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-next>] [day] [month] [year] [list]
Date:   Thu, 10 Jun 2021 18:58:04 +0530
From:   vhiren@...eaurora.com
To:     bjorn.andersson@...aro.org, adrian.hunter@...el.com,
        robh+dt@...nel.org, ulf.hansson@...aro.org,
        asutoshd@...eaurora.org, stummala@...eaurora.org,
        sayalil@...eaurora.org, rampraka@...eaurora.org,
        vbadigan@...eaurora.org, sboyd@...nel.org,
        georgi.djakov@...aro.org, mka@...omium.org
Cc:     linux-mmc@...r.kernel.org, linux-kernel@...r.kernel.org,
        linux-arm-msm@...r.kernel.org, devicetree@...r.kernel.org,
        agross@...nel.org, linux-mmc-owner@...r.kernel.org,
        nitirawa@...eaurora.org
Subject: Re : Query on UFS Lun detection

Hi Community,

We are using UFS as a boot device, and we encountered an error, 
mentioned below while booting through UFS.


APPS Crash - Kernel panic - not syncing: VFS: Unable to mount root fs on 
unknown-block(8,7)

Here in Kernel command line we are passing "root=sda7", this is our 
Filesystem partition to be mounted.

While debugging the same we observed the partition which is suppose to 
be mounted as Filesystem is not enumerated  itself.
The reason behind not enumerating of partition is, the other Lun got 
detected first instead of one which is having FS partition.

We are using async probe, hence we know the detection will happen in 
async way,


In working case :  it looks like below

sd 0:0:0:0: sda1 sda2 sda3 sda4 …..sda8

sd 0:0:0:1: sdb1 sdb2
Here, sd 0:0:0:0 Lun got detected first and enumerated as "sdaN", we 
have FS partition as "sda7" and so we don’t see above mentioned error.


In non working case : it looks like below

sd 0:0:0:0: sdb1 sdb2 sdb3 sdb4 …..sdb8

sd 0:0:0:1: sda1 sda2
Here, sd 0:0:0:1 Lun got detected first and enumerated as "sdaN" but we 
don’t have FS partition in this Lun. Instead FS partition now became 
"sdb7".


Observation here is, which ever Lun gets detected first it will have 
"sda" enumeration and so on.
However, sdx’s driver scsi/sd.c uses async probe to improve performance, 
meaning sdx disks are created in parallel. Whoever calls 
device_add_disk() first becomes sda, the next become sdb….


Is there any way, where we can fix this in the Lun detection part of 
code?

Thank You,
Regards,
Hiren Gohel

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ