[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Qemu-devel] [Bug 646427] Re: VirtFS mapped symlinks resolved wrong
From: |
Venkateswararao Jujjuri \(JV\) |
Subject: |
[Qemu-devel] [Bug 646427] Re: VirtFS mapped symlinks resolved wrong |
Date: |
Fri, 24 Sep 2010 17:07:55 -0000 |
Responding to multiple issues raised above.
- Let us track this bug for the security/symlink issue and open a different
defect for the mount issue.
- I tried to reproduce this defect with the QEMU mainline. I don't see the
problem. More description on
the environment (Host OS/Guest OS/ QEMU command line etc) and how to
reproduce the problem will
be helpful.
address@hidden mmnt]# cat file1
omg it is important
address@hidden mmnt]# ln -sf file1 sym1
address@hidden mmnt]# ls -l
total 16
-rw-r--r-- 1 root root 20 2010-09-24 01:24 file1
lrwxrwxrwx 1 root root 6 2010-09-24 01:24 sym1 -> file1
address@hidden mmnt]# logout
- Moshroom:
- There are multiple security models here, if you use 'passthrough' same type
of files are created on the
host/guest. But if you use 'mapped' security model this is how we do it.
This security model is for cloud use case scenarios. Please read more in
this post
http://www.mail-archive.com/address@hidden/msg31452.html
- No need to use lsetattr/lgetattr here because we always create regular
files in this model.
--
VirtFS mapped symlinks resolved wrong
https://bugs.launchpad.net/bugs/646427
You received this bug notification because you are a member of qemu-
devel-ml, which is subscribed to QEMU.
Status in QEMU: New
Bug description:
I tried to map a folder with qemu-kvm qemu-kvm-0.13.0-rc1-3-gc9c2179 (this is
v0.13.0-rc1-16667-gc9c2179). I mounted it first in passthrough mode and saw all
symlinks as expected. Then I used mapped and noticed that the target of a
symlink changed to the actual data inside the linked folder as target instead
of the target filename.
So for example if you have a file abc with the content "omg, this is important
stuff" and a symlink lnkme -> abc then it wiill look inside the kvm with
mounted 9p virtio (mapped) like that:
lnkme -> omg, this is important stuff
Another problem I noticed was that I cannot mount it (mapped or passthrough)
using /etc/rc.local or /etc/fstab, but after login as root using
mount /mnt
or
mount -t 9p -o trans=virtio host_share /mnt
(the same stuff i tried inside /etc/rc.local)
The only output on a failed mount in rc.local/fstab I get is
[ 15.035920] device: '9p-1': device_add
[ 15.038180] 9p: no channels available
[ 15.038937] device: '9p-1': device_unregister
[ 15.049123] device: '9p-1': device_create_release
- [Qemu-devel] [PATCH v2 00/19] Monitor: split HMP and QMP dispatch tables, Luiz Capitulino, 2010/09/30
- [Qemu-devel] [PATCH 01/19] Monitor: Introduce search_dispatch_table(), Luiz Capitulino, 2010/09/30
- [Qemu-devel] [PATCH 02/19] QMP: handle_qmp_command(): Move 'cmd' sanity check, Luiz Capitulino, 2010/09/30
- [Qemu-devel] [PATCH 03/19] QMP: Don't use do_info(), Luiz Capitulino, 2010/09/30
- [Qemu-devel] [PATCH 04/19] Monitor: Drop QMP bits from do_info(), Luiz Capitulino, 2010/09/30
- [Qemu-devel] [PATCH 05/19] Monitor: Drop is_async_return(), Luiz Capitulino, 2010/09/30
- [Qemu-devel] [PATCH 06/19] Monitor: Convert do_info() back to HMP, Luiz Capitulino, 2010/09/30
- [Qemu-devel] [PATCH 07/19] Monitor: Introduce the qmp-commands.hx file, Luiz Capitulino, 2010/09/30
- [Qemu-devel] [PATCH 09/19] QMP: Introduce command dispatch table, Luiz Capitulino, 2010/09/30
- [Qemu-devel] [PATCH 08/19] QMP: Introduce qmp_find_cmd(), Luiz Capitulino, 2010/09/30
- [Qemu-devel] [PATCH 10/19] QMP: Introduce query commands dispatch table, Luiz Capitulino, 2010/09/30