lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date:   Fri, 01 Nov 2019 09:18:06 +0800
From:   cang@...eaurora.org
To:     Mark Salyzyn <salyzyn@...roid.com>
Cc:     asutoshd@...eaurora.org, nguyenb@...eaurora.org,
        rnayak@...eaurora.org, linux-scsi@...r.kernel.org,
        kernel-team@...roid.com, saravanak@...gle.com, salyzyn@...gle.com,
        Alim Akhtar <alim.akhtar@...sung.com>,
        Avri Altman <avri.altman@....com>,
        Pedro Sousa <pedrom.sousa@...opsys.com>,
        "James E.J. Bottomley" <jejb@...ux.ibm.com>,
        "Martin K. Petersen" <martin.petersen@...cle.com>,
        Stanley Chu <stanley.chu@...iatek.com>,
        Bean Huo <beanhuo@...ron.com>,
        Tomas Winkler <tomas.winkler@...el.com>,
        Subhash Jadavani <subhashj@...eaurora.org>,
        Arnd Bergmann <arnd@...db.de>,
        Bjorn Andersson <bjorn.andersson@...aro.org>,
        open list <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v1 1/2] scsi: ufs: Introduce a vops for resetting host
 controller

On 2019-10-31 22:44, Mark Salyzyn wrote:
> On 10/22/19 9:13 PM, Can Guo wrote:
>> Some UFS host controllers need their specific implementations of 
>> resetting
>> to get them into a good state. Provide a new vops to allow the 
>> platform
>> driver to implement this own reset operation.
>> 
>> Signed-off-by: Can Guo <cang@...eaurora.org>
>> ---
>>   drivers/scsi/ufs/ufshcd.c | 16 ++++++++++++++++
>>   drivers/scsi/ufs/ufshcd.h | 10 ++++++++++
>>   2 files changed, 26 insertions(+)
>> 
>> diff --git a/drivers/scsi/ufs/ufshcd.c b/drivers/scsi/ufs/ufshcd.c
>> index c28c144..161e3c4 100644
>> --- a/drivers/scsi/ufs/ufshcd.c
>> +++ b/drivers/scsi/ufs/ufshcd.c
>> @@ -3859,6 +3859,14 @@ static int ufshcd_link_recovery(struct ufs_hba 
>> *hba)
>>   	ufshcd_set_eh_in_progress(hba);
>>   	spin_unlock_irqrestore(hba->host->host_lock, flags);
>>   +	ret = ufshcd_vops_full_reset(hba);
>> +	if (ret)
>> +		dev_warn(hba->dev, "%s: full reset returned %d\n",
>> +				  __func__, ret);
>> +
>> +	/* Reset the attached device */
>> +	ufshcd_vops_device_reset(hba);
>> +
>>   	ret = ufshcd_host_reset_and_restore(hba);
>>     	spin_lock_irqsave(hba->host->host_lock, flags);
> 
> In all your cases, especially after this adjustment,
> ufshcd_vops_full_reset is called blindly (+error checking message)
> before ufshcd_vops_device_reset. What about dropping the .full_reset
> (should really have been called .hw_reset or .host_reset) addition to
> the vops, just adding ufshcd_vops_device_reset call here before
> ufshcd_host_reset_and_restore, and in the driver folding the
> ufshcd_vops_full_reset code into the .device_reset handler?
> 
> Would that be workable? It would be simpler if so.
> 
> I can see a desire for the heads up
> (ufshcd_vops_full_reset+)ufshcd_vops_device_reset calls before
> ufshcd_host_reset_and_restore because that function will spin 10
> seconds waiting for a response from a standardized register, that
> itself could be hardware locked up requiring product specific reset
> procedures. But if that is the case, then what about all the other
> calls to ufshcd_host_reset_and_restore in this file that are not
> provided the heads up? My guess is that the host device only
> demonstrated issues in the ufshcd_link_recovery handling path? Are you
> sure this is the only path that tickles the controller into a hardware
> lockup state?
> 
> Sincerely -- Mark Salyzyn

Hi Mark Salyzyn,

Folding the "full_reset" vops inito "device_reset" vops is one choice 
for now. Shall do that.
Your guess is correct. the head up is needed in ufshcd_link_recovery() 
path because
link is already in bad state when we are here, expeically after hibern8 
exit fails.
So we need a full reset to PHY and host controller here before 
host_reset_and_restore.
But other calls to host_reset_and_restore are under good conditions.

Regards,
Can Guo.

Powered by blists - more mailing lists