linux-kernel - 42c9c06bec x86: ACPI: use ioremap_early() instead of __va()/_

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]

Message-ID: <87r6gft753.fsf@saeurebad.de>
Date:	Fri, 18 Jan 2008 17:45:12 +0100
From:	Johannes Weiner <hannes@...urebad.de>
To:	Ingo Molnar <mingo@...e.hu>
Cc:	Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: 42c9c06bec x86: ACPI: use ioremap_early() instead of __va()/__pa()

Hi Ingo,

with the commit

	42c9c06bec2f48002d5b6573c8700461120070a9
	x86: ACPI: use ioremap_early() instead of __va()/__pa()

you made __acpi_map_table(), which is non-__init, call early_ioremap()
which is __init on 64bit (init_64.c) but non-__init on 32bit
(ioremap_32.c).  This results in section mismatches on 32bit.

The version I have here is v2.6.24-rc8-725-g0f80375.

I had a look into ioremap_32.c and as it seems no __init function should
be called at all in the callpath down from early_ioremap() after the
__init sections are tossed out; but it is decided during runtime with
the check of after_paging_init and so the illegal references still
exist.

What is the solution here?  I have not a real clue of the code, but
perhaps we can make early_ioremap REALLY __init only and use something
different from non-init code?

	Hannes
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/