linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Bhaumik Bhatt <bbhatt@codeaurora.org>
To: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>,
	mgautam@codeaurora.org
Cc: linux-arm-msm@vger.kernel.org, hemantk@codeaurora.org,
	jhugo@codeaurora.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH v1 06/10] bus: mhi: core: Improve shutdown handling after link down detection
Date: Tue, 13 Oct 2020 11:40:10 -0700	[thread overview]
Message-ID: <3793225dd0932d0b2a9bb5b3df293fce@codeaurora.org> (raw)
In-Reply-To: <20201009161950.GG4810@Mani-XPS-13-9360>

On 2020-10-09 09:19, Manivannan Sadhasivam wrote:
> On Fri, Sep 18, 2020 at 07:02:31PM -0700, Bhaumik Bhatt wrote:
>> If MHI were to attempt a device shutdown following an assumption
> 
> MHI host? And is this really an assumption or it is definite that the
> link is inaccessible. Please clarify!
> 
Will update it to MHI host.

I say assumption because we act based on the "graceful" flag passed by
the MHI controller driver. Link may be accessible but the controller has
instructed MHI not to do any further accesses. They may decide to set 
the
flag as "false" if they see any read/access issues, an actual link down
interrupt from the bus driver or a sideband GPIO signal declaring that a
software assert occurred on the device.

MHI host sees that power down attempt as ungraceful and assumes that the
controller has decided that it's either a link down or a fatal error.

Once an "ungraceful" power down attempt is made, MHI host moves to the
LD_ERR_FATAL_DETECT pm_state and without this patch, it moved from an
LD_ERR_FATAL_DETECT to SHUTDOWN_PROCESS state, where SHUTDOWN_PROCESS
is defined as a register accessible state by the MHI_REG_ACCESS_VALID()
macro.

If someone were to do a call that needed a register access from their
.remove() callback, for example, we could see a bus error.

Moves from MHI_PM_M3 to SHUTDOWN_PROCESS and LD_ERR_FATAL_DETECT to
SHUTDOWN_PROCESS are not allowed by use of this patch.

I'll add better wording and explanation.

>> that the device is inaccessible, the host currently moves to a state
>> where device register accesses are allowed when they should not be.
>> This would end up allowing accesses to the device register space when
>> the link is inaccessible and can result in bus errors observed on the
>> host. Improve shutdown handling to prevent these outcomes and do not
>> move the MHI PM state to a register accessible state after device is
>> assumed to be inaccessible.
>> 
> 
> Apparently you are introducing a new device transition state but your
> commit description doesn't state that :/
> 
> Thanks,
> Mani
> 
Sure. I will add that.

>> Signed-off-by: Bhaumik Bhatt <bbhatt@codeaurora.org>
>> ---
>>  drivers/bus/mhi/core/init.c     |  1 +
>>  drivers/bus/mhi/core/internal.h |  1 +
>>  drivers/bus/mhi/core/pm.c       | 18 +++++++++++++-----
>>  3 files changed, 15 insertions(+), 5 deletions(-)
>> 
>> diff --git a/drivers/bus/mhi/core/init.c b/drivers/bus/mhi/core/init.c
>> index 9ae4c19..fa33dde 100644
>> --- a/drivers/bus/mhi/core/init.c
>> +++ b/drivers/bus/mhi/core/init.c
>> @@ -37,6 +37,7 @@ const char * const 
>> dev_state_tran_str[DEV_ST_TRANSITION_MAX] = {
>>  	[DEV_ST_TRANSITION_MISSION_MODE] = "MISSION_MODE",
>>  	[DEV_ST_TRANSITION_SYS_ERR] = "SYS_ERR",
>>  	[DEV_ST_TRANSITION_DISABLE] = "DISABLE",
>> +	[DEV_ST_TRANSITION_FATAL] = "FATAL SHUTDOWN",
>>  };
>> 
>>  const char * const mhi_state_str[MHI_STATE_MAX] = {
>> diff --git a/drivers/bus/mhi/core/internal.h 
>> b/drivers/bus/mhi/core/internal.h
>> index 7989269..f3b9e5a 100644
>> --- a/drivers/bus/mhi/core/internal.h
>> +++ b/drivers/bus/mhi/core/internal.h
>> @@ -388,6 +388,7 @@ enum dev_st_transition {
>>  	DEV_ST_TRANSITION_MISSION_MODE,
>>  	DEV_ST_TRANSITION_SYS_ERR,
>>  	DEV_ST_TRANSITION_DISABLE,
>> +	DEV_ST_TRANSITION_FATAL,
>>  	DEV_ST_TRANSITION_MAX,
>>  };
>> 
>> diff --git a/drivers/bus/mhi/core/pm.c b/drivers/bus/mhi/core/pm.c
>> index 3462d82..bce1f62 100644
>> --- a/drivers/bus/mhi/core/pm.c
>> +++ b/drivers/bus/mhi/core/pm.c
>> @@ -37,9 +37,10 @@
>>   *     M0 -> FW_DL_ERR
>>   *     M0 -> M3_ENTER -> M3 -> M3_EXIT --> M0
>>   * L1: SYS_ERR_DETECT -> SYS_ERR_PROCESS --> POR
>> - * L2: SHUTDOWN_PROCESS -> DISABLE
>> + * L2: SHUTDOWN_PROCESS -> LD_ERR_FATAL_DETECT
>> + *     SHUTDOWN_PROCESS -> DISABLE
>>   * L3: LD_ERR_FATAL_DETECT <--> LD_ERR_FATAL_DETECT
>> - *     LD_ERR_FATAL_DETECT -> SHUTDOWN_PROCESS
>> + *     LD_ERR_FATAL_DETECT -> DISABLE
>>   */
>>  static struct mhi_pm_transitions const dev_state_transitions[] = {
>>  	/* L0 States */
>> @@ -72,7 +73,7 @@ static struct mhi_pm_transitions const 
>> dev_state_transitions[] = {
>>  	{
>>  		MHI_PM_M3,
>>  		MHI_PM_M3_EXIT | MHI_PM_SYS_ERR_DETECT |
>> -		MHI_PM_SHUTDOWN_PROCESS | MHI_PM_LD_ERR_FATAL_DETECT
>> +		MHI_PM_LD_ERR_FATAL_DETECT
>>  	},
>>  	{
>>  		MHI_PM_M3_EXIT,
>> @@ -103,7 +104,7 @@ static struct mhi_pm_transitions const 
>> dev_state_transitions[] = {
>>  	/* L3 States */
>>  	{
>>  		MHI_PM_LD_ERR_FATAL_DETECT,
>> -		MHI_PM_LD_ERR_FATAL_DETECT | MHI_PM_SHUTDOWN_PROCESS
>> +		MHI_PM_LD_ERR_FATAL_DETECT | MHI_PM_DISABLE
>>  	},
>>  };
>> 
>> @@ -670,6 +671,10 @@ void mhi_pm_st_worker(struct work_struct *work)
>>  			mhi_pm_disable_transition
>>  				(mhi_cntrl, MHI_PM_SHUTDOWN_PROCESS);
>>  			break;
>> +		case DEV_ST_TRANSITION_FATAL:
>> +			mhi_pm_disable_transition
>> +				(mhi_cntrl, MHI_PM_LD_ERR_FATAL_DETECT);
>> +			break;
>>  		default:
>>  			break;
>>  		}
>> @@ -1039,6 +1044,7 @@ EXPORT_SYMBOL_GPL(mhi_async_power_up);
>>  void mhi_power_down(struct mhi_controller *mhi_cntrl, bool graceful)
>>  {
>>  	enum mhi_pm_state cur_state;
>> +	enum dev_st_transition next_state = DEV_ST_TRANSITION_DISABLE;
>>  	struct device *dev = &mhi_cntrl->mhi_dev->dev;
>> 
>>  	/* If it's not a graceful shutdown, force MHI to linkdown state */
>> @@ -1053,9 +1059,11 @@ void mhi_power_down(struct mhi_controller 
>> *mhi_cntrl, bool graceful)
>>  			dev_dbg(dev, "Failed to move to state: %s from: %s\n",
>>  				to_mhi_pm_state_str(MHI_PM_LD_ERR_FATAL_DETECT),
>>  				to_mhi_pm_state_str(mhi_cntrl->pm_state));
>> +		else
>> +			next_state = DEV_ST_TRANSITION_FATAL;
>>  	}
>> 
>> -	mhi_queue_state_transition(mhi_cntrl, DEV_ST_TRANSITION_DISABLE);
>> +	mhi_queue_state_transition(mhi_cntrl, next_state);
>> 
>>  	/* Wait for shutdown to complete */
>>  	flush_work(&mhi_cntrl->st_worker);
>> --
>> The Qualcomm Innovation Center, Inc. is a member of the Code Aurora 
>> Forum,
>> a Linux Foundation Collaborative Project
>> 

-- 
The Qualcomm Innovation Center, Inc. is a member of the Code Aurora 
Forum,
a Linux Foundation Collaborative Project

  reply	other threads:[~2020-10-13 18:40 UTC|newest]

Thread overview: 30+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-09-19  2:02 [PATCH v1 00/10] Bug fixes and improvements for MHI power operations Bhaumik Bhatt
2020-09-19  2:02 ` [PATCH v1 01/10] bus: mhi: core: Use appropriate names for firmware load functions Bhaumik Bhatt
2020-10-09 15:23   ` Manivannan Sadhasivam
2020-09-19  2:02 ` [PATCH v1 02/10] bus: mhi: core: Move to using high priority workqueue Bhaumik Bhatt
2020-10-09 15:49   ` Manivannan Sadhasivam
2020-09-19  2:02 ` [PATCH v1 03/10] bus: mhi: core: Skip device wake in error or shutdown states Bhaumik Bhatt
2020-10-09 15:54   ` Manivannan Sadhasivam
2020-09-19  2:02 ` [PATCH v1 04/10] bus: mhi: core: Use the IRQF_ONESHOT flag for the BHI interrupt line Bhaumik Bhatt
2020-10-09 15:57   ` Manivannan Sadhasivam
2020-10-09 16:04     ` Jeffrey Hugo
2020-10-12 23:52       ` Bhaumik Bhatt
2020-09-19  2:02 ` [PATCH v1 05/10] bus: mhi: core: Disable IRQs when powering down Bhaumik Bhatt
2020-10-09 16:02   ` Manivannan Sadhasivam
2020-10-13  0:03     ` Bhaumik Bhatt
2020-10-10 23:45   ` Manu Gautam
2020-09-19  2:02 ` [PATCH v1 06/10] bus: mhi: core: Improve shutdown handling after link down detection Bhaumik Bhatt
2020-10-09 16:19   ` Manivannan Sadhasivam
2020-10-13 18:40     ` Bhaumik Bhatt [this message]
2020-10-11  0:06   ` Manu Gautam
2020-09-19  2:02 ` [PATCH v1 07/10] bus: mhi: core: Move to SYS_ERROR regardless of RDDM capability Bhaumik Bhatt
2020-10-09 16:32   ` Manivannan Sadhasivam
2020-10-14  1:06     ` Bhaumik Bhatt
2020-09-19  2:02 ` [PATCH v1 08/10] bus: mhi: core: Move to an error state on any firmware load failure Bhaumik Bhatt
2020-10-09 16:42   ` Manivannan Sadhasivam
2020-10-14  1:37     ` Bhaumik Bhatt
2020-09-19  2:02 ` [PATCH v1 09/10] bus: mhi: core: Move to an error state on mission mode failure Bhaumik Bhatt
2020-10-09 16:44   ` Manivannan Sadhasivam
2020-09-19  2:02 ` [PATCH v1 10/10] bus: mhi: core: Mark device inactive soon after host issues a shutdown Bhaumik Bhatt
2020-10-09 17:18   ` Manivannan Sadhasivam
2020-10-11  0:17   ` Manu Gautam

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=3793225dd0932d0b2a9bb5b3df293fce@codeaurora.org \
    --to=bbhatt@codeaurora.org \
    --cc=hemantk@codeaurora.org \
    --cc=jhugo@codeaurora.org \
    --cc=linux-arm-msm@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=manivannan.sadhasivam@linaro.org \
    --cc=mgautam@codeaurora.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
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).