linux-kernel - Re: [PATCH]: MTRR: fix 32-bit ioctls on x64

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: <5c8016cf0701160959l1489b56t980779f1f6e49848@mail.gmail.com>
Date:	Tue, 16 Jan 2007 17:59:15 +0000
From:	"Giuliano Procida" <giuliano.procida@...glemail.com>
To:	"Mikael Pettersson" <mikpe@...uu.se>, rgooch@...f.csiro.au
Cc:	linux-kernel@...r.kernel.org
Subject: Re: [PATCH]: MTRR: fix 32-bit ioctls on x64_32

Hi.

On 16/01/07, Mikael Pettersson <mikpe@...uu.se> wrote:
> On Tue, 16 Jan 2007 08:14:30 +0000, Giuliano Procida wrote:
> These #ifdefs are too ugly.

Agreed that the #ifdefs are rather ugly, but they were the smallest change.
Whoever wrote the original compat changes was relying on the IOC
constants being different for 32- and 64-bit userspace. This allowed the
lazy reuse of the whole ioctl function rather than having to write a complete
replacement compat_ioctl for fops.

> Since you apparently just add aliases for the case labels,
> and do no actual code changes, why not
> 1. make the new cases unconditional, or

The constants are only visible under CONFIG_COMPAT. I think they
should stay that way.

> 2. invoke a translation function before the switch which
>    maps the MTRRCIOC32_ constants to what the kernel uses

Other things I considered:

3. write a wrapper compat function that calls the original, make the
original pure 64-bit
4. macroise the case labels away somehow
5. update cmd in place (cannot do this as it re-used in the third switch)
6. add real_cmd and switch on that instead (your 2.),
   requires yet another switch and #ifdef.

It might be nicer to decode the IOC constants and use action, size and R/W flags
to control all the switches. Not sure myself.

Giuliano.
-
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/