linux-kernel - Re: [PATCH 0/2] scsi:donot skip lun if inquiry returns PQ=1 for all hosts

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] [day] [month] [year] [list]

Date:   Wed, 28 Dec 2022 17:35:51 +0800
From:   Wenchao Hao <haowenchao@...wei.com>
To:     Christoph Hellwig <hch@...radead.org>
CC:     "Martin K . Petersen" <martin.petersen@...cle.com>,
        Mike Christie <michael.christie@...cle.com>,
        "James E . J . Bottomley" <jejb@...ux.ibm.com>,
        Lee Duncan <lduncan@...e.com>, Chris Leech <cleech@...hat.com>,
        "open-iscsi@...glegroups.com" <open-iscsi@...glegroups.com>,
        "linux-scsi@...r.kernel.org" <linux-scsi@...r.kernel.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "liuzhiqiang (I)" <liuzhiqiang26@...wei.com>,
        linfeilong <linfeilong@...wei.com>,
        Andrey Melnikov <temnota.am@...il.com>
Subject: Re: [PATCH 0/2] scsi:donot skip lun if inquiry returns PQ=1 for all
 hosts

On 2022/12/23 23:54, Christoph Hellwig wrote:
> On Fri, Dec 16, 2022 at 07:41:26PM +0800, Wenchao Hao wrote:
>> For nothing, I want the iscsi_tcp transport do not skip PQ=1 default
>> as what it did before commit 948e922fc4461 ("scsi: core: map PQ=1,
>> PDT=other values to SCSI_SCAN_TARGET_PRESENT").
> 
> Well, that commit was very much intentional and is now three an a half
> years old, so we've not just going to partially revert it on a
> per-transport basis when it is in no way transport related.
> 
> If you can come up with a good enough rationale we could do the
> sysfs override, but so far the reason mostly seems to be "I want"
> and not anctual explanation of why it is useful.
> .

Sorry, I did not describe in detail.

Here is the reason of my patches.

1. The SPC did not tell PQ=1 means the addressed lun is invalid explicitly.
2. scsi_bus_match() would prevent luns with PQ=1 be handled by any scsi
   drivers, so the only influence if we do not skip luns with PQ=1 is we
   would add an scsi_device and an sg device.

The reason I force a specific behavior for iSCSI:

1. This issue is occurred with iSCSI, it means there are scenarios where
   targets would return PQ=1 for an valid LUN which should not be skipped.
2. The changes for iSCSI could be tested

I did not change other transports' behavior is because I do not know if PQ=1
would be returned, and I do not have the related environment. If other
transports like adaptec raid also needs these changes, they can override
the default value by other patches.

Thanks.