linux-kernel - Commit 'MIPS: Malta: Use syscon-reboot driver to reboot' in -next and changed reset behavior

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

Date:   Fri, 14 Oct 2016 00:09:01 -0700
From:   Guenter Roeck <linux@...ck-us.net>
To:     Paul Burton <paul.burton@...tec.com>
Cc:     Linux MIPS Mailing List <linux-mips@...ux-mips.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        Ralf Baechle <ralf@...ux-mips.org>
Subject: Commit 'MIPS: Malta: Use syscon-reboot driver to reboot' in -next and
 changed reset behavior

Hi Paul,

with commit 'MIPS: Malta: Use syscon-reboot driver to reboot' in -next, the value written
into the reset register is changed from 0x42 to 0x4d. Is this change on purpose,
or a copy-and-paste error from the SEAD3 changes ?

Reason for asking is that qemu only accepts a value of 0x42, which causes the reset
in qemu to fail. Question is if qemu or the new reset value is wrong.

What values are valid ? Can you shed a light ?

Second question is endianness. Even when changing the value to 0x42, the system still
did not reboot for me. After a while I found out that I needed to add "big-endian;"
to the syscon node when running a big endian image. However, when running a
little endian image, "big-endian" did not work. In that case, I had to use the default,
which is little endian.

Which makes me really wonder how this is expected to work. Does the real hardware accept
any value written into the reset register ?

Thanks,
Guenter