Hi
On Mon, 30 Jul 2007 10:47:26 -0400
"John W. Linville" <linville@xxxxxxxxxxxxx> wrote:
> Have you tested the both the old and the new wireless-tools on
> a kernel with your patch applied? Both 32- and 64-bit versions?
> Do they all work equally well?
Sorry for my late reply.
OK, I tested the following this week:
- kernel 2.6.23-rc2 (geoff's PS3 tree)
- the patch applied(*1) and unpatched
(*1) another patch also applied. See later section
- wirlesstools version 28 and 29 pre22
- 64bit and 32bit compiled
- iwpriv and iwconfig
Test
- boot each kernel of two versions (patched and unpatched) on PS3
- associate zd1211rw to an access point with wpa_supplicant (64bit
compiled)
- run 'iwconfig eth1' and 'iwpriv eth1'
four versions
wirelesstools v28/32bit, v28/64bit
wirelesstools v29/32bit, v29/64bit
- compare the outputs
As I expected, 32bit iwconfig and iwpriv did not show proper result on
unpatched kernel, like:
> [root@localhost wireless_tools.28.32]# ./iwpriv eth1
> eth1 no private ioctls.
>
> [root@localhost wireless_tools.28.32]# ./iwconfig eth1
> Warning: Driver for device eth1 has been compiled with version 22
> of Wireless Extension, while this program supports up to version 20.
> Some things may be broken...
>
> eth1 IEEE 802.11b/g ESSID:off/any Nickname:"zd1211"
> Mode:Managed Frequency:2.462 GHz Access Point: 00:16:01:3A:F5:FD
> Bit Rate:24 Mb/s
> Encryption key:<too big>
> [root@localhost wireless_tools.29.32]# ./iwpriv eth1
> eth1 no private ioctls.
>
> [root@localhost wireless_tools.29.32]# ./iwconfig eth1
> eth1 IEEE 802.11b/g ESSID:off/any Nickname:"zd1211"
> Mode:Managed Frequency:2.462 GHz Access Point: 00:16:01:3A:F5:FD
> Bit Rate=24 Mb/s
> Encryption key:<too big>
On the patched kernel, 32bit iwconfig and iwpriv showed appropriate
outputs and were same as 64bit's output:
> [root@localhost wireless_tools.28.32]# ./iwpriv eth1
> eth1 Available private ioctls :
> set_regdomain (8BE0) : set 1 byte & get 0
> get_regdomain (8BE1) : set 0 & get 1 byte
>
> [root@localhost wireless_tools.28.32]# ./iwconfig eth1
> Warning: Driver for device eth1 has been compiled with version 22
> of Wireless Extension, while this program supports up to version 20.
> Some things may be broken...
>
> eth1 IEEE 802.11b/g ESSID:"planexuser" Nickname:"zd1211"
> Mode:Managed Frequency:2.462 GHz Access Point: 00:16:01:3A:F5:FD
> Bit Rate:24 Mb/s
> Encryption key:9741-5654-B44F-7C71-2B3A-F918-1908-431F Security mode:open
> Link Quality=81/100 Signal level=45/100
> Rx invalid nwid:0 Rx invalid crypt:0 Rx invalid frag:0
> Tx excessive retries:0 Invalid misc:0 Missed beacon:0
> [root@localhost wireless_tools.28.64]# ./iwpriv eth1
> eth1 Available private ioctls :
> set_regdomain (8BE0) : set 1 byte & get 0
> get_regdomain (8BE1) : set 0 & get 1 byte
>
> [root@localhost wireless_tools.28.64]# ./iwconfig eth1
> Warning: Driver for device eth1 has been compiled with version 22
> of Wireless Extension, while this program supports up to version 20.
> Some things may be broken...
>
> eth1 IEEE 802.11b/g ESSID:"planexuser" Nickname:"zd1211"
> Mode:Managed Frequency:2.462 GHz Access Point: 00:16:01:3A:F5:FD
> Bit Rate=24 Mb/s
> Encryption key:9741-5654-B44F-7C71-2B3A-F918-1908-431F Security mode:open
> Link Quality=82/100 Signal level=45/100
> Rx invalid nwid:0 Rx invalid crypt:0 Rx invalid frag:0
> Tx excessive retries:0 Invalid misc:0 Missed beacon:0
So I think SIOCGIWPRIV and SOICGIWSTAT issues were fixed without
regressions.
While this testing, I've found another problem for iocompat32.
fs/compat_ioctl.c:do_wireless_ioctl() copies the contents of the
argument of the 32bit ioctls into allocated area. But it does not copy
back the contents on the return from 64bit ioctl calls. So the caller
would get the unmodified argument even if the ioctl handler modified it.
For example, SIOCGIWENCODE could not return proper key length to the
caller, then iwconfig would show the strange result:
> Encryption key:<too big>
The applied patch fixes the issue.
As I mentioned early in this mail, I did testing showed above with this
patch and found no obvious issue, but I did not have time/test-suite to
do comprehensive testing. Please someone do review this new patch and
test more.
regards
--
Masakazu MOKUNO
Attachment:
fix_do_wirless_ioctl.diff
Description: Binary data
