stable-tracking test image
werner at openmoko.org
Sun Nov 16 06:36:58 CET 2008
"Marco Trevisan (Trevi?o)" wrote:
> I've got them all the times I've tried it
Can you actually associate to an AP or does it not even get that far ?
I wonder if there's some signal integrity or timing issue. I've
attached a patch for a kernel where WLAN uses Linux-SDIO. This patch
lets you adjust the maximum clock frequency of the SD interface.
Boot with s3cmci.f_max=10000000
Then the kernel should say
s3c2440-sdi s3c2440-sdi: f_max lowered from 50000000 to 10000000 Hz
and the interface frequency should be limited accordingly.
Without restriction, the SDIO interface runs at 25MHz. The minimum
limit is 195312Hz. If you try to set f_max below this, the option is
ignored (and no "f_max lowered" line is printed).
For convenience, I made a snapshot of the kernel:
This kernel lags behind stable-tracking, so please don't use it for
anything but this WLAN test. The rootfs is the latest from
To set s3cmci.f_max in the kernel parameter line, if using u-boot, you
can use the "envi" command from the rootfs below. It reads the u-boot
environment, presents it for editing in vi, and then writes it back
The rootfs gives you a shell on the serial console very early. SSH
comes up a bit later. iwconfig works as usual. For DHCP, there is
dhclient. Invocation: dhclient eth0
Note that the resolver is hard-configured to OpenDNS, that /etc/hosts
is my local hosts file, and that SSH's authorized keys give me access
to anything running that rootfs ;-)
Alternatively, you can of course also change the command line in
u-boot directly (or use Qi). Then you don't need my rootfs.
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 845 bytes
Desc: not available
Url : http://lists.openmoko.org/pipermail/openmoko-kernel/attachments/20081116/c32b31a7/attachment.patch
More information about the openmoko-kernel