Benoit Audouard wrote:
On Thu, May 19, 2005 22:08, matthieu castet said:
Hi,
Benoit Audouard wrote:
Hi,
After svetljo patch,
https://mail.gna.org/public/eagleusb-dev/2005-05/msg00001.html
here is another one from the forum :
http://forum.eagle-usb.org/viewtopic.php?p=20541#20541
I'm not sure how cpu_to_be16 works ?
seems the hexadecimal value is byte-reversed to make it work on ppc,
would
this be compatible with x86 ?
Does cpu_to_be16 exist in kernel-2.4 or only since kernel-2.6.11 ?
yes, but there were a change in usb policy : before kernel 2.6.11 the
usb-layer already done the conversion, now it is to the driver to to it.
ah ;-) you may give a url to the guy currently writing an usb driver (see
link to the forum).
le16_to_cpu don't do anything on littke endian cpu (intel, ...) and swap
high and low byte on big endian cpu (ppc, ...)
ok I'll take cpu_to_le16 for coherency's sake...
In case of doubt I've not yet applied them yet... Matthieu, more ideas ?
The patch is safe to apply.
If you don't aplly it, I will do it tomorow.
ok I can apply it now (+ svetljo's), if you can review it as I'll surely
apply both of them a bit "blindly" ;-) (two pairs of eyes are better than
none).
https://mail.gna.org/public/eagleusb-dev/2005-05/msg00001.html is ok
for http://forum.eagle-usb.org/viewtopic.php?p=20541#20541 we should use
htons insted of cpu_to_be16 (even if it does the same thing.
We also should have used linux/include/linux/if_ether.h constant insted
of defined ours...
@+
Matthieu
