diff options
author | Kirill A. Shutemov <kirill.shutemov@linux.intel.com> | 2014-01-28 17:08:27 +0200 |
---|---|---|
committer | Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> | 2014-02-02 22:09:16 +0530 |
commit | f8b7ee38b3ed4ec2da5cc0529cf0cf82c8589805 (patch) | |
tree | d0bfd4a2449640fff86149e9d456062f68d14664 /hw/arm/pxa2xx.c | |
parent | 1a9978a51a17a26fd07e1b46f31ceeb9772fc191 (diff) | |
download | hqemu-f8b7ee38b3ed4ec2da5cc0529cf0cf82c8589805.zip hqemu-f8b7ee38b3ed4ec2da5cc0529cf0cf82c8589805.tar.gz |
hw/9pfs: fix P9_STATS_GEN handling
Currently we fail getattr request altogether if we can't read
P9_STATS_GEN for some reason. It breaks valid use cases:
E.g let's assume we have non-readable directory with execution bit set
on host and we export it to client over 9p On host we can chdir into
directory, but not open directory on read and list content.
But if client will try to call getattr (as part of chdir(2)) for the
directory it will fail with -EACCES. It happens because we try to open
the directory on read to call ioctl(FS_IOC_GETVERSION), it fails and we
return the error code to client.
It's excessive. The solution is to make P9_STATS_GEN failure non-fatal
for getattr request. Just don't set P9_STATS_GEN flag in result mask on
failure.
Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Diffstat (limited to 'hw/arm/pxa2xx.c')
0 files changed, 0 insertions, 0 deletions