linux-kernel - Re: [PATCH 2/2] scsi: ufs: Stop clearing unit attentions

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]

Message-ID: <f4f81b75-9b0b-7734-ebfa-14bd1b935c54@interlog.com>
Date:   Fri, 1 Oct 2021 13:21:39 -0400
From:   Douglas Gilbert <dgilbert@...erlog.com>
To:     Bart Van Assche <bvanassche@....org>,
        Adrian Hunter <adrian.hunter@...el.com>,
        Jaegeuk Kim <jaegeuk@...nel.org>, linux-kernel@...r.kernel.org,
        linux-scsi@...r.kernel.org, martin.petersen@...cle.com
Cc:     Bart Van Assche <bvanassche@...gle.com>
Subject: Re: [PATCH 2/2] scsi: ufs: Stop clearing unit attentions

On 2021-10-01 12:59 p.m., Bart Van Assche wrote:
> On 9/30/21 11:52 PM, Adrian Hunter wrote:
>> Finally, there is another thing to change.  The reason
>> ufshcd_suspend_prepare() does a runtime resume of sdev_rpmb is because the
>> UAC clear would wait for an async runtime resume, which will never happen
>> during system suspend because the PM workqueue gets frozen.  So with the
>> removal of UAC clear, ufshcd_suspend_prepare() and ufshcd_resume_complete()
>> should be updated also, to leave rpmb alone.

Somewhat related ...

Since there was some confusion among the members of T10 of what precisely
the RPM bit meant, in SPC-6 revision (draft), a new "HOT PLUGGABLE" two
bit field was introduced into the standard INQUIRY response:

                 Table 151 — HOT PLUGGABLE field

Code   Description
00b    No information is provided regarding whether SCSI target device is hot
        pluggable.
01b    The SCSI target device is designed to be removed from a SCSI domain as
        a single object (i.e., concurrent removal of the SCSI target ports,
        logical units, and all other objects contained in that SCSI target
        device (see SAM-6)) while that SCSI domain continues to operate for
        all other SCSI target devices, if any, in that SCSI domain.
10b    The SCSI target device is not designed to be removed from a SCSI
        domain while that SCSI domain continues to operate.
11b    Reserved

That field is bits 5 and 4 of byte 1 of the response.

Perhaps we should be adding provision for this new field.

Doug Gilbert

> 
> Is the following change what you have in mind?
> 
> 
> diff --git a/drivers/scsi/ufs/ufshcd.c b/drivers/scsi/ufs/ufshcd.c
> index 0a28cc4c09d8..0743f54e55f9 100644
> --- a/drivers/scsi/ufs/ufshcd.c
> +++ b/drivers/scsi/ufs/ufshcd.c
> @@ -9648,10 +9648,6 @@ void ufshcd_resume_complete(struct device *dev)
>           ufshcd_rpm_put(hba);
>           hba->complete_put = false;
>       }
> -    if (hba->rpmb_complete_put) {
> -        ufshcd_rpmb_rpm_put(hba);
> -        hba->rpmb_complete_put = false;
> -    }
>   }
>   EXPORT_SYMBOL_GPL(ufshcd_resume_complete);
> 
> @@ -9674,10 +9670,6 @@ int ufshcd_suspend_prepare(struct device *dev)
>           }
>           hba->complete_put = true;
>       }
> -    if (hba->sdev_rpmb) {
> -        ufshcd_rpmb_rpm_get_sync(hba);
> -        hba->rpmb_complete_put = true;
> -    }
>       return 0;
>   }
>   EXPORT_SYMBOL_GPL(ufshcd_suspend_prepare);
> diff --git a/drivers/scsi/ufs/ufshcd.h b/drivers/scsi/ufs/ufshcd.h
> index 86b615023ecb..5ecfcd8cae0a 100644
> --- a/drivers/scsi/ufs/ufshcd.h
> +++ b/drivers/scsi/ufs/ufshcd.h
> @@ -921,7 +921,6 @@ struct ufs_hba {
>   #endif
>       u32 luns_avail;
>       bool complete_put;
> -    bool rpmb_complete_put;
>   };
> 
>   /* Returns true if clocks can be gated. Otherwise false */
> 
> 
> 
> Thanks,
> 
> Bart.