From: Konstantin Khlebnikov <khlebnikov@openvz.org>
To: Matt Helsley <matthltc@us.ibm.com>
Cc: "linux-mm@kvack.org" <linux-mm@kvack.org>,
Andrew Morton <akpm@linux-foundation.org>,
"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
Oleg Nesterov <oleg@redhat.com>, Eric Paris <eparis@redhat.com>,
"linux-security-module@vger.kernel.org"
<linux-security-module@vger.kernel.org>,
"oprofile-list@lists.sf.net" <oprofile-list@lists.sf.net>,
Linus Torvalds <torvalds@linux-foundation.org>,
Al Viro <viro@zeniv.linux.org.uk>
Subject: Re: [PATCH 6/7] mm: kill vma flag VM_EXECUTABLE
Date: Tue, 03 Apr 2012 09:06:12 +0400 [thread overview]
Message-ID: <4F7A8544.2020603@openvz.org> (raw)
In-Reply-To: <20120402231837.GC32299@count0.beaverton.ibm.com>
Matt Helsley wrote:
> On Sat, Mar 31, 2012 at 01:29:29PM +0400, Konstantin Khlebnikov wrote:
>> Currently the kernel sets mm->exe_file during sys_execve() and then tracks
>> number of vmas with VM_EXECUTABLE flag in mm->num_exe_file_vmas, as soon as
>> this counter drops to zero kernel resets mm->exe_file to NULL. Plus it resets
>> mm->exe_file at last mmput() when mm->mm_users drops to zero.
>>
>> Vma with VM_EXECUTABLE flag appears after mapping file with flag MAP_EXECUTABLE,
>> such vmas can appears only at sys_execve() or after vma splitting, because
>> sys_mmap ignores this flag. Usually binfmt module sets mm->exe_file and mmaps
>> some executable vmas with this file, they hold mm->exe_file while task is running.
>>
>> comment from v2.6.25-6245-g925d1c4 ("procfs task exe symlink"),
>> where all this stuff was introduced:
>>
>>> The kernel implements readlink of /proc/pid/exe by getting the file from
>>> the first executable VMA. Then the path to the file is reconstructed and
>>> reported as the result.
>>>
>>> Because of the VMA walk the code is slightly different on nommu systems.
>>> This patch avoids separate /proc/pid/exe code on nommu systems. Instead of
>>> walking the VMAs to find the first executable file-backed VMA we store a
>>> reference to the exec'd file in the mm_struct.
>>>
>>> That reference would prevent the filesystem holding the executable file
>>> from being unmounted even after unmapping the VMAs. So we track the number
>>> of VM_EXECUTABLE VMAs and drop the new reference when the last one is
>>> unmapped. This avoids pinning the mounted filesystem.
>>
>> So, this logic is hooked into every file mmap/unmmap and vma split/merge just to
>> fix some hypothetical pinning fs from umounting by mm which already unmapped all
>> its executable files, but still alive. Does anyone know any real world example?
>> mm can be borrowed by swapoff or some get_task_mm() user, but it's not a big problem.
>>
>> Thus, we can remove all this stuff together with VM_EXECUTABLE flag and
>> keep mm->exe_file alive till final mmput().
>>
>> After that we can access current->mm->exe_file without any locks
>> (after checking current->mm and mm->exe_file for NULL)
>>
>> Some code around security and oprofile still uses VM_EXECUTABLE for retrieving
>> task's executable file, after this patch they will use mm->exe_file directly.
>> In tomoyo and audit mm is always current->mm, oprofile uses get_task_mm().
>
> Perhaps I'm missing something but it seems like you ought to split
> this into two patches. The first could fix up the cell, tile, etc. arch
> code to use the exe_file reference rather than walk the VMAs. Then the
> second patch could remove the unusual logic used to allow userspace to unpin
> the mount and we could continue to discuss that separately. It would
> also make the git log somewhat cleaner I think...
Ok, I'll resend this patch as independent patch-set,
anyway I need to return mm->mmap_sem locking back.
next prev parent reply other threads:[~2012-04-03 5:06 UTC|newest]
Thread overview: 52+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-03-31 9:25 [PATCH 0/7] mm: vma->vm_flags diet Konstantin Khlebnikov
2012-03-31 9:29 ` [PATCH 1/7] mm, x86, PAT: rework linear pfn-mmap tracking Konstantin Khlebnikov
2012-03-31 17:09 ` [PATCH 1/7 v2] " Konstantin Khlebnikov
2012-04-03 0:46 ` [x86 PAT PATCH 0/2] x86 PAT vm_flag code refactoring Suresh Siddha
2012-04-03 0:46 ` [x86 PAT PATCH 1/2] x86, pat: remove the dependency on 'vm_pgoff' in track/untrack pfn vma routines Suresh Siddha
2012-04-03 5:37 ` Konstantin Khlebnikov
2012-04-03 23:31 ` Suresh Siddha
2012-04-04 4:43 ` Konstantin Khlebnikov
2012-04-05 11:56 ` Konstantin Khlebnikov
2012-04-06 0:01 ` [v3 VM_PAT PATCH 0/3] x86 VM_PAT series Suresh Siddha
2012-04-06 0:01 ` [v3 VM_PAT PATCH 1/3] x86, pat: remove the dependency on 'vm_pgoff' in track/untrack pfn vma routines Suresh Siddha
2012-04-06 0:01 ` [v3 VM_PAT PATCH 2/3] x86, pat: separate the pfn attribute tracking for remap_pfn_range and vm_insert_pfn Suresh Siddha
2012-04-06 0:01 ` [v3 VM_PAT PATCH 3/3] mm, x86, PAT: rework linear pfn-mmap tracking Suresh Siddha
2012-04-03 0:46 ` [x86 PAT PATCH 2/2] " Suresh Siddha
2012-04-03 5:48 ` Konstantin Khlebnikov
2012-04-03 5:55 ` Konstantin Khlebnikov
2012-04-03 6:03 ` [x86 PAT PATCH 0/2] x86 PAT vm_flag code refactoring Konstantin Khlebnikov
2012-04-03 23:14 ` Suresh Siddha
2012-04-04 4:40 ` Konstantin Khlebnikov
2012-03-31 9:29 ` [PATCH 2/7] mm: introduce vma flag VM_ARCH_1 Konstantin Khlebnikov
2012-03-31 22:25 ` Benjamin Herrenschmidt
2012-03-31 9:29 ` [PATCH 3/7] mm: kill vma flag VM_CAN_NONLINEAR Konstantin Khlebnikov
2012-03-31 17:01 ` Linus Torvalds
2012-03-31 9:29 ` [PATCH 4/7] mm: kill vma flag VM_INSERTPAGE Konstantin Khlebnikov
2012-03-31 9:29 ` [PATCH 5/7] mm, drm/udl: fixup vma flags on mmap Konstantin Khlebnikov
2012-03-31 9:29 ` [PATCH 6/7] mm: kill vma flag VM_EXECUTABLE Konstantin Khlebnikov
2012-03-31 20:13 ` Oleg Nesterov
2012-03-31 20:39 ` Cyrill Gorcunov
2012-04-02 9:46 ` Konstantin Khlebnikov
2012-04-02 9:54 ` Cyrill Gorcunov
2012-04-02 10:13 ` Konstantin Khlebnikov
2012-04-02 14:48 ` Oleg Nesterov
2012-04-02 16:02 ` Cyrill Gorcunov
2012-04-02 16:19 ` Konstantin Khlebnikov
2012-04-02 16:27 ` Cyrill Gorcunov
2012-04-02 17:14 ` Konstantin Khlebnikov
2012-04-02 18:05 ` Cyrill Gorcunov
2012-04-02 23:04 ` Matt Helsley
2012-04-03 5:10 ` Konstantin Khlebnikov
2012-04-03 18:16 ` Matt Helsley
2012-04-03 19:32 ` Cyrill Gorcunov
2012-04-05 20:29 ` Matt Helsley
2012-04-05 20:53 ` Cyrill Gorcunov
2012-04-05 21:04 ` Konstantin Khlebnikov
2012-04-05 21:44 ` Matt Helsley
2012-04-05 21:55 ` Linus Torvalds
2012-04-06 4:36 ` Konstantin Khlebnikov
2012-04-02 23:18 ` Matt Helsley
2012-04-03 5:06 ` Konstantin Khlebnikov [this message]
2012-04-06 22:48 ` Andrew Morton
2012-03-31 9:29 ` [PATCH 7/7] mm: move madvise vma flags to the end Konstantin Khlebnikov
2012-03-31 14:06 ` [PATCH 0/7] mm: vma->vm_flags diet Andi Kleen
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=4F7A8544.2020603@openvz.org \
--to=khlebnikov@openvz.org \
--cc=akpm@linux-foundation.org \
--cc=eparis@redhat.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=linux-security-module@vger.kernel.org \
--cc=matthltc@us.ibm.com \
--cc=oleg@redhat.com \
--cc=oprofile-list@lists.sf.net \
--cc=torvalds@linux-foundation.org \
--cc=viro@zeniv.linux.org.uk \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).