qemu-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Qemu-devel] [PATCH] virtio-9p: Fix a memory leak


From: Venkateswararao Jujjuri
Subject: Re: [Qemu-devel] [PATCH] virtio-9p: Fix a memory leak
Date: Sun, 08 May 2011 10:40:56 -0700
User-agent: Mozilla/5.0 (X11; U; Linux i686 (x86_64); en-US; rv:1.9.2.17) Gecko/20110414 Thunderbird/3.1.10

On 05/07/2011 01:27 PM, Pedro Scarapicchia Junior wrote:
Hi Stefan,

Thanks for the comment.

I believe that it is possible to release the memory at v9fs_walk. However v9fs_walk_complete() is called from two another functions: v9fs_walk_post_newfid_lstat()   and  v9fs_walk_post_oldfid_lstat(). Placing qemu_free at end of v9fs_walk_complete() solve memory leak in both cases.

Venkateswararao, what's your opinion?
I agree with Pedro. Given the state machine model the v9fs_walk() returns from multiple places.
We can take this patch for now. A new patch set is getting brewed converting this whole method with
coroutines + glib thread pools. That will replace all these *post* functions with more of a
readable format. Again thanks for the patch.

- JV

Best regards,

Pedro 

On Sat, May 7, 2011 at 5:56 AM, Stefan Weil <address@hidden> wrote:
Am 07.05.2011 10:34, schrieb Stefan Hajnoczi:

On Fri, May 6, 2011 at 8:05 PM, Pedro Scarapicchia Junior
<address@hidden> wrote:
At v9fs_walk_complete(), the memory allocated at v9fs_walk() is not being
released leading system to crash due out of memory.

This patch releases structure V9fsWalkState after v9fs_walk is complete.

Signed-off-by: Pedro Scarapicchia Junior <address@hidden>
---
 hw/9pfs/virtio-9p.c |    2 ++
 1 files changed, 2 insertions(+), 0 deletions(-)

Thanks for this patch. I suggest CCing Venkateswararao Jujjuri (JV)
<address@hidden>, the virtio-9p maintainer (see MAINTAINERS
file), on future patches so he can pick them up quickly.

Stefan

Releasing the memory in v9fs_walk() were it was allocated
would be cleaner and easier to review. Is this not possible?

Stefan W.





reply via email to

[Prev in Thread] Current Thread [Next in Thread]