linux-kernel - Re: [PATCH 2/2 v3] initramfs: print error and shell out for unsupported content

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: <533AA1A0.10802@ahsoftware.de>
Date:	Tue, 01 Apr 2014 13:23:12 +0200
From:	Alexander Holler <holler@...oftware.de>
To:	Michal Marek <mmarek@...e.cz>, linux-kernel@...r.kernel.org
CC:	Andrew Morton <akpm@...ux-foundation.org>,
	Levente Kurusa <levex@...ux.com>, stable@...r.kernel.org
Subject: Re: [PATCH 2/2 v3] initramfs: print error and shell out for unsupported
 content

Am 31.03.2014 22:31, schrieb Michal Marek:
> Dne 22.3.2014 19:22, Alexander Holler napsal(a):
>> Am 22.03.2014 13:29, schrieb Alexander Holler:
>>> The initramfs generation is broken for file and directory names which
>>> contain
>>> colons, spaces and other unusual characters. Print an error and don't
>>> try to
>>> continue.
>>
>> (...)
>>> +    # Files and directories with spaces and colons are unsupported.
>>> +    local unsupported=$(find "${srcdir}" -regex '.*\(:\|
>>> \|\n\|\r\|\t\).*')
>>
>> I've just noted that -regex isn't POSIX. I don't know the kernel rules
>> regarding this, and I don't care. But it might be a blocker for this patch.
>
> The bigger problem is that there is no C-style quoting in regexps or
> character classes matches any file with 'n', 'r' or 't' in its name. How
> about
>
>    -name '*[:[:space:]]*'
>
> ?

Hmm, I wasn't aware that find supports such expressions, and I don't 
know how compatible that is. It would do the trick too.

But as I already said, trying to use arbitrary filenames in a Makefile 
doesn't really work. There are still other problems, e.g. filenames with 
a / (directory separator), *, or some of the special make variables like $<.

So the really working solution would be to get rid of that generated 
(and hidden) include. Modifying gen_init_cpio.c to browse and collect 
all filenames and types itself doesn't look like much work (an evening 
should be enough).

The drawback is that such a solution would build the initramfs every 
time make is called (while CONFIG_INITRAMFS_SOURCE points to a 
directory), but I think that time is negligible.
Another drawback is that it wouldn't be a simple patch anymore. Thus it 
wouldn't qualify for any stable tree (and nobody promises such a patch 
would be accepted at all, so it might be a waste of time at all).

Regards,

Alexander Holler

--
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/