On 02/17/2014 10:15 AM, Hans de Goede wrote:
> Hi Arend,
> 
> I've spend my entire Sunday debugging the bcm43362 on the cubietruck
> not working. I did end up finding several bugs in our sunxi-mmc driver (*),
> and with those fixes it works like a charm with a 3.13 kernel + this
> fix:
> http://git.kernel.org/cgit/linux/kernel/git/linville/wireless-testing.git/commit/drivers/net/wireless/brcm80211?id=79c868e5ada93601c8107a6d08fbb1e0d9348b94
> 
> NB, you should probably send this fix to stable.

I assumed this was not needed because caller uses small chunks so the
loop is always done once. I will look into this.

> But the *exact* same mmc host code does not work in 3.14-rc3, the device
> never responds to the first SD_IO_RW_EXTENDED command (32 64 bytes blocks,
> so 2048 bytes) issued as part of the firmware uploading, the command gets
> submitted to the mmc host controller, but the host controller never generates
> any interrupts (not even a timeout interrupt, which it normally handles in
> hardware too).
> 
> Note that the "register writes" done through 4 data byte SD_IO_RW_EXTENDED
> commands before the firmware upload work fine.

Noted. So it are all F1 operations with 64 byte block size. Need to dive
in the code.

> Unfortunately I cannot bisect this because we need a ton of patches on
> top of upstream to get mmc working for now.
> 
> Thanks & Regards,
> 
> Hans
> 
> 
> *) Which I need to clean up and post, I plan to do this later today,
> first I need to focus on $dayjob, so if you've a cubietruck yourself you
> may want to wait a bit with testing.

Nope. Did not push my supervisor too much to get one :-)

Regards,
Arend

-- 
You received this message because you are subscribed to the Google Groups 
"linux-sunxi" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/groups/opt_out.

Reply via email to