Re: [PATCH v2 0/5] Fix ELF / FDPIC ELF core dumping, and use mmap_sem properly in there - Russell King

From: Russell King - ARM Linux admin <linux@armlinux.org.uk>
To: Jann Horn <jannh@google.com>, Nicolas Pitre <nico@fluxnic.net>
Cc: Rich Felker <dalias@libc.org>,
	linux-c6x-dev@linux-c6x.org,
	Yoshinori Sato <ysato@users.sourceforge.jp>,
	linux-sh@vger.kernel.org, linux-kernel@vger.kernel.org,
	Oleg Nesterov <oleg@redhat.com>,
	linux-mm@kvack.org, Alexander Viro <viro@zeniv.linux.org.uk>,
	Mark Salter <msalter@redhat.com>,
	linux-fsdevel@vger.kernel.org,
	Andrew Morton <akpm@linux-foundation.org>,
	Aurelien Jacquiot <jacquiot.aurelien@gmail.com>,
	Linus Torvalds <torvalds@linux-foundation.org>,
	Christoph Hellwig <hch@lst.de>,
	linux-arm-kernel@lists.infradead.org,
	"Eric W . Biederman" <ebiederm@xmission.com>
Subject: Re: [PATCH v2 0/5] Fix ELF / FDPIC ELF core dumping, and use mmap_sem properly in there
Date: Wed, 29 Apr 2020 21:56:20 +0000	[thread overview]
Message-ID: <20200429215620.GM1551@shell.armlinux.org.uk> (raw)
In-Reply-To: <20200429214954.44866-1-jannh@google.com>

On Wed, Apr 29, 2020 at 11:49:49PM +0200, Jann Horn wrote:
> At the moment, we have that rather ugly mmget_still_valid() helper to
> work around <https://crbug.com/project-zero/1790>: ELF core dumping
> doesn't take the mmap_sem while traversing the task's VMAs, and if
> anything (like userfaultfd) then remotely messes with the VMA tree,
> fireworks ensue. So at the moment we use mmget_still_valid() to bail
> out in any writers that might be operating on a remote mm's VMAs.
> 
> With this series, I'm trying to get rid of the need for that as
> cleanly as possible.
> In particular, I want to avoid holding the mmap_sem across unbounded
> sleeps.
> 
> 
> Patches 1, 2 and 3 are relatively unrelated cleanups in the core
> dumping code.
> 
> Patches 4 and 5 implement the main change: Instead of repeatedly
> accessing the VMA list with sleeps in between, we snapshot it at the
> start with proper locking, and then later we just use our copy of
> the VMA list. This ensures that the kernel won't crash, that VMA
> metadata in the coredump is consistent even in the presence of
> concurrent modifications, and that any virtual addresses that aren't
> being concurrently modified have their contents show up in the core
> dump properly.
> 
> The disadvantage of this approach is that we need a bit more memory
> during core dumping for storing metadata about all VMAs.
> 
> After this series has landed, we should be able to rip out
> mmget_still_valid().
> 
> 
> Testing done so far:
> 
>  - Creating a simple core dump on X86-64 still works.
>  - The created coredump on X86-64 opens in GDB, and both the stack and the
>    exectutable look vaguely plausible.
>  - 32-bit ARM compiles with FDPIC support, both with MMU and !MMU config.
> 
> I'm CCing some folks from the architectures that use FDPIC in case
> anyone wants to give this a spin.

I've never had any reason to use FDPIC, and I don't have any binaries
that would use it.  Nicolas Pitre added ARM support, so I guess he
would be the one to talk to about it.  (Added Nicolas.)

-- 
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
FTTC broadband for 0.8mile line in suburbia: sync at 10.2Mbps down 587kbps up