020-03-11 15:47 [PATCH v2 1/2] sdhci: tegra: Implement Tegra specific set_timeout callback Sowjanya Komatineni
@ 2020-03-11 15:47 ` Sowjanya Komatineni
  0 siblings, 0 replies; 16+ messages in thread
From: Sowjanya Komatineni @ 2020-03-11 15:47 UTC (permalink / raw)
  To: adrian.hunter, ulf.hansson, baolin.wang, kstewart, tglx,
	bradleybolen, gregkh, thierry.reding, jonathanh, skomatineni
  Cc: anrao, linux-tegra, linux-kernel, linux-mmc

Tegra sdhci host supports HW busy detection of the device busy
signaling over data0 lane.

So, this patch enables host capability MMC_CAP_wAIT_WHILE_BUSY.

Signed-off-by: Sowjanya Komatineni <skomatineni@nvidia.com>
---
 drivers/mmc/host/sdhci-tegra.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/mmc/host/sdhci-tegra.c b/drivers/mmc/host/sdhci-tegra.c
index fa8f6a4..1c381f8 100644
--- a/drivers/mmc/host/sdhci-tegra.c
+++ b/drivers/mmc/host/sdhci-tegra.c
@@ -1580,6 +1580,8 @@ static int sdhci_tegra_probe(struct platform_device *pdev)
 	if (rc)
 		goto err_parse_dt;
 
+	host->mmc->caps |= MMC_CAP_WAIT_WHILE_BUSY;
+
 	if (tegra_host->soc_data->nvquirks & NVQUIRK_ENABLE_DDR50)
 		host->mmc->caps |= MMC_CAP_1_8V_DDR;
 
-- 
2.7.4


^ permalink raw reply	[flat|nested] 16+ messages in thread

* Re: [PATCH v2 1/2] sdhci: tegra: Implement Tegra specific set_timeout callback
  2020-04-17  8:04         ` Ulf Hansson
@ 2020-04-17 18:29           ` Sowjanya Komatineni
  0 siblings, 0 replies; 16+ messages in thread
From: Sowjanya Komatineni @ 2020-04-17 18:29 UTC (permalink / raw)
  To: Ulf Hansson
  Cc: Naresh Kamboju, Greg Kroah-Hartman, Sasha Levin, Adrian Hunter,
	(Exiting) Baolin Wang, Kate Stewart, Thomas Gleixner,
	Bradley Bolen, Thierry Reding, Jon Hunter, Aniruddha Tvs Rao,
	linux-tegra, open list, linux-mmc, lkft-triage, linux- stable


On 4/17/20 1:04 AM, Ulf Hansson wrote:
> External email: Use caution opening links or attachments
>
>
> On Thu, 16 Apr 2020 at 21:39, Sowjanya Komatineni
> <skomatineni@nvidia.com> wrote:
>>
>> On 4/16/20 9:29 AM, Sowjanya Komatineni wrote:
>>> On 4/16/20 3:59 AM, Ulf Hansson wrote:
>>>> External email: Use caution opening links or attachments
>>>>
>>>>
>>>> On Wed, 15 Apr 2020 at 19:55, Naresh Kamboju
>>>> <naresh.kamboju@linaro.org> wrote:
>>>>> On Fri, 13 Mar 2020 at 06:41, Sowjanya Komatineni
>>>>> <skomatineni@nvidia.com> wrote:
>>>>>> Tegra host supports HW busy detection and timeouts based on the
>>>>>> count programmed in SDHCI_TIMEOUT_CONTROL register and max busy
>>>>>> timeout it supports is 11s in finite busy wait mode.
>>>>>>
>>>>>> Some operations like SLEEP_AWAKE, ERASE and flush cache through
>>>>>> SWITCH commands take longer than 11s and Tegra host supports
>>>>>> infinite HW busy wait mode where HW waits forever till the card
>>>>>> is busy without HW timeout.
>>>>>>
>>>>>> This patch implements Tegra specific set_timeout sdhci_ops to allow
>>>>>> switching between finite and infinite HW busy detection wait modes
>>>>>> based on the device command expected operation time.
>>>>>>
>>>>>> Signed-off-by: Sowjanya Komatineni <skomatineni@nvidia.com>
>>>>>> ---
>>>>>>    drivers/mmc/host/sdhci-tegra.c | 31 +++++++++++++++++++++++++++++++
>>>>>>    1 file changed, 31 insertions(+)
>>>>>>
>>>>>> diff --git a/drivers/mmc/host/sdhci-tegra.c
>>>>>> b/drivers/mmc/host/sdhci-tegra.c
>>>>>> index a25c3a4..fa8f6a4 100644
>>>>>> --- a/drivers/mmc/host/sdhci-tegra.c
>>>>>> +++ b/drivers/mmc/host/sdhci-tegra.c
>>>>>> @@ -45,6 +45,7 @@
>>>>>>    #define SDHCI_TEGRA_CAP_OVERRIDES_DQS_TRIM_SHIFT       8
>>>>>>
>>>>>>    #define SDHCI_TEGRA_VENDOR_MISC_CTRL 0x120
>>>>>> +#define SDHCI_MISC_CTRL_ERASE_TIMEOUT_LIMIT BIT(0)
>>>>>>    #define SDHCI_MISC_CTRL_ENABLE_SDR104                  0x8
>>>>>>    #define SDHCI_MISC_CTRL_ENABLE_SDR50 0x10
>>>>>>    #define SDHCI_MISC_CTRL_ENABLE_SDHCI_SPEC_300 0x20
>>>>>> @@ -1227,6 +1228,34 @@ static u32 sdhci_tegra_cqhci_irq(struct
>>>>>> sdhci_host *host, u32 intmask)
>>>>>>           return 0;
>>>>>>    }
>>>>>>
>>>>>> +static void tegra_sdhci_set_timeout(struct sdhci_host *host,
>>>>>> +                                   struct mmc_command *cmd)
>>>>>> +{
>>>>>> +       u32 val;
>>>>>> +
>>>>>> +       /*
>>>>>> +        * HW busy detection timeout is based on programmed data
>>>>>> timeout
>>>>>> +        * counter and maximum supported timeout is 11s which may
>>>>>> not be
>>>>>> +        * enough for long operations like cache flush, sleep
>>>>>> awake, erase.
>>>>>> +        *
>>>>>> +        * ERASE_TIMEOUT_LIMIT bit of VENDOR_MISC_CTRL register allows
>>>>>> +        * host controller to wait for busy state until the card is
>>>>>> busy
>>>>>> +        * without HW timeout.
>>>>>> +        *
>>>>>> +        * So, use infinite busy wait mode for operations that may
>>>>>> take
>>>>>> +        * more than maximum HW busy timeout of 11s otherwise use
>>>>>> finite
>>>>>> +        * busy wait mode.
>>>>>> +        */
>>>>>> +       val = sdhci_readl(host, SDHCI_TEGRA_VENDOR_MISC_CTRL);
>>>>>> +       if (cmd && cmd->busy_timeout >= 11 * HZ)
>>>>>> +               val |= SDHCI_MISC_CTRL_ERASE_TIMEOUT_LIMIT;
>>>>>> +       else
>>>>>> +               val &= ~SDHCI_MISC_CTRL_ERASE_TIMEOUT_LIMIT;
>>>>>> +       sdhci_writel(host, val, SDHCI_TEGRA_VENDOR_MISC_CTRL);
>>>>>> +
>>>>>> +       __sdhci_set_timeout(host, cmd);
>>>>> kernel build on arm and arm64 architecture failed on stable-rc 4.19
>>>>> (arm), 5.4 (arm64) and 5.5 (arm64)
>>>>>
>>>>> drivers/mmc/host/sdhci-tegra.c: In function 'tegra_sdhci_set_timeout':
>>>>> drivers/mmc/host/sdhci-tegra.c:1256:2: error: implicit declaration of
>>>>> function '__sdhci_set_timeout'; did you mean
>>>>> 'tegra_sdhci_set_timeout'? [-Werror=implicit-function-declaration]
>>>>>     __sdhci_set_timeout(host, cmd);
>>>>>     ^~~~~~~~~~~~~~~~~~~
>>>>>     tegra_sdhci_set_timeout
>>>>>
>>>>> Full build log,
>>>>> https://ci.linaro.org/view/lkft/job/openembedded-lkft-linux-stable-rc-5.5/DISTRO=lkft,MACHINE=am57xx-evm,label=docker-lkft/83/consoleText
>>>>>
>>>>> https://ci.linaro.org/view/lkft/job/openembedded-lkft-linux-stable-rc-5.4/DISTRO=lkft,MACHINE=juno,label=docker-lkft/158/consoleText
>>>>>
>>>>> https://ci.linaro.org/view/lkft/job/openembedded-lkft-linux-stable-rc-4.19/DISTRO=lkft,MACHINE=am57xx-evm,label=docker-lkft/511/consoleText
>>>>>
>>>>>
>>>>> - Naresh
>>>> Thanks for reporting! What a mess.
>>>>
>>>> It turns out that the commit that was queued for stable that is
>>>> causing the above errors, also requires another commit.
>>>>
>>>> The commit that was queued:
>>>> 5e958e4aacf4 ("sdhci: tegra: Implement Tegra specific set_timeout
>>>> callback")
>>>>
>>>> The additional commit needed (which was added in v5.6-rc1):
>>>> 7d76ed77cfbd ("mmc: sdhci: Refactor sdhci_set_timeout()")
>>>>
>>>> However, the above commit needs a manual backport (quite trivial, but
>>>> still) for the relevant stable kernels, to allow it to solve the build
>>>> problems.
>>>>
>>>> Greg, Sasha - I suggest you to drop the offending commit from the
>>>> stable kernels, for now. I think it's better to let Sowjanya deal with
>>>> the backports, then send them in small series instead.
>>>>
>>>> Kind regards
>>>> Uffe
>>> Hi Ufee,
>>>
>>> Will back-porting below commit cause any issues to other vendors?
>>>
>>> 7d76ed77cfbd ("mmc: sdhci: Refactor sdhci_set_timeout()")
>>>
>> sdhci-tegra driver in 4.19 is using same sdhci_ops for Tegra114 and
>> Tegra210 and separate sdhci_ops for T210 started from 4.20.
>>
>> 5e958e4aacf4 ("sdhci: tegra: Implement Tegra specific set_timeout callback")
>>
>> So above commit can't be applied to 4.19. So probably a separate patch
>> need to be created to apply for 4.19 and back port above commit along
>> with its dependency commit (7d76ed77cfbd ("mmc: sdhci: Refactor
>> sdhci_set_timeout()") for 5.4 and 5.4.
> Alright, seems reasonable. Just keep me/Adrian on cc when/if you post
> the patches so we can ack them.
>
> Kind regards
> Uffe

Sure, will send patches to backport both below changes to 4.19

5e958e4aacf4 ("sdhci: tegra: Implement Tegra specific set_timeout
callback")

7d76ed77cfbd ("mmc: sdhci: Refactor sdhci_set_timeout()")

But on 5.5 and on 5.4.33,  patch "mmc: sdhci: Refactor 
sdhci_set_timeout() already committed but not on 5.4.32 and prior.

So, not sure why kbuild reported error on 5.4 and 5.5 with 
__sdhci_set_timeout when this patch is already committed on 5.4.33 
(26711cc7e064) and 5.5.18 (71bab39fa67d)

Can you please help clarify this?


drivers/mmc/host/sdhci-tegra.c: In function 'tegra_sdhci_set_timeout':
drivers/mmc/host/sdhci-tegra.c:1256:2: error: implicit declaration of
function '__sdhci_set_timeout'; did you mean
'tegra_sdhci_set_timeout'? [-Werror=implicit-function-declaration]
    __sdhci_set_timeout(host, cmd);


Thanks

Sowjanya



^ permalink raw reply	[flat|nested] 16+ messages in thread

* Re: [PATCH v2 1/2] sdhci: tegra: Implement Tegra specific set_timeout callback
  2020-04-16 19:38       ` Sowjanya Komatineni
@ 2020-04-17  8:04         ` Ulf Hansson
  2020-04-17 18:29           ` Sowjanya Komatineni
  0 siblings, 1 reply; 16+ messages in thread
From: Ulf Hansson @ 2020-04-17  8:04 UTC (permalink / raw)
  To: Sowjanya Komatineni
  Cc: Naresh Kamboju, Greg Kroah-Hartman, Sasha Levin, Adrian Hunter,
	(Exiting) Baolin Wang, Kate Stewart, Thomas Gleixner,
	Bradley Bolen, Thierry Reding, Jon Hunter, Aniruddha Tvs Rao,
	linux-tegra, open list, linux-mmc, lkft-triage, linux- stable

On Thu, 16 Apr 2020 at 21:39, Sowjanya Komatineni
<skomatineni@nvidia.com> wrote:
>
>
> On 4/16/20 9:29 AM, Sowjanya Komatineni wrote:
> >
> > On 4/16/20 3:59 AM, Ulf Hansson wrote:
> >> External email: Use caution opening links or attachments
> >>
> >>
> >> On Wed, 15 Apr 2020 at 19:55, Naresh Kamboju
> >> <naresh.kamboju@linaro.org> wrote:
> >>> On Fri, 13 Mar 2020 at 06:41, Sowjanya Komatineni
> >>> <skomatineni@nvidia.com> wrote:
> >>>> Tegra host supports HW busy detection and timeouts based on the
> >>>> count programmed in SDHCI_TIMEOUT_CONTROL register and max busy
> >>>> timeout it supports is 11s in finite busy wait mode.
> >>>>
> >>>> Some operations like SLEEP_AWAKE, ERASE and flush cache through
> >>>> SWITCH commands take longer than 11s and Tegra host supports
> >>>> infinite HW busy wait mode where HW waits forever till the card
> >>>> is busy without HW timeout.
> >>>>
> >>>> This patch implements Tegra specific set_timeout sdhci_ops to allow
> >>>> switching between finite and infinite HW busy detection wait modes
> >>>> based on the device command expected operation time.
> >>>>
> >>>> Signed-off-by: Sowjanya Komatineni <skomatineni@nvidia.com>
> >>>> ---
> >>>>   drivers/mmc/host/sdhci-tegra.c | 31 +++++++++++++++++++++++++++++++
> >>>>   1 file changed, 31 insertions(+)
> >>>>
> >>>> diff --git a/drivers/mmc/host/sdhci-tegra.c
> >>>> b/drivers/mmc/host/sdhci-tegra.c
> >>>> index a25c3a4..fa8f6a4 100644
> >>>> --- a/drivers/mmc/host/sdhci-tegra.c
> >>>> +++ b/drivers/mmc/host/sdhci-tegra.c
> >>>> @@ -45,6 +45,7 @@
> >>>>   #define SDHCI_TEGRA_CAP_OVERRIDES_DQS_TRIM_SHIFT       8
> >>>>
> >>>>   #define SDHCI_TEGRA_VENDOR_MISC_CTRL 0x120
> >>>> +#define SDHCI_MISC_CTRL_ERASE_TIMEOUT_LIMIT BIT(0)
> >>>>   #define SDHCI_MISC_CTRL_ENABLE_SDR104                  0x8
> >>>>   #define SDHCI_MISC_CTRL_ENABLE_SDR50 0x10
> >>>>   #define SDHCI_MISC_CTRL_ENABLE_SDHCI_SPEC_300 0x20
> >>>> @@ -1227,6 +1228,34 @@ static u32 sdhci_tegra_cqhci_irq(struct
> >>>> sdhci_host *host, u32 intmask)
> >>>>          return 0;
> >>>>   }
> >>>>
> >>>> +static void tegra_sdhci_set_timeout(struct sdhci_host *host,
> >>>> +                                   struct mmc_command *cmd)
> >>>> +{
> >>>> +       u32 val;
> >>>> +
> >>>> +       /*
> >>>> +        * HW busy detection timeout is based on programmed data
> >>>> timeout
> >>>> +        * counter and maximum supported timeout is 11s which may
> >>>> not be
> >>>> +        * enough for long operations like cache flush, sleep
> >>>> awake, erase.
> >>>> +        *
> >>>> +        * ERASE_TIMEOUT_LIMIT bit of VENDOR_MISC_CTRL register allows
> >>>> +        * host controller to wait for busy state until the card is
> >>>> busy
> >>>> +        * without HW timeout.
> >>>> +        *
> >>>> +        * So, use infinite busy wait mode for operations that may
> >>>> take
> >>>> +        * more than maximum HW busy timeout of 11s otherwise use
> >>>> finite
> >>>> +        * busy wait mode.
> >>>> +        */
> >>>> +       val = sdhci_readl(host, SDHCI_TEGRA_VENDOR_MISC_CTRL);
> >>>> +       if (cmd && cmd->busy_timeout >= 11 * HZ)
> >>>> +               val |= SDHCI_MISC_CTRL_ERASE_TIMEOUT_LIMIT;
> >>>> +       else
> >>>> +               val &= ~SDHCI_MISC_CTRL_ERASE_TIMEOUT_LIMIT;
> >>>> +       sdhci_writel(host, val, SDHCI_TEGRA_VENDOR_MISC_CTRL);
> >>>> +
> >>>> +       __sdhci_set_timeout(host, cmd);
> >>> kernel build on arm and arm64 architecture failed on stable-rc 4.19
> >>> (arm), 5.4 (arm64) and 5.5 (arm64)
> >>>
> >>> drivers/mmc/host/sdhci-tegra.c: In function 'tegra_sdhci_set_timeout':
> >>> drivers/mmc/host/sdhci-tegra.c:1256:2: error: implicit declaration of
> >>> function '__sdhci_set_timeout'; did you mean
> >>> 'tegra_sdhci_set_timeout'? [-Werror=implicit-function-declaration]
> >>>    __sdhci_set_timeout(host, cmd);
> >>>    ^~~~~~~~~~~~~~~~~~~
> >>>    tegra_sdhci_set_timeout
> >>>
> >>> Full build log,
> >>> https://ci.linaro.org/view/lkft/job/openembedded-lkft-linux-stable-rc-5.5/DISTRO=lkft,MACHINE=am57xx-evm,label=docker-lkft/83/consoleText
> >>>
> >>> https://ci.linaro.org/view/lkft/job/openembedded-lkft-linux-stable-rc-5.4/DISTRO=lkft,MACHINE=juno,label=docker-lkft/158/consoleText
> >>>
> >>> https://ci.linaro.org/view/lkft/job/openembedded-lkft-linux-stable-rc-4.19/DISTRO=lkft,MACHINE=am57xx-evm,label=docker-lkft/511/consoleText
> >>>
> >>>
> >>> - Naresh
> >> Thanks for reporting! What a mess.
> >>
> >> It turns out that the commit that was queued for stable that is
> >> causing the above errors, also requires another commit.
> >>
> >> The commit that was queued:
> >> 5e958e4aacf4 ("sdhci: tegra: Implement Tegra specific set_timeout
> >> callback")
> >>
> >> The additional commit needed (which was added in v5.6-rc1):
> >> 7d76ed77cfbd ("mmc: sdhci: Refactor sdhci_set_timeout()")
> >>
> >> However, the above commit needs a manual backport (quite trivial, but
> >> still) for the relevant stable kernels, to allow it to solve the build
> >> problems.
> >>
> >> Greg, Sasha - I suggest you to drop the offending commit from the
> >> stable kernels, for now. I think it's better to let Sowjanya deal with
> >> the backports, then send them in small series instead.
> >>
> >> Kind regards
> >> Uffe
> >
> > Hi Ufee,
> >
> > Will back-porting below commit cause any issues to other vendors?
> >
> > 7d76ed77cfbd ("mmc: sdhci: Refactor sdhci_set_timeout()")
> >
> sdhci-tegra driver in 4.19 is using same sdhci_ops for Tegra114 and
> Tegra210 and separate sdhci_ops for T210 started from 4.20.
>
> 5e958e4aacf4 ("sdhci: tegra: Implement Tegra specific set_timeout callback")
>
> So above commit can't be applied to 4.19. So probably a separate patch
> need to be created to apply for 4.19 and back port above commit along
> with its dependency commit (7d76ed77cfbd ("mmc: sdhci: Refactor
> sdhci_set_timeout()") for 5.4 and 5.4.

Alright, seems reasonable. Just keep me/Adrian on cc when/if you post
the patches so we can ack them.

Kind regards
Uffe

^ permalink raw reply	[flat|nested] 16+ messages in thread

* Re: [PATCH v2 1/2] sdhci: tegra: Implement Tegra specific set_timeout callback
  2020-04-16 16:29     ` Sowjanya Komatineni
@ 2020-04-16 19:38       ` Sowjanya Komatineni
  2020-04-17  8:04         ` Ulf Hansson
  0 siblings, 1 reply; 16+ messages in thread
From: Sowjanya Komatineni @ 2020-04-16 19:38 UTC (permalink / raw)
  To: Ulf Hansson, Naresh Kamboju, Greg Kroah-Hartman, Sasha Levin
  Cc: Adrian Hunter, (Exiting) Baolin Wang, Kate Stewart,
	Thomas Gleixner, Bradley Bolen, Thierry Reding, Jon Hunter,
	Aniruddha Tvs Rao, linux-tegra, open list, linux-mmc,
	lkft-triage, linux- stable


On 4/16/20 9:29 AM, Sowjanya Komatineni wrote:
>
> On 4/16/20 3:59 AM, Ulf Hansson wrote:
>> External email: Use caution opening links or attachments
>>
>>
>> On Wed, 15 Apr 2020 at 19:55, Naresh Kamboju 
>> <naresh.kamboju@linaro.org> wrote:
>>> On Fri, 13 Mar 2020 at 06:41, Sowjanya Komatineni
>>> <skomatineni@nvidia.com> wrote:
>>>> Tegra host supports HW busy detection and timeouts based on the
>>>> count programmed in SDHCI_TIMEOUT_CONTROL register and max busy
>>>> timeout it supports is 11s in finite busy wait mode.
>>>>
>>>> Some operations like SLEEP_AWAKE, ERASE and flush cache through
>>>> SWITCH commands take longer than 11s and Tegra host supports
>>>> infinite HW busy wait mode where HW waits forever till the card
>>>> is busy without HW timeout.
>>>>
>>>> This patch implements Tegra specific set_timeout sdhci_ops to allow
>>>> switching between finite and infinite HW busy detection wait modes
>>>> based on the device command expected operation time.
>>>>
>>>> Signed-off-by: Sowjanya Komatineni <skomatineni@nvidia.com>
>>>> ---
>>>>   drivers/mmc/host/sdhci-tegra.c | 31 +++++++++++++++++++++++++++++++
>>>>   1 file changed, 31 insertions(+)
>>>>
>>>> diff --git a/drivers/mmc/host/sdhci-tegra.c 
>>>> b/drivers/mmc/host/sdhci-tegra.c
>>>> index a25c3a4..fa8f6a4 100644
>>>> --- a/drivers/mmc/host/sdhci-tegra.c
>>>> +++ b/drivers/mmc/host/sdhci-tegra.c
>>>> @@ -45,6 +45,7 @@
>>>>   #define SDHCI_TEGRA_CAP_OVERRIDES_DQS_TRIM_SHIFT       8
>>>>
>>>>   #define SDHCI_TEGRA_VENDOR_MISC_CTRL 0x120
>>>> +#define SDHCI_MISC_CTRL_ERASE_TIMEOUT_LIMIT BIT(0)
>>>>   #define SDHCI_MISC_CTRL_ENABLE_SDR104                  0x8
>>>>   #define SDHCI_MISC_CTRL_ENABLE_SDR50 0x10
>>>>   #define SDHCI_MISC_CTRL_ENABLE_SDHCI_SPEC_300 0x20
>>>> @@ -1227,6 +1228,34 @@ static u32 sdhci_tegra_cqhci_irq(struct 
>>>> sdhci_host *host, u32 intmask)
>>>>          return 0;
>>>>   }
>>>>
>>>> +static void tegra_sdhci_set_timeout(struct sdhci_host *host,
>>>> +                                   struct mmc_command *cmd)
>>>> +{
>>>> +       u32 val;
>>>> +
>>>> +       /*
>>>> +        * HW busy detection timeout is based on programmed data 
>>>> timeout
>>>> +        * counter and maximum supported timeout is 11s which may 
>>>> not be
>>>> +        * enough for long operations like cache flush, sleep 
>>>> awake, erase.
>>>> +        *
>>>> +        * ERASE_TIMEOUT_LIMIT bit of VENDOR_MISC_CTRL register allows
>>>> +        * host controller to wait for busy state until the card is 
>>>> busy
>>>> +        * without HW timeout.
>>>> +        *
>>>> +        * So, use infinite busy wait mode for operations that may 
>>>> take
>>>> +        * more than maximum HW busy timeout of 11s otherwise use 
>>>> finite
>>>> +        * busy wait mode.
>>>> +        */
>>>> +       val = sdhci_readl(host, SDHCI_TEGRA_VENDOR_MISC_CTRL);
>>>> +       if (cmd && cmd->busy_timeout >= 11 * HZ)
>>>> +               val |= SDHCI_MISC_CTRL_ERASE_TIMEOUT_LIMIT;
>>>> +       else
>>>> +               val &= ~SDHCI_MISC_CTRL_ERASE_TIMEOUT_LIMIT;
>>>> +       sdhci_writel(host, val, SDHCI_TEGRA_VENDOR_MISC_CTRL);
>>>> +
>>>> +       __sdhci_set_timeout(host, cmd);
>>> kernel build on arm and arm64 architecture failed on stable-rc 4.19
>>> (arm), 5.4 (arm64) and 5.5 (arm64)
>>>
>>> drivers/mmc/host/sdhci-tegra.c: In function 'tegra_sdhci_set_timeout':
>>> drivers/mmc/host/sdhci-tegra.c:1256:2: error: implicit declaration of
>>> function '__sdhci_set_timeout'; did you mean
>>> 'tegra_sdhci_set_timeout'? [-Werror=implicit-function-declaration]
>>>    __sdhci_set_timeout(host, cmd);
>>>    ^~~~~~~~~~~~~~~~~~~
>>>    tegra_sdhci_set_timeout
>>>
>>> Full build log,
>>> https://ci.linaro.org/view/lkft/job/openembedded-lkft-linux-stable-rc-5.5/DISTRO=lkft,MACHINE=am57xx-evm,label=docker-lkft/83/consoleText 
>>>
>>> https://ci.linaro.org/view/lkft/job/openembedded-lkft-linux-stable-rc-5.4/DISTRO=lkft,MACHINE=juno,label=docker-lkft/158/consoleText 
>>>
>>> https://ci.linaro.org/view/lkft/job/openembedded-lkft-linux-stable-rc-4.19/DISTRO=lkft,MACHINE=am57xx-evm,label=docker-lkft/511/consoleText 
>>>
>>>
>>> - Naresh
>> Thanks for reporting! What a mess.
>>
>> It turns out that the commit that was queued for stable that is
>> causing the above errors, also requires another commit.
>>
>> The commit that was queued:
>> 5e958e4aacf4 ("sdhci: tegra: Implement Tegra specific set_timeout 
>> callback")
>>
>> The additional commit needed (which was added in v5.6-rc1):
>> 7d76ed77cfbd ("mmc: sdhci: Refactor sdhci_set_timeout()")
>>
>> However, the above commit needs a manual backport (quite trivial, but
>> still) for the relevant stable kernels, to allow it to solve the build
>> problems.
>>
>> Greg, Sasha - I suggest you to drop the offending commit from the
>> stable kernels, for now. I think it's better to let Sowjanya deal with
>> the backports, then send them in small series instead.
>>
>> Kind regards
>> Uffe
>
> Hi Ufee,
>
> Will back-porting below commit cause any issues to other vendors?
>
> 7d76ed77cfbd ("mmc: sdhci: Refactor sdhci_set_timeout()")
>
sdhci-tegra driver in 4.19 is using same sdhci_ops for Tegra114 and 
Tegra210 and separate sdhci_ops for T210 started from 4.20.

5e958e4aacf4 ("sdhci: tegra: Implement Tegra specific set_timeout callback")

So above commit can't be applied to 4.19. So probably a separate patch 
need to be created to apply for 4.19 and back port above commit along 
with its dependency commit (7d76ed77cfbd ("mmc: sdhci: Refactor 
sdhci_set_timeout()") for 5.4 and 5.4.


> Thanks
Sowjanya
>

^ permalink raw reply	[flat|nested] 16+ messages in thread

* Re: [PATCH v2 1/2] sdhci: tegra: Implement Tegra specific set_timeout callback
  2020-04-16 10:59   ` Ulf Hansson
  2020-04-16 11:37     ` Greg Kroah-Hartman
@ 2020-04-16 16:29     ` Sowjanya Komatineni
  2020-04-16 19:38       ` Sowjanya Komatineni
  1 sibling, 1 reply; 16+ messages in thread
From: Sowjanya Komatineni @ 2020-04-16 16:29 UTC (permalink / raw)
  To: Ulf Hansson, Naresh Kamboju, Greg Kroah-Hartman, Sasha Levin
  Cc: Adrian Hunter, (Exiting) Baolin Wang, Kate Stewart,
	Thomas Gleixner, Bradley Bolen, Thierry Reding, Jon Hunter,
	Aniruddha Tvs Rao, linux-tegra, open list, linux-mmc,
	lkft-triage, linux- stable


On 4/16/20 3:59 AM, Ulf Hansson wrote:
> External email: Use caution opening links or attachments
>
>
> On Wed, 15 Apr 2020 at 19:55, Naresh Kamboju <naresh.kamboju@linaro.org> wrote:
>> On Fri, 13 Mar 2020 at 06:41, Sowjanya Komatineni
>> <skomatineni@nvidia.com> wrote:
>>> Tegra host supports HW busy detection and timeouts based on the
>>> count programmed in SDHCI_TIMEOUT_CONTROL register and max busy
>>> timeout it supports is 11s in finite busy wait mode.
>>>
>>> Some operations like SLEEP_AWAKE, ERASE and flush cache through
>>> SWITCH commands take longer than 11s and Tegra host supports
>>> infinite HW busy wait mode where HW waits forever till the card
>>> is busy without HW timeout.
>>>
>>> This patch implements Tegra specific set_timeout sdhci_ops to allow
>>> switching between finite and infinite HW busy detection wait modes
>>> based on the device command expected operation time.
>>>
>>> Signed-off-by: Sowjanya Komatineni <skomatineni@nvidia.com>
>>> ---
>>>   drivers/mmc/host/sdhci-tegra.c | 31 +++++++++++++++++++++++++++++++
>>>   1 file changed, 31 insertions(+)
>>>
>>> diff --git a/drivers/mmc/host/sdhci-tegra.c b/drivers/mmc/host/sdhci-tegra.c
>>> index a25c3a4..fa8f6a4 100644
>>> --- a/drivers/mmc/host/sdhci-tegra.c
>>> +++ b/drivers/mmc/host/sdhci-tegra.c
>>> @@ -45,6 +45,7 @@
>>>   #define SDHCI_TEGRA_CAP_OVERRIDES_DQS_TRIM_SHIFT       8
>>>
>>>   #define SDHCI_TEGRA_VENDOR_MISC_CTRL                   0x120
>>> +#define SDHCI_MISC_CTRL_ERASE_TIMEOUT_LIMIT            BIT(0)
>>>   #define SDHCI_MISC_CTRL_ENABLE_SDR104                  0x8
>>>   #define SDHCI_MISC_CTRL_ENABLE_SDR50                   0x10
>>>   #define SDHCI_MISC_CTRL_ENABLE_SDHCI_SPEC_300          0x20
>>> @@ -1227,6 +1228,34 @@ static u32 sdhci_tegra_cqhci_irq(struct sdhci_host *host, u32 intmask)
>>>          return 0;
>>>   }
>>>
>>> +static void tegra_sdhci_set_timeout(struct sdhci_host *host,
>>> +                                   struct mmc_command *cmd)
>>> +{
>>> +       u32 val;
>>> +
>>> +       /*
>>> +        * HW busy detection timeout is based on programmed data timeout
>>> +        * counter and maximum supported timeout is 11s which may not be
>>> +        * enough for long operations like cache flush, sleep awake, erase.
>>> +        *
>>> +        * ERASE_TIMEOUT_LIMIT bit of VENDOR_MISC_CTRL register allows
>>> +        * host controller to wait for busy state until the card is busy
>>> +        * without HW timeout.
>>> +        *
>>> +        * So, use infinite busy wait mode for operations that may take
>>> +        * more than maximum HW busy timeout of 11s otherwise use finite
>>> +        * busy wait mode.
>>> +        */
>>> +       val = sdhci_readl(host, SDHCI_TEGRA_VENDOR_MISC_CTRL);
>>> +       if (cmd && cmd->busy_timeout >= 11 * HZ)
>>> +               val |= SDHCI_MISC_CTRL_ERASE_TIMEOUT_LIMIT;
>>> +       else
>>> +               val &= ~SDHCI_MISC_CTRL_ERASE_TIMEOUT_LIMIT;
>>> +       sdhci_writel(host, val, SDHCI_TEGRA_VENDOR_MISC_CTRL);
>>> +
>>> +       __sdhci_set_timeout(host, cmd);
>> kernel build on arm and arm64 architecture failed on stable-rc 4.19
>> (arm), 5.4 (arm64) and 5.5 (arm64)
>>
>> drivers/mmc/host/sdhci-tegra.c: In function 'tegra_sdhci_set_timeout':
>> drivers/mmc/host/sdhci-tegra.c:1256:2: error: implicit declaration of
>> function '__sdhci_set_timeout'; did you mean
>> 'tegra_sdhci_set_timeout'? [-Werror=implicit-function-declaration]
>>    __sdhci_set_timeout(host, cmd);
>>    ^~~~~~~~~~~~~~~~~~~
>>    tegra_sdhci_set_timeout
>>
>> Full build log,
>> https://ci.linaro.org/view/lkft/job/openembedded-lkft-linux-stable-rc-5.5/DISTRO=lkft,MACHINE=am57xx-evm,label=docker-lkft/83/consoleText
>> https://ci.linaro.org/view/lkft/job/openembedded-lkft-linux-stable-rc-5.4/DISTRO=lkft,MACHINE=juno,label=docker-lkft/158/consoleText
>> https://ci.linaro.org/view/lkft/job/openembedded-lkft-linux-stable-rc-4.19/DISTRO=lkft,MACHINE=am57xx-evm,label=docker-lkft/511/consoleText
>>
>> - Naresh
> Thanks for reporting! What a mess.
>
> It turns out that the commit that was queued for stable that is
> causing the above errors, also requires another commit.
>
> The commit that was queued:
> 5e958e4aacf4 ("sdhci: tegra: Implement Tegra specific set_timeout callback")
>
> The additional commit needed (which was added in v5.6-rc1):
> 7d76ed77cfbd ("mmc: sdhci: Refactor sdhci_set_timeout()")
>
> However, the above commit needs a manual backport (quite trivial, but
> still) for the relevant stable kernels, to allow it to solve the build
> problems.
>
> Greg, Sasha - I suggest you to drop the offending commit from the
> stable kernels, for now. I think it's better to let Sowjanya deal with
> the backports, then send them in small series instead.
>
> Kind regards
> Uffe

Hi Ufee,

Will back-porting below commit cause any issues to other vendors?

7d76ed77cfbd ("mmc: sdhci: Refactor sdhci_set_timeout()")

Thanks
Sowjanya


^ permalink raw reply	[flat|nested] 16+ messages in thread

* Re: [PATCH v2 1/2] sdhci: tegra: Implement Tegra specific set_timeout callback
  2020-04-16 10:59   ` Ulf Hansson
@ 2020-04-16 11:37     ` Greg Kroah-Hartman
  2020-04-16 16:29     ` Sowjanya Komatineni
  1 sibling, 0 replies; 16+ messages in thread
From: Greg Kroah-Hartman @ 2020-04-16 11:37 UTC (permalink / raw)
  To: Ulf Hansson
  Cc: Naresh Kamboju, Sasha Levin, Sowjanya Komatineni, Adrian Hunter,
	(Exiting) Baolin Wang, Kate Stewart, Thomas Gleixner,
	Bradley Bolen, Thierry Reding, Jon Hunter, Aniruddha Tvs Rao,
	linux-tegra, open list, linux-mmc, lkft-triage, linux- stable

On Thu, Apr 16, 2020 at 12:59:06PM +0200, Ulf Hansson wrote:
> On Wed, 15 Apr 2020 at 19:55, Naresh Kamboju <naresh.kamboju@linaro.org> wrote:
> >
> > On Fri, 13 Mar 2020 at 06:41, Sowjanya Komatineni
> > <skomatineni@nvidia.com> wrote:
> > >
> > > Tegra host supports HW busy detection and timeouts based on the
> > > count programmed in SDHCI_TIMEOUT_CONTROL register and max busy
> > > timeout it supports is 11s in finite busy wait mode.
> > >
> > > Some operations like SLEEP_AWAKE, ERASE and flush cache through
> > > SWITCH commands take longer than 11s and Tegra host supports
> > > infinite HW busy wait mode where HW waits forever till the card
> > > is busy without HW timeout.
> > >
> > > This patch implements Tegra specific set_timeout sdhci_ops to allow
> > > switching between finite and infinite HW busy detection wait modes
> > > based on the device command expected operation time.
> > >
> > > Signed-off-by: Sowjanya Komatineni <skomatineni@nvidia.com>
> > > ---
> > >  drivers/mmc/host/sdhci-tegra.c | 31 +++++++++++++++++++++++++++++++
> > >  1 file changed, 31 insertions(+)
> > >
> > > diff --git a/drivers/mmc/host/sdhci-tegra.c b/drivers/mmc/host/sdhci-tegra.c
> > > index a25c3a4..fa8f6a4 100644
> > > --- a/drivers/mmc/host/sdhci-tegra.c
> > > +++ b/drivers/mmc/host/sdhci-tegra.c
> > > @@ -45,6 +45,7 @@
> > >  #define SDHCI_TEGRA_CAP_OVERRIDES_DQS_TRIM_SHIFT       8
> > >
> > >  #define SDHCI_TEGRA_VENDOR_MISC_CTRL                   0x120
> > > +#define SDHCI_MISC_CTRL_ERASE_TIMEOUT_LIMIT            BIT(0)
> > >  #define SDHCI_MISC_CTRL_ENABLE_SDR104                  0x8
> >
> > >  #define SDHCI_MISC_CTRL_ENABLE_SDR50                   0x10
> > >  #define SDHCI_MISC_CTRL_ENABLE_SDHCI_SPEC_300          0x20
> > > @@ -1227,6 +1228,34 @@ static u32 sdhci_tegra_cqhci_irq(struct sdhci_host *host, u32 intmask)
> > >         return 0;
> > >  }
> > >
> > > +static void tegra_sdhci_set_timeout(struct sdhci_host *host,
> > > +                                   struct mmc_command *cmd)
> > > +{
> > > +       u32 val;
> > > +
> > > +       /*
> > > +        * HW busy detection timeout is based on programmed data timeout
> > > +        * counter and maximum supported timeout is 11s which may not be
> > > +        * enough for long operations like cache flush, sleep awake, erase.
> > > +        *
> > > +        * ERASE_TIMEOUT_LIMIT bit of VENDOR_MISC_CTRL register allows
> > > +        * host controller to wait for busy state until the card is busy
> > > +        * without HW timeout.
> > > +        *
> > > +        * So, use infinite busy wait mode for operations that may take
> > > +        * more than maximum HW busy timeout of 11s otherwise use finite
> > > +        * busy wait mode.
> > > +        */
> > > +       val = sdhci_readl(host, SDHCI_TEGRA_VENDOR_MISC_CTRL);
> > > +       if (cmd && cmd->busy_timeout >= 11 * HZ)
> > > +               val |= SDHCI_MISC_CTRL_ERASE_TIMEOUT_LIMIT;
> > > +       else
> > > +               val &= ~SDHCI_MISC_CTRL_ERASE_TIMEOUT_LIMIT;
> > > +       sdhci_writel(host, val, SDHCI_TEGRA_VENDOR_MISC_CTRL);
> > > +
> > > +       __sdhci_set_timeout(host, cmd);
> >
> > kernel build on arm and arm64 architecture failed on stable-rc 4.19
> > (arm), 5.4 (arm64) and 5.5 (arm64)
> >
> > drivers/mmc/host/sdhci-tegra.c: In function 'tegra_sdhci_set_timeout':
> > drivers/mmc/host/sdhci-tegra.c:1256:2: error: implicit declaration of
> > function '__sdhci_set_timeout'; did you mean
> > 'tegra_sdhci_set_timeout'? [-Werror=implicit-function-declaration]
> >   __sdhci_set_timeout(host, cmd);
> >   ^~~~~~~~~~~~~~~~~~~
> >   tegra_sdhci_set_timeout
> >
> > Full build log,
> > https://ci.linaro.org/view/lkft/job/openembedded-lkft-linux-stable-rc-5.5/DISTRO=lkft,MACHINE=am57xx-evm,label=docker-lkft/83/consoleText
> > https://ci.linaro.org/view/lkft/job/openembedded-lkft-linux-stable-rc-5.4/DISTRO=lkft,MACHINE=juno,label=docker-lkft/158/consoleText
> > https://ci.linaro.org/view/lkft/job/openembedded-lkft-linux-stable-rc-4.19/DISTRO=lkft,MACHINE=am57xx-evm,label=docker-lkft/511/consoleText
> >
> > - Naresh
> 
> Thanks for reporting! What a mess.
> 
> It turns out that the commit that was queued for stable that is
> causing the above errors, also requires another commit.
> 
> The commit that was queued:
> 5e958e4aacf4 ("sdhci: tegra: Implement Tegra specific set_timeout callback")
> 
> The additional commit needed (which was added in v5.6-rc1):
> 7d76ed77cfbd ("mmc: sdhci: Refactor sdhci_set_timeout()")
> 
> However, the above commit needs a manual backport (quite trivial, but
> still) for the relevant stable kernels, to allow it to solve the build
> problems.
> 
> Greg, Sasha - I suggest you to drop the offending commit from the
> stable kernels, for now. I think it's better to let Sowjanya deal with
> the backports, then send them in small series instead.

Thanks for this, now dropped.

greg k-h

^ permalink raw reply	[flat|nested] 16+ messages in thread

* Re: [PATCH v2 1/2] sdhci: tegra: Implement Tegra specific set_timeout callback
  2020-04-15 17:55 ` Naresh Kamboju
@ 2020-04-16 10:59   ` Ulf Hansson
  2020-04-16 11:37     ` Greg Kroah-Hartman
  2020-04-16 16:29     ` Sowjanya Komatineni
  0 siblings, 2 replies; 16+ messages in thread
From: Ulf Hansson @ 2020-04-16 10:59 UTC (permalink / raw)
  To: Naresh Kamboju, Greg Kroah-Hartman, Sasha Levin, Sowjanya Komatineni
  Cc: Adrian Hunter, (Exiting) Baolin Wang, Kate Stewart,
	Thomas Gleixner, Bradley Bolen, Thierry Reding, Jon Hunter,
	Aniruddha Tvs Rao, linux-tegra, open list, linux-mmc,
	lkft-triage, linux- stable

On Wed, 15 Apr 2020 at 19:55, Naresh Kamboju <naresh.kamboju@linaro.org> wrote:
>
> On Fri, 13 Mar 2020 at 06:41, Sowjanya Komatineni
> <skomatineni@nvidia.com> wrote:
> >
> > Tegra host supports HW busy detection and timeouts based on the
> > count programmed in SDHCI_TIMEOUT_CONTROL register and max busy
> > timeout it supports is 11s in finite busy wait mode.
> >
> > Some operations like SLEEP_AWAKE, ERASE and flush cache through
> > SWITCH commands take longer than 11s and Tegra host supports
> > infinite HW busy wait mode where HW waits forever till the card
> > is busy without HW timeout.
> >
> > This patch implements Tegra specific set_timeout sdhci_ops to allow
> > switching between finite and infinite HW busy detection wait modes
> > based on the device command expected operation time.
> >
> > Signed-off-by: Sowjanya Komatineni <skomatineni@nvidia.com>
> > ---
> >  drivers/mmc/host/sdhci-tegra.c | 31 +++++++++++++++++++++++++++++++
> >  1 file changed, 31 insertions(+)
> >
> > diff --git a/drivers/mmc/host/sdhci-tegra.c b/drivers/mmc/host/sdhci-tegra.c
> > index a25c3a4..fa8f6a4 100644
> > --- a/drivers/mmc/host/sdhci-tegra.c
> > +++ b/drivers/mmc/host/sdhci-tegra.c
> > @@ -45,6 +45,7 @@
> >  #define SDHCI_TEGRA_CAP_OVERRIDES_DQS_TRIM_SHIFT       8
> >
> >  #define SDHCI_TEGRA_VENDOR_MISC_CTRL                   0x120
> > +#define SDHCI_MISC_CTRL_ERASE_TIMEOUT_LIMIT            BIT(0)
> >  #define SDHCI_MISC_CTRL_ENABLE_SDR104                  0x8
>
> >  #define SDHCI_MISC_CTRL_ENABLE_SDR50                   0x10
> >  #define SDHCI_MISC_CTRL_ENABLE_SDHCI_SPEC_300          0x20
> > @@ -1227,6 +1228,34 @@ static u32 sdhci_tegra_cqhci_irq(struct sdhci_host *host, u32 intmask)
> >         return 0;
> >  }
> >
> > +static void tegra_sdhci_set_timeout(struct sdhci_host *host,
> > +                                   struct mmc_command *cmd)
> > +{
> > +       u32 val;
> > +
> > +       /*
> > +        * HW busy detection timeout is based on programmed data timeout
> > +        * counter and maximum supported timeout is 11s which may not be
> > +        * enough for long operations like cache flush, sleep awake, erase.
> > +        *
> > +        * ERASE_TIMEOUT_LIMIT bit of VENDOR_MISC_CTRL register allows
> > +        * host controller to wait for busy state until the card is busy
> > +        * without HW timeout.
> > +        *
> > +        * So, use infinite busy wait mode for operations that may take
> > +        * more than maximum HW busy timeout of 11s otherwise use finite
> > +        * busy wait mode.
> > +        */
> > +       val = sdhci_readl(host, SDHCI_TEGRA_VENDOR_MISC_CTRL);
> > +       if (cmd && cmd->busy_timeout >= 11 * HZ)
> > +               val |= SDHCI_MISC_CTRL_ERASE_TIMEOUT_LIMIT;
> > +       else
> > +               val &= ~SDHCI_MISC_CTRL_ERASE_TIMEOUT_LIMIT;
> > +       sdhci_writel(host, val, SDHCI_TEGRA_VENDOR_MISC_CTRL);
> > +
> > +       __sdhci_set_timeout(host, cmd);
>
> kernel build on arm and arm64 architecture failed on stable-rc 4.19
> (arm), 5.4 (arm64) and 5.5 (arm64)
>
> drivers/mmc/host/sdhci-tegra.c: In function 'tegra_sdhci_set_timeout':
> drivers/mmc/host/sdhci-tegra.c:1256:2: error: implicit declaration of
> function '__sdhci_set_timeout'; did you mean
> 'tegra_sdhci_set_timeout'? [-Werror=implicit-function-declaration]
>   __sdhci_set_timeout(host, cmd);
>   ^~~~~~~~~~~~~~~~~~~
>   tegra_sdhci_set_timeout
>
> Full build log,
> https://ci.linaro.org/view/lkft/job/openembedded-lkft-linux-stable-rc-5.5/DISTRO=lkft,MACHINE=am57xx-evm,label=docker-lkft/83/consoleText
> https://ci.linaro.org/view/lkft/job/openembedded-lkft-linux-stable-rc-5.4/DISTRO=lkft,MACHINE=juno,label=docker-lkft/158/consoleText
> https://ci.linaro.org/view/lkft/job/openembedded-lkft-linux-stable-rc-4.19/DISTRO=lkft,MACHINE=am57xx-evm,label=docker-lkft/511/consoleText
>
> - Naresh

Thanks for reporting! What a mess.

It turns out that the commit that was queued for stable that is
causing the above errors, also requires another commit.

The commit that was queued:
5e958e4aacf4 ("sdhci: tegra: Implement Tegra specific set_timeout callback")

The additional commit needed (which was added in v5.6-rc1):
7d76ed77cfbd ("mmc: sdhci: Refactor sdhci_set_timeout()")

However, the above commit needs a manual backport (quite trivial, but
still) for the relevant stable kernels, to allow it to solve the build
problems.

Greg, Sasha - I suggest you to drop the offending commit from the
stable kernels, for now. I think it's better to let Sowjanya deal with
the backports, then send them in small series instead.

Kind regards
Uffe

^ permalink raw reply	[flat|nested] 16+ messages in thread

* Re: [PATCH v2 1/2] sdhci: tegra: Implement Tegra specific set_timeout callback
  2020-03-11  0:20 Sowjanya Komatineni
  2020-03-12 13:08 ` Ulf Hansson
@ 2020-04-15 17:55 ` Naresh Kamboju
  2020-04-16 10:59   ` Ulf Hansson
  1 sibling, 1 reply; 16+ messages in thread
From: Naresh Kamboju @ 2020-04-15 17:55 UTC (permalink / raw)
  To: Sowjanya Komatineni, Greg Kroah-Hartman, Sasha Levin
  Cc: Adrian Hunter, Ulf Hansson, baolin.wang, Kate Stewart,
	Thomas Gleixner, bradleybolen, thierry.reding, Jon Hunter, anrao,
	linux-tegra, open list, linux-mmc, lkft-triage, linux- stable

On Fri, 13 Mar 2020 at 06:41, Sowjanya Komatineni
<skomatineni@nvidia.com> wrote:
>
> Tegra host supports HW busy detection and timeouts based on the
> count programmed in SDHCI_TIMEOUT_CONTROL register and max busy
> timeout it supports is 11s in finite busy wait mode.
>
> Some operations like SLEEP_AWAKE, ERASE and flush cache through
> SWITCH commands take longer than 11s and Tegra host supports
> infinite HW busy wait mode where HW waits forever till the card
> is busy without HW timeout.
>
> This patch implements Tegra specific set_timeout sdhci_ops to allow
> switching between finite and infinite HW busy detection wait modes
> based on the device command expected operation time.
>
> Signed-off-by: Sowjanya Komatineni <skomatineni@nvidia.com>
> ---
>  drivers/mmc/host/sdhci-tegra.c | 31 +++++++++++++++++++++++++++++++
>  1 file changed, 31 insertions(+)
>
> diff --git a/drivers/mmc/host/sdhci-tegra.c b/drivers/mmc/host/sdhci-tegra.c
> index a25c3a4..fa8f6a4 100644
> --- a/drivers/mmc/host/sdhci-tegra.c
> +++ b/drivers/mmc/host/sdhci-tegra.c
> @@ -45,6 +45,7 @@
>  #define SDHCI_TEGRA_CAP_OVERRIDES_DQS_TRIM_SHIFT       8
>
>  #define SDHCI_TEGRA_VENDOR_MISC_CTRL                   0x120
> +#define SDHCI_MISC_CTRL_ERASE_TIMEOUT_LIMIT            BIT(0)
>  #define SDHCI_MISC_CTRL_ENABLE_SDR104                  0x8

>  #define SDHCI_MISC_CTRL_ENABLE_SDR50                   0x10
>  #define SDHCI_MISC_CTRL_ENABLE_SDHCI_SPEC_300          0x20
> @@ -1227,6 +1228,34 @@ static u32 sdhci_tegra_cqhci_irq(struct sdhci_host *host, u32 intmask)
>         return 0;
>  }
>
> +static void tegra_sdhci_set_timeout(struct sdhci_host *host,
> +                                   struct mmc_command *cmd)
> +{
> +       u32 val;
> +
> +       /*
> +        * HW busy detection timeout is based on programmed data timeout
> +        * counter and maximum supported timeout is 11s which may not be
> +        * enough for long operations like cache flush, sleep awake, erase.
> +        *
> +        * ERASE_TIMEOUT_LIMIT bit of VENDOR_MISC_CTRL register allows
> +        * host controller to wait for busy state until the card is busy
> +        * without HW timeout.
> +        *
> +        * So, use infinite busy wait mode for operations that may take
> +        * more than maximum HW busy timeout of 11s otherwise use finite
> +        * busy wait mode.
> +        */
> +       val = sdhci_readl(host, SDHCI_TEGRA_VENDOR_MISC_CTRL);
> +       if (cmd && cmd->busy_timeout >= 11 * HZ)
> +               val |= SDHCI_MISC_CTRL_ERASE_TIMEOUT_LIMIT;
> +       else
> +               val &= ~SDHCI_MISC_CTRL_ERASE_TIMEOUT_LIMIT;
> +       sdhci_writel(host, val, SDHCI_TEGRA_VENDOR_MISC_CTRL);
> +
> +       __sdhci_set_timeout(host, cmd);

kernel build on arm and arm64 architecture failed on stable-rc 4.19
(arm), 5.4 (arm64) and 5.5 (arm64)

drivers/mmc/host/sdhci-tegra.c: In function 'tegra_sdhci_set_timeout':
drivers/mmc/host/sdhci-tegra.c:1256:2: error: implicit declaration of
function '__sdhci_set_timeout'; did you mean
'tegra_sdhci_set_timeout'? [-Werror=implicit-function-declaration]
  __sdhci_set_timeout(host, cmd);
  ^~~~~~~~~~~~~~~~~~~
  tegra_sdhci_set_timeout

Full build log,
https://ci.linaro.org/view/lkft/job/openembedded-lkft-linux-stable-rc-5.5/DISTRO=lkft,MACHINE=am57xx-evm,label=docker-lkft/83/consoleText
https://ci.linaro.org/view/lkft/job/openembedded-lkft-linux-stable-rc-5.4/DISTRO=lkft,MACHINE=juno,label=docker-lkft/158/consoleText
https://ci.linaro.org/view/lkft/job/openembedded-lkft-linux-stable-rc-4.19/DISTRO=lkft,MACHINE=am57xx-evm,label=docker-lkft/511/consoleText

- Naresh

^ permalink raw reply	[flat|nested] 16+ messages in thread

* Re: [PATCH v2 1/2] sdhci: tegra: Implement Tegra specific set_timeout callback
  2020-03-12 15:35     ` Ulf Hansson
@ 2020-03-13  1:19       ` Sowjanya Komatineni
  0 siblings, 0 replies; 16+ messages in thread
From: Sowjanya Komatineni @ 2020-03-13  1:19 UTC (permalink / raw)
  To: Ulf Hansson
  Cc: Adrian Hunter, Bradley Bolen, Thierry Reding, Jon Hunter,
	Aniruddha Tvs Rao, linux-tegra, Linux Kernel Mailing List,
	linux-mmc


On 3/12/20 8:35 AM, Ulf Hansson wrote:
> External email: Use caution opening links or attachments
>
>
> On Thu, 12 Mar 2020 at 16:28, Sowjanya Komatineni
> <skomatineni@nvidia.com> wrote:
>>
>> On 3/12/20 6:08 AM, Ulf Hansson wrote:
>>> External email: Use caution opening links or attachments
>>>
>>>
>>> -trimmed cc list
>>>
>>> On Thu, 12 Mar 2020 at 00:51, Sowjanya Komatineni
>>> <skomatineni@nvidia.com> wrote:
>>>> Tegra host supports HW busy detection and timeouts based on the
>>>> count programmed in SDHCI_TIMEOUT_CONTROL register and max busy
>>>> timeout it supports is 11s in finite busy wait mode.
>>>>
>>>> Some operations like SLEEP_AWAKE, ERASE and flush cache through
>>>> SWITCH commands take longer than 11s and Tegra host supports
>>>> infinite HW busy wait mode where HW waits forever till the card
>>>> is busy without HW timeout.
>>>>
>>>> This patch implements Tegra specific set_timeout sdhci_ops to allow
>>>> switching between finite and infinite HW busy detection wait modes
>>>> based on the device command expected operation time.
>>>>
>>>> Signed-off-by: Sowjanya Komatineni <skomatineni@nvidia.com>
>>> Applied for next, thanks!
>>>
>>> We should probably tag this for stable as well, don't you think?
>>>
>>> Kind regards
>>> Uffe
>> Yes, we need this for stable as well. As this is applied for next, looks
>> like can't re-send patch with tag.
>>
>> Can you please help to add tag if you don't mind?
> Yes, I will amend the change to add the stable tag, no worries!
>
> Thanks for confirming!
>
> [...]
>
> Kind regards
> Uffe

Thanks Uffe. Somehow patches I sent without mail server configured on 
git 2 days back went out now.

So, please kindly ignore v1 patches and also v2 patches that got resent 
which are actually the patches sent before configuring git mail server 
by mistake 2 days ago.

Sorry about that.

Thanks

Sowjanya


^ permalink raw reply	[flat|nested] 16+ messages in thread

* Re: [PATCH v2 1/2] sdhci: tegra: Implement Tegra specific set_timeout callback
  2020-03-12 15:28   ` Sowjanya Komatineni
@ 2020-03-12 15:35     ` Ulf Hansson
  2020-03-13  1:19       ` Sowjanya Komatineni
  0 siblings, 1 reply; 16+ messages in thread
From: Ulf Hansson @ 2020-03-12 15:35 UTC (permalink / raw)
  To: Sowjanya Komatineni
  Cc: Adrian Hunter, Bradley Bolen, Thierry Reding, Jon Hunter,
	Aniruddha Tvs Rao, linux-tegra, Linux Kernel Mailing List,
	linux-mmc

On Thu, 12 Mar 2020 at 16:28, Sowjanya Komatineni
<skomatineni@nvidia.com> wrote:
>
>
> On 3/12/20 6:08 AM, Ulf Hansson wrote:
> > External email: Use caution opening links or attachments
> >
> >
> > -trimmed cc list
> >
> > On Thu, 12 Mar 2020 at 00:51, Sowjanya Komatineni
> > <skomatineni@nvidia.com> wrote:
> >> Tegra host supports HW busy detection and timeouts based on the
> >> count programmed in SDHCI_TIMEOUT_CONTROL register and max busy
> >> timeout it supports is 11s in finite busy wait mode.
> >>
> >> Some operations like SLEEP_AWAKE, ERASE and flush cache through
> >> SWITCH commands take longer than 11s and Tegra host supports
> >> infinite HW busy wait mode where HW waits forever till the card
> >> is busy without HW timeout.
> >>
> >> This patch implements Tegra specific set_timeout sdhci_ops to allow
> >> switching between finite and infinite HW busy detection wait modes
> >> based on the device command expected operation time.
> >>
> >> Signed-off-by: Sowjanya Komatineni <skomatineni@nvidia.com>
> > Applied for next, thanks!
> >
> > We should probably tag this for stable as well, don't you think?
> >
> > Kind regards
> > Uffe
>
> Yes, we need this for stable as well. As this is applied for next, looks
> like can't re-send patch with tag.
>
> Can you please help to add tag if you don't mind?

Yes, I will amend the change to add the stable tag, no worries!

Thanks for confirming!

[...]

Kind regards
Uffe

^ permalink raw reply	[flat|nested] 16+ messages in thread

* Re: [PATCH v2 1/2] sdhci: tegra: Implement Tegra specific set_timeout callback
  2020-03-12 13:08 ` Ulf Hansson
@ 2020-03-12 15:28   ` Sowjanya Komatineni
  2020-03-12 15:35     ` Ulf Hansson
  0 siblings, 1 reply; 16+ messages in thread
From: Sowjanya Komatineni @ 2020-03-12 15:28 UTC (permalink / raw)
  To: Ulf Hansson
  Cc: Adrian Hunter, Bradley Bolen, Thierry Reding, Jon Hunter,
	Aniruddha Tvs Rao, linux-tegra, Linux Kernel Mailing List,
	linux-mmc


On 3/12/20 6:08 AM, Ulf Hansson wrote:
> External email: Use caution opening links or attachments
>
>
> -trimmed cc list
>
> On Thu, 12 Mar 2020 at 00:51, Sowjanya Komatineni
> <skomatineni@nvidia.com> wrote:
>> Tegra host supports HW busy detection and timeouts based on the
>> count programmed in SDHCI_TIMEOUT_CONTROL register and max busy
>> timeout it supports is 11s in finite busy wait mode.
>>
>> Some operations like SLEEP_AWAKE, ERASE and flush cache through
>> SWITCH commands take longer than 11s and Tegra host supports
>> infinite HW busy wait mode where HW waits forever till the card
>> is busy without HW timeout.
>>
>> This patch implements Tegra specific set_timeout sdhci_ops to allow
>> switching between finite and infinite HW busy detection wait modes
>> based on the device command expected operation time.
>>
>> Signed-off-by: Sowjanya Komatineni <skomatineni@nvidia.com>
> Applied for next, thanks!
>
> We should probably tag this for stable as well, don't you think?
>
> Kind regards
> Uffe

Yes, we need this for stable as well. As this is applied for next, looks 
like can't re-send patch with tag.

Can you please help to add tag if you don't mind?

>
>
>> ---
>>   drivers/mmc/host/sdhci-tegra.c | 31 +++++++++++++++++++++++++++++++
>>   1 file changed, 31 insertions(+)
>>
>> diff --git a/drivers/mmc/host/sdhci-tegra.c b/drivers/mmc/host/sdhci-tegra.c
>> index a25c3a4..fa8f6a4 100644
>> --- a/drivers/mmc/host/sdhci-tegra.c
>> +++ b/drivers/mmc/host/sdhci-tegra.c
>> @@ -45,6 +45,7 @@
>>   #define SDHCI_TEGRA_CAP_OVERRIDES_DQS_TRIM_SHIFT       8
>>
>>   #define SDHCI_TEGRA_VENDOR_MISC_CTRL                   0x120
>> +#define SDHCI_MISC_CTRL_ERASE_TIMEOUT_LIMIT            BIT(0)
>>   #define SDHCI_MISC_CTRL_ENABLE_SDR104                  0x8
>>   #define SDHCI_MISC_CTRL_ENABLE_SDR50                   0x10
>>   #define SDHCI_MISC_CTRL_ENABLE_SDHCI_SPEC_300          0x20
>> @@ -1227,6 +1228,34 @@ static u32 sdhci_tegra_cqhci_irq(struct sdhci_host *host, u32 intmask)
>>          return 0;
>>   }
>>
>> +static void tegra_sdhci_set_timeout(struct sdhci_host *host,
>> +                                   struct mmc_command *cmd)
>> +{
>> +       u32 val;
>> +
>> +       /*
>> +        * HW busy detection timeout is based on programmed data timeout
>> +        * counter and maximum supported timeout is 11s which may not be
>> +        * enough for long operations like cache flush, sleep awake, erase.
>> +        *
>> +        * ERASE_TIMEOUT_LIMIT bit of VENDOR_MISC_CTRL register allows
>> +        * host controller to wait for busy state until the card is busy
>> +        * without HW timeout.
>> +        *
>> +        * So, use infinite busy wait mode for operations that may take
>> +        * more than maximum HW busy timeout of 11s otherwise use finite
>> +        * busy wait mode.
>> +        */
>> +       val = sdhci_readl(host, SDHCI_TEGRA_VENDOR_MISC_CTRL);
>> +       if (cmd && cmd->busy_timeout >= 11 * HZ)
>> +               val |= SDHCI_MISC_CTRL_ERASE_TIMEOUT_LIMIT;
>> +       else
>> +               val &= ~SDHCI_MISC_CTRL_ERASE_TIMEOUT_LIMIT;
>> +       sdhci_writel(host, val, SDHCI_TEGRA_VENDOR_MISC_CTRL);
>> +
>> +       __sdhci_set_timeout(host, cmd);
>> +}
>> +
>>   static const struct cqhci_host_ops sdhci_tegra_cqhci_ops = {
>>          .write_l    = tegra_cqhci_writel,
>>          .enable = sdhci_tegra_cqe_enable,
>> @@ -1366,6 +1395,7 @@ static const struct sdhci_ops tegra210_sdhci_ops = {
>>          .set_uhs_signaling = tegra_sdhci_set_uhs_signaling,
>>          .voltage_switch = tegra_sdhci_voltage_switch,
>>          .get_max_clock = tegra_sdhci_get_max_clock,
>> +       .set_timeout = tegra_sdhci_set_timeout,
>>   };
>>
>>   static const struct sdhci_pltfm_data sdhci_tegra210_pdata = {
>> @@ -1403,6 +1433,7 @@ static const struct sdhci_ops tegra186_sdhci_ops = {
>>          .voltage_switch = tegra_sdhci_voltage_switch,
>>          .get_max_clock = tegra_sdhci_get_max_clock,
>>          .irq = sdhci_tegra_cqhci_irq,
>> +       .set_timeout = tegra_sdhci_set_timeout,
>>   };
>>
>>   static const struct sdhci_pltfm_data sdhci_tegra186_pdata = {
>> --
>> 2.7.4
>>

^ permalink raw reply	[flat|nested] 16+ messages in thread

* Re: [PATCH v2 1/2] sdhci: tegra: Implement Tegra specific set_timeout callback
  2020-03-11  0:20 Sowjanya Komatineni
@ 2020-03-12 13:08 ` Ulf Hansson
  2020-03-12 15:28   ` Sowjanya Komatineni
  2020-04-15 17:55 ` Naresh Kamboju
  1 sibling, 1 reply; 16+ messages in thread
From: Ulf Hansson @ 2020-03-12 13:08 UTC (permalink / raw)
  To: Sowjanya Komatineni
  Cc: Adrian Hunter, Bradley Bolen, Thierry Reding, Jon Hunter,
	Aniruddha Tvs Rao, linux-tegra, Linux Kernel Mailing List,
	linux-mmc

-trimmed cc list

On Thu, 12 Mar 2020 at 00:51, Sowjanya Komatineni
<skomatineni@nvidia.com> wrote:
>
> Tegra host supports HW busy detection and timeouts based on the
> count programmed in SDHCI_TIMEOUT_CONTROL register and max busy
> timeout it supports is 11s in finite busy wait mode.
>
> Some operations like SLEEP_AWAKE, ERASE and flush cache through
> SWITCH commands take longer than 11s and Tegra host supports
> infinite HW busy wait mode where HW waits forever till the card
> is busy without HW timeout.
>
> This patch implements Tegra specific set_timeout sdhci_ops to allow
> switching between finite and infinite HW busy detection wait modes
> based on the device command expected operation time.
>
> Signed-off-by: Sowjanya Komatineni <skomatineni@nvidia.com>

Applied for next, thanks!

We should probably tag this for stable as well, don't you think?

Kind regards
Uffe



> ---
>  drivers/mmc/host/sdhci-tegra.c | 31 +++++++++++++++++++++++++++++++
>  1 file changed, 31 insertions(+)
>
> diff --git a/drivers/mmc/host/sdhci-tegra.c b/drivers/mmc/host/sdhci-tegra.c
> index a25c3a4..fa8f6a4 100644
> --- a/drivers/mmc/host/sdhci-tegra.c
> +++ b/drivers/mmc/host/sdhci-tegra.c
> @@ -45,6 +45,7 @@
>  #define SDHCI_TEGRA_CAP_OVERRIDES_DQS_TRIM_SHIFT       8
>
>  #define SDHCI_TEGRA_VENDOR_MISC_CTRL                   0x120
> +#define SDHCI_MISC_CTRL_ERASE_TIMEOUT_LIMIT            BIT(0)
>  #define SDHCI_MISC_CTRL_ENABLE_SDR104                  0x8
>  #define SDHCI_MISC_CTRL_ENABLE_SDR50                   0x10
>  #define SDHCI_MISC_CTRL_ENABLE_SDHCI_SPEC_300          0x20
> @@ -1227,6 +1228,34 @@ static u32 sdhci_tegra_cqhci_irq(struct sdhci_host *host, u32 intmask)
>         return 0;
>  }
>
> +static void tegra_sdhci_set_timeout(struct sdhci_host *host,
> +                                   struct mmc_command *cmd)
> +{
> +       u32 val;
> +
> +       /*
> +        * HW busy detection timeout is based on programmed data timeout
> +        * counter and maximum supported timeout is 11s which may not be
> +        * enough for long operations like cache flush, sleep awake, erase.
> +        *
> +        * ERASE_TIMEOUT_LIMIT bit of VENDOR_MISC_CTRL register allows
> +        * host controller to wait for busy state until the card is busy
> +        * without HW timeout.
> +        *
> +        * So, use infinite busy wait mode for operations that may take
> +        * more than maximum HW busy timeout of 11s otherwise use finite
> +        * busy wait mode.
> +        */
> +       val = sdhci_readl(host, SDHCI_TEGRA_VENDOR_MISC_CTRL);
> +       if (cmd && cmd->busy_timeout >= 11 * HZ)
> +               val |= SDHCI_MISC_CTRL_ERASE_TIMEOUT_LIMIT;
> +       else
> +               val &= ~SDHCI_MISC_CTRL_ERASE_TIMEOUT_LIMIT;
> +       sdhci_writel(host, val, SDHCI_TEGRA_VENDOR_MISC_CTRL);
> +
> +       __sdhci_set_timeout(host, cmd);
> +}
> +
>  static const struct cqhci_host_ops sdhci_tegra_cqhci_ops = {
>         .write_l    = tegra_cqhci_writel,
>         .enable = sdhci_tegra_cqe_enable,
> @@ -1366,6 +1395,7 @@ static const struct sdhci_ops tegra210_sdhci_ops = {
>         .set_uhs_signaling = tegra_sdhci_set_uhs_signaling,
>         .voltage_switch = tegra_sdhci_voltage_switch,
>         .get_max_clock = tegra_sdhci_get_max_clock,
> +       .set_timeout = tegra_sdhci_set_timeout,
>  };
>
>  static const struct sdhci_pltfm_data sdhci_tegra210_pdata = {
> @@ -1403,6 +1433,7 @@ static const struct sdhci_ops tegra186_sdhci_ops = {
>         .voltage_switch = tegra_sdhci_voltage_switch,
>         .get_max_clock = tegra_sdhci_get_max_clock,
>         .irq = sdhci_tegra_cqhci_irq,
> +       .set_timeout = tegra_sdhci_set_timeout,
>  };
>
>  static const struct sdhci_pltfm_data sdhci_tegra186_pdata = {
> --
> 2.7.4
>

^ permalink raw reply	[flat|nested] 16+ messages in thread

* [PATCH v2 1/2] sdhci: tegra: Implement Tegra specific set_timeout callback
@ 2020-03-11  8:56 Sowjanya Komatineni
  0 siblings, 0 replies; 16+ messages in thread
From: Sowjanya Komatineni @ 2020-03-11  8:56 UTC (permalink / raw)
  To: adrian.hunter, ulf.hansson, baolin.wang, kstewart, tglx,
	bradleybolen, gregkh, thierry.reding, jonathanh, skomatineni
  Cc: anrao, linux-tegra, linux-kernel, linux-mmc

Tegra host supports HW busy detection and timeouts based on the
count programmed in SDHCI_TIMEOUT_CONTROL register and max busy
timeout it supports is 11s in finite busy wait mode.

Some operations like SLEEP_AWAKE, ERASE and flush cache through
SWITCH commands take longer than 11s and Tegra host supports
infinite HW busy wait mode where HW waits forever till the card
is busy without HW timeout.

This patch implements Tegra specific set_timeout sdhci_ops to allow
switching between finite and infinite HW busy detection wait modes
based on the device command expected operation time.

Signed-off-by: Sowjanya Komatineni <skomatineni@nvidia.com>
---
 drivers/mmc/host/sdhci-tegra.c | 31 +++++++++++++++++++++++++++++++
 1 file changed, 31 insertions(+)

diff --git a/drivers/mmc/host/sdhci-tegra.c b/drivers/mmc/host/sdhci-tegra.c
index a25c3a4..fa8f6a4 100644
--- a/drivers/mmc/host/sdhci-tegra.c
+++ b/drivers/mmc/host/sdhci-tegra.c
@@ -45,6 +45,7 @@
 #define SDHCI_TEGRA_CAP_OVERRIDES_DQS_TRIM_SHIFT	8
 
 #define SDHCI_TEGRA_VENDOR_MISC_CTRL			0x120
+#define SDHCI_MISC_CTRL_ERASE_TIMEOUT_LIMIT		BIT(0)
 #define SDHCI_MISC_CTRL_ENABLE_SDR104			0x8
 #define SDHCI_MISC_CTRL_ENABLE_SDR50			0x10
 #define SDHCI_MISC_CTRL_ENABLE_SDHCI_SPEC_300		0x20
@@ -1227,6 +1228,34 @@ static u32 sdhci_tegra_cqhci_irq(struct sdhci_host *host, u32 intmask)
 	return 0;
 }
 
+static void tegra_sdhci_set_timeout(struct sdhci_host *host,
+				    struct mmc_command *cmd)
+{
+	u32 val;
+
+	/*
+	 * HW busy detection timeout is based on programmed data timeout
+	 * counter and maximum supported timeout is 11s which may not be
+	 * enough for long operations like cache flush, sleep awake, erase.
+	 *
+	 * ERASE_TIMEOUT_LIMIT bit of VENDOR_MISC_CTRL register allows
+	 * host controller to wait for busy state until the card is busy
+	 * without HW timeout.
+	 *
+	 * So, use infinite busy wait mode for operations that may take
+	 * more than maximum HW busy timeout of 11s otherwise use finite
+	 * busy wait mode.
+	 */
+	val = sdhci_readl(host, SDHCI_TEGRA_VENDOR_MISC_CTRL);
+	if (cmd && cmd->busy_timeout >= 11 * HZ)
+		val |= SDHCI_MISC_CTRL_ERASE_TIMEOUT_LIMIT;
+	else
+		val &= ~SDHCI_MISC_CTRL_ERASE_TIMEOUT_LIMIT;
+	sdhci_writel(host, val, SDHCI_TEGRA_VENDOR_MISC_CTRL);
+
+	__sdhci_set_timeout(host, cmd);
+}
+
 static const struct cqhci_host_ops sdhci_tegra_cqhci_ops = {
 	.write_l    = tegra_cqhci_writel,
 	.enable	= sdhci_tegra_cqe_enable,
@@ -1366,6 +1395,7 @@ static const struct sdhci_ops tegra210_sdhci_ops = {
 	.set_uhs_signaling = tegra_sdhci_set_uhs_signaling,
 	.voltage_switch = tegra_sdhci_voltage_switch,
 	.get_max_clock = tegra_sdhci_get_max_clock,
+	.set_timeout = tegra_sdhci_set_timeout,
 };
 
 static const struct sdhci_pltfm_data sdhci_tegra210_pdata = {
@@ -1403,6 +1433,7 @@ static const struct sdhci_ops tegra186_sdhci_ops = {
 	.voltage_switch = tegra_sdhci_voltage_switch,
 	.get_max_clock = tegra_sdhci_get_max_clock,
 	.irq = sdhci_tegra_cqhci_irq,
+	.set_timeout = tegra_sdhci_set_timeout,
 };
 
 static const struct sdhci_pltfm_data sdhci_tegra186_pdata = {
-- 
2.7.4


^ permalink raw reply	[flat|nested] 16+ messages in thread

* [PATCH v2 1/2] sdhci: tegra: Implement Tegra specific set_timeout callback
@ 2020-03-11  8:54 Sowjanya Komatineni
  0 siblings, 0 replies; 16+ messages in thread
From: Sowjanya Komatineni @ 2020-03-11  8:54 UTC (permalink / raw)
  To: adrian.hunter, ulf.hansson, baolin.wang, kstewart, tglx,
	bradleybolen, gregkh, thierry.reding, jonathanh, skomatineni
  Cc: anrao, linux-tegra, linux-kernel, linux-mmc

Tegra host supports HW busy detection and timeouts based on the
count programmed in SDHCI_TIMEOUT_CONTROL register and max busy
timeout it supports is 11s in finite busy wait mode.

Some operations like SLEEP_AWAKE, ERASE and flush cache through
SWITCH commands take longer than 11s and Tegra host supports
infinite HW busy wait mode where HW waits forever till the card
is busy without HW timeout.

This patch implements Tegra specific set_timeout sdhci_ops to allow
switching between finite and infinite HW busy detection wait modes
based on the device command expected operation time.

Signed-off-by: Sowjanya Komatineni <skomatineni@nvidia.com>
---
 drivers/mmc/host/sdhci-tegra.c | 31 +++++++++++++++++++++++++++++++
 1 file changed, 31 insertions(+)

diff --git a/drivers/mmc/host/sdhci-tegra.c b/drivers/mmc/host/sdhci-tegra.c
index a25c3a4..fa8f6a4 100644
--- a/drivers/mmc/host/sdhci-tegra.c
+++ b/drivers/mmc/host/sdhci-tegra.c
@@ -45,6 +45,7 @@
 #define SDHCI_TEGRA_CAP_OVERRIDES_DQS_TRIM_SHIFT	8
 
 #define SDHCI_TEGRA_VENDOR_MISC_CTRL			0x120
+#define SDHCI_MISC_CTRL_ERASE_TIMEOUT_LIMIT		BIT(0)
 #define SDHCI_MISC_CTRL_ENABLE_SDR104			0x8
 #define SDHCI_MISC_CTRL_ENABLE_SDR50			0x10
 #define SDHCI_MISC_CTRL_ENABLE_SDHCI_SPEC_300		0x20
@@ -1227,6 +1228,34 @@ static u32 sdhci_tegra_cqhci_irq(struct sdhci_host *host, u32 intmask)
 	return 0;
 }
 
+static void tegra_sdhci_set_timeout(struct sdhci_host *host,
+				    struct mmc_command *cmd)
+{
+	u32 val;
+
+	/*
+	 * HW busy detection timeout is based on programmed data timeout
+	 * counter and maximum supported timeout is 11s which may not be
+	 * enough for long operations like cache flush, sleep awake, erase.
+	 *
+	 * ERASE_TIMEOUT_LIMIT bit of VENDOR_MISC_CTRL register allows
+	 * host controller to wait for busy state until the card is busy
+	 * without HW timeout.
+	 *
+	 * So, use infinite busy wait mode for operations that may take
+	 * more than maximum HW busy timeout of 11s otherwise use finite
+	 * busy wait mode.
+	 */
+	val = sdhci_readl(host, SDHCI_TEGRA_VENDOR_MISC_CTRL);
+	if (cmd && cmd->busy_timeout >= 11 * HZ)
+		val |= SDHCI_MISC_CTRL_ERASE_TIMEOUT_LIMIT;
+	else
+		val &= ~SDHCI_MISC_CTRL_ERASE_TIMEOUT_LIMIT;
+	sdhci_writel(host, val, SDHCI_TEGRA_VENDOR_MISC_CTRL);
+
+	__sdhci_set_timeout(host, cmd);
+}
+
 static const struct cqhci_host_ops sdhci_tegra_cqhci_ops = {
 	.write_l    = tegra_cqhci_writel,
 	.enable	= sdhci_tegra_cqe_enable,
@@ -1366,6 +1395,7 @@ static const struct sdhci_ops tegra210_sdhci_ops = {
 	.set_uhs_signaling = tegra_sdhci_set_uhs_signaling,
 	.voltage_switch = tegra_sdhci_voltage_switch,
 	.get_max_clock = tegra_sdhci_get_max_clock,
+	.set_timeout = tegra_sdhci_set_timeout,
 };
 
 static const struct sdhci_pltfm_data sdhci_tegra210_pdata = {
@@ -1403,6 +1433,7 @@ static const struct sdhci_ops tegra186_sdhci_ops = {
 	.voltage_switch = tegra_sdhci_voltage_switch,
 	.get_max_clock = tegra_sdhci_get_max_clock,
 	.irq = sdhci_tegra_cqhci_irq,
+	.set_timeout = tegra_sdhci_set_timeout,
 };
 
 static const struct sdhci_pltfm_data sdhci_tegra186_pdata = {
-- 
2.7.4


^ permalink raw reply	[flat|nested] 16+ messages in thread

* [PATCH v2 1/2] sdhci: tegra: Implement Tegra specific set_timeout callback
@ 2020-03-11  0:20 Sowjanya Komatineni
  2020-03-12 13:08 ` Ulf Hansson
  2020-04-15 17:55 ` Naresh Kamboju
  0 siblings, 2 replies; 16+ messages in thread
From: Sowjanya Komatineni @ 2020-03-11  0:20 UTC (permalink / raw)
  To: adrian.hunter, ulf.hansson, baolin.wang, kstewart, tglx,
	bradleybolen, gregkh, thierry.reding, jonathanh
  Cc: skomatineni, anrao, linux-tegra, linux-kernel, linux-mmc

Tegra host supports HW busy detection and timeouts based on the
count programmed in SDHCI_TIMEOUT_CONTROL register and max busy
timeout it supports is 11s in finite busy wait mode.

Some operations like SLEEP_AWAKE, ERASE and flush cache through
SWITCH commands take longer than 11s and Tegra host supports
infinite HW busy wait mode where HW waits forever till the card
is busy without HW timeout.

This patch implements Tegra specific set_timeout sdhci_ops to allow
switching between finite and infinite HW busy detection wait modes
based on the device command expected operation time.

Signed-off-by: Sowjanya Komatineni <skomatineni@nvidia.com>
---
 drivers/mmc/host/sdhci-tegra.c | 31 +++++++++++++++++++++++++++++++
 1 file changed, 31 insertions(+)

diff --git a/drivers/mmc/host/sdhci-tegra.c b/drivers/mmc/host/sdhci-tegra.c
index a25c3a4..fa8f6a4 100644
--- a/drivers/mmc/host/sdhci-tegra.c
+++ b/drivers/mmc/host/sdhci-tegra.c
@@ -45,6 +45,7 @@
 #define SDHCI_TEGRA_CAP_OVERRIDES_DQS_TRIM_SHIFT	8
 
 #define SDHCI_TEGRA_VENDOR_MISC_CTRL			0x120
+#define SDHCI_MISC_CTRL_ERASE_TIMEOUT_LIMIT		BIT(0)
 #define SDHCI_MISC_CTRL_ENABLE_SDR104			0x8
 #define SDHCI_MISC_CTRL_ENABLE_SDR50			0x10
 #define SDHCI_MISC_CTRL_ENABLE_SDHCI_SPEC_300		0x20
@@ -1227,6 +1228,34 @@ static u32 sdhci_tegra_cqhci_irq(struct sdhci_host *host, u32 intmask)
 	return 0;
 }
 
+static void tegra_sdhci_set_timeout(struct sdhci_host *host,
+				    struct mmc_command *cmd)
+{
+	u32 val;
+
+	/*
+	 * HW busy detection timeout is based on programmed data timeout
+	 * counter and maximum supported timeout is 11s which may not be
+	 * enough for long operations like cache flush, sleep awake, erase.
+	 *
+	 * ERASE_TIMEOUT_LIMIT bit of VENDOR_MISC_CTRL register allows
+	 * host controller to wait for busy state until the card is busy
+	 * without HW timeout.
+	 *
+	 * So, use infinite busy wait mode for operations that may take
+	 * more than maximum HW busy timeout of 11s otherwise use finite
+	 * busy wait mode.
+	 */
+	val = sdhci_readl(host, SDHCI_TEGRA_VENDOR_MISC_CTRL);
+	if (cmd && cmd->busy_timeout >= 11 * HZ)
+		val |= SDHCI_MISC_CTRL_ERASE_TIMEOUT_LIMIT;
+	else
+		val &= ~SDHCI_MISC_CTRL_ERASE_TIMEOUT_LIMIT;
+	sdhci_writel(host, val, SDHCI_TEGRA_VENDOR_MISC_CTRL);
+
+	__sdhci_set_timeout(host, cmd);
+}
+
 static const struct cqhci_host_ops sdhci_tegra_cqhci_ops = {
 	.write_l    = tegra_cqhci_writel,
 	.enable	= sdhci_tegra_cqe_enable,
@@ -1366,6 +1395,7 @@ static const struct sdhci_ops tegra210_sdhci_ops = {
 	.set_uhs_signaling = tegra_sdhci_set_uhs_signaling,
 	.voltage_switch = tegra_sdhci_voltage_switch,
 	.get_max_clock = tegra_sdhci_get_max_clock,
+	.set_timeout = tegra_sdhci_set_timeout,
 };
 
 static const struct sdhci_pltfm_data sdhci_tegra210_pdata = {
@@ -1403,6 +1433,7 @@ static const struct sdhci_ops tegra186_sdhci_ops = {
 	.voltage_switch = tegra_sdhci_voltage_switch,
 	.get_max_clock = tegra_sdhci_get_max_clock,
 	.irq = sdhci_tegra_cqhci_irq,
+	.set_timeout = tegra_sdhci_set_timeout,
 };
 
 static const struct sdhci_pltfm_data sdhci_tegra186_pdata = {
-- 
2.7.4


^ permalink raw reply	[flat|nested] 16+ messages in thread

end of thread, other threads:[~2020-04-17 18:30 UTC | newest]

Thread overview: 16+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-03-11 15:47 [PATCH v2 1/2] sdhci: tegra: Implement Tegra specific set_timeout callback Sowjanya Komatineni
2020-03-11 15:47 ` [PATCH v2 2/2] sdhci: tegra: Enable MMC_CAP_WAIT_WHILE_BUSY host capability Sowjanya Komatineni
  -- strict thread matches above, loose matches on Subject: below --
2020-03-11  8:56 [PATCH v2 1/2] sdhci: tegra: Implement Tegra specific set_timeout callback Sowjanya Komatineni
2020-03-11  8:54 Sowjanya Komatineni
2020-03-11  0:20 Sowjanya Komatineni
2020-03-12 13:08 ` Ulf Hansson
2020-03-12 15:28   ` Sowjanya Komatineni
2020-03-12 15:35     ` Ulf Hansson
2020-03-13  1:19       ` Sowjanya Komatineni
2020-04-15 17:55 ` Naresh Kamboju
2020-04-16 10:59   ` Ulf Hansson
2020-04-16 11:37     ` Greg Kroah-Hartman
2020-04-16 16:29     ` Sowjanya Komatineni
2020-04-16 19:38       ` Sowjanya Komatineni
2020-04-17  8:04         ` Ulf Hansson
2020-04-17 18:29           ` Sowjanya Komatineni

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).