[RFC PATCH] UFS: Remove the timeout while send command

[neil.armstrong at linaro.org]

Hi,

On 04/03/2025 06:15, Abbarapu, Venkatesh wrote:
> [AMD Official Use Only - AMD Internal Distribution Only]
> 
> Hi Neil,
> 
>> -----Original Message-----
>> From: neil.armstrong at linaro.org <neil.armstrong at linaro.org>
>> Sent: Friday, February 28, 2025 8:55 PM
>> To: Abbarapu, Venkatesh <venkatesh.abbarapu at amd.com>; u-boot at lists.denx.de
>> Cc: Simek, Michal <michal.simek at amd.com>; bhupesh.linux at gmail.com; n-
>> francis at ti.com; git (AMD-Xilinx) <git at amd.com>
>> Subject: Re: [RFC PATCH] UFS: Remove the timeout while send command
>>
>> Hi,
>>
>> On 27/02/2025 07:04, Venkatesh Yadav Abbarapu wrote:
>>> Based on the UFS speed mode(HS or PWM) the timeout might vary and can
>>> result timeouts.
>>
>> Can't you bump the QUERY_REQ_TIMEOUT to the PWM timeout delay ?
> 
> Is there any way we can get the PWM timeout delay?
> I think it also depends on the data size we are going to transfer right?
> Any generic way to find the timeouts for PWM or HS modes.

I don't have access to the UFS spec, but perhaps Linux does the calculation here ?

Neil

> 
> Thanks
> Venkatesh
>>
>> Neil
>>
>>>
>>> ufs-versal2-pltfm ufs at f10b0000: Timedout waiting for UTP response
>>> ufs-versal2-pltfm ufs at f10b0000: OCS error from controller = f
>>>
>>> It's a temporary workaround to solve the timeout issue.
>>>
>>> Signed-off-by: Venkatesh Yadav Abbarapu <venkatesh.abbarapu at amd.com>
>>> ---
>>>    drivers/ufs/ufs.c | 9 ---------
>>>    1 file changed, 9 deletions(-)
>>>
>>> diff --git a/drivers/ufs/ufs.c b/drivers/ufs/ufs.c index
>>> f7d8c40c448..c3acbd709e7 100644
>>> --- a/drivers/ufs/ufs.c
>>> +++ b/drivers/ufs/ufs.c
>>> @@ -858,7 +858,6 @@ static int ufshcd_comp_devman_upiu(struct ufs_hba
>>> *hba,
>>>
>>>    static int ufshcd_send_command(struct ufs_hba *hba, unsigned int task_tag)
>>>    {
>>> -   unsigned long start;
>>>      u32 intr_status;
>>>      u32 enabled_intr_status;
>>>
>>> @@ -867,19 +866,11 @@ static int ufshcd_send_command(struct ufs_hba *hba,
>> unsigned int task_tag)
>>>      /* Make sure doorbell reg is updated before reading interrupt status */
>>>      wmb();
>>>
>>> -   start = get_timer(0);
>>>      do {
>>>              intr_status = ufshcd_readl(hba, REG_INTERRUPT_STATUS);
>>>              enabled_intr_status = intr_status & hba->intr_mask;
>>>              ufshcd_writel(hba, intr_status, REG_INTERRUPT_STATUS);
>>>
>>> -           if (get_timer(start) > QUERY_REQ_TIMEOUT) {
>>> -                   dev_err(hba->dev,
>>> -                           "Timedout waiting for UTP response\n");
>>> -
>>> -                   return -ETIMEDOUT;
>>> -           }
>>> -
>>>              if (enabled_intr_status & UFSHCD_ERROR_MASK) {
>>>                      dev_err(hba->dev, "Error in status:%08x\n",
>>>                              enabled_intr_status);
>