Re: [PATCH 4/6] scsi: ufs: fix LINERESET on hibern8

[Jaegeuk Kim]

Please ignore this patch.
Thanks.

On 09/15, Jaegeuk Kim wrote:
> From: Jaegeuk Kim <jaegeuk@xxxxxxxxxx>
> 
> When testing infinite test to read sysfs entries of UFS, I got a UFS timeout
> with the following kernel message.
> 
> query: dev_cmd_send: seq_no=78082 tag=31, idn=2
> query: ufshcd_wait_for_dev_cmd: dev_cmd request timedout, tag 31
> query: __ufshcd_query_descriptor: opcode 0x01 for idn 2 failed, index 0, err = -11
>  --  hibern8: dme: dme_send: cmd_id=0x17 idn=0
> query: ufshcd_query_descriptor: failed with error -11, retries 3
>  --  hibern8: ufshcd_update_uic_error: LINERESET during hibern8 enter
>  --  hibern8: __ufshcd_uic_hibern8_enter: hibern8 enter failed. ret = -110
> 
> The problem is casued by hibern8 command issued by ufshcd_suspend(), which is
> not aware of query command. If autohibern8 is enabled, we actually don't need
> to issue hibern8 command by suspend.
> 
> Cc: Alim Akhtar <alim.akhtar@xxxxxxxxxxx>
> Cc: Avri Altman <avri.altman@xxxxxxx>
> Signed-off-by: Jaegeuk Kim <jaegeuk@xxxxxxxxxx>
> ---
>  drivers/scsi/ufs/ufshcd.c | 20 ++++++++++++++++++--
>  drivers/scsi/ufs/ufshcd.h |  1 +
>  2 files changed, 19 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/scsi/ufs/ufshcd.c b/drivers/scsi/ufs/ufshcd.c
> index 848e33ec40639..bdc82cc3824aa 100644
> --- a/drivers/scsi/ufs/ufshcd.c
> +++ b/drivers/scsi/ufs/ufshcd.c
> @@ -3079,8 +3079,12 @@ int ufshcd_query_descriptor_retry(struct ufs_hba *hba,
>  	int retries;
>  
>  	for (retries = QUERY_REQ_RETRIES; retries > 0; retries--) {
> -		err = __ufshcd_query_descriptor(hba, opcode, idn, index,
> +		err = -EAGAIN;
> +		down_read(&hba->query_lock);
> +		if (!ufshcd_is_link_hibern8(hba))
> +			err = __ufshcd_query_descriptor(hba, opcode, idn, index,
>  						selector, desc_buf, buf_len);
> +		up_read(&hba->query_lock);
>  		if (!err || err == -EINVAL)
>  			break;
>  	}
> @@ -8263,8 +8267,8 @@ static int ufshcd_suspend(struct ufs_hba *hba, enum ufs_pm_op pm_op)
>  	enum ufs_pm_level pm_lvl;
>  	enum ufs_dev_pwr_mode req_dev_pwr_mode;
>  	enum uic_link_state req_link_state;
> +	bool need_upwrite = false;
>  
> -	hba->pm_op_in_progress = 1;
>  	if (!ufshcd_is_shutdown_pm(pm_op)) {
>  		pm_lvl = ufshcd_is_runtime_pm(pm_op) ?
>  			 hba->rpm_lvl : hba->spm_lvl;
> @@ -8275,6 +8279,15 @@ static int ufshcd_suspend(struct ufs_hba *hba, enum ufs_pm_op pm_op)
>  		req_link_state = UIC_LINK_OFF_STATE;
>  	}
>  
> +	if (ufshcd_is_runtime_pm(pm_op) &&
> +			req_link_state == UIC_LINK_HIBERN8_STATE &&
> +			hba->capabilities & MASK_AUTO_HIBERN8_SUPPORT) {
> +		need_upwrite = true;
> +		if (!down_write_trylock(&hba->query_lock))
> +			return -EBUSY;
> +	}
> +	hba->pm_op_in_progress = 1;
> +
>  	/*
>  	 * If we can't transition into any of the low power modes
>  	 * just gate the clocks.
> @@ -8403,6 +8416,8 @@ static int ufshcd_suspend(struct ufs_hba *hba, enum ufs_pm_op pm_op)
>  	}
>  
>  	hba->pm_op_in_progress = 0;
> +	if (need_upwrite)
> +		up_write(&hba->query_lock);
>  
>  	if (ret)
>  		ufshcd_update_reg_hist(&hba->ufs_stats.suspend_err, (u32)ret);
> @@ -8894,6 +8909,7 @@ int ufshcd_init(struct ufs_hba *hba, void __iomem *mmio_base, unsigned int irq)
>  	mutex_init(&hba->dev_cmd.lock);
>  
>  	init_rwsem(&hba->clk_scaling_lock);
> +	init_rwsem(&hba->query_lock);
>  
>  	ufshcd_init_clk_gating(hba);
>  
> diff --git a/drivers/scsi/ufs/ufshcd.h b/drivers/scsi/ufs/ufshcd.h
> index 363589c0bd370..6f8e05eaf9661 100644
> --- a/drivers/scsi/ufs/ufshcd.h
> +++ b/drivers/scsi/ufs/ufshcd.h
> @@ -754,6 +754,7 @@ struct ufs_hba {
>  	bool is_urgent_bkops_lvl_checked;
>  
>  	struct rw_semaphore clk_scaling_lock;
> +	struct rw_semaphore query_lock;
>  	unsigned char desc_size[QUERY_DESC_IDN_MAX];
>  	atomic_t scsi_block_reqs_cnt;
>  
> -- 
> 2.28.0.618.gf4bc123cb7-goog