Re: [PATCH] EHCI/USBMS corrections

[Aleš Nesrsta]

Dne 16.12.2013 15:42, Melki Christian (consultant) napsal(a):
> I think there was more than that.
> Several lines was more than 80 chars long with diff additions. Cutting and 
> pasting them that was wrapped by some client leads to errors.
Yes, you are right.
You can find patch in attachment - I hope in this form it will be OK.
Sorry for inconvenience related to this patch.

>> ... I'd like to try too. Just to get some
>> generic feel for the patch, not that I have any current major problems.
The problem is random and depends most probably on speed of CPU and EHCI
DMA access. On my desktop PC it usually does not happen but it often
happened on my old notebook where I tried replicate the problem "flash
drive timing out, can't boot vmlinuz/initrd (coreboot payload)".

The problem never happens if "ehci" debugging is on (debug="ehci" or
debug="all") - it is because in this case the debugging texts make
sufficient delay between setup of transfer QH/TD(s) and first test of QH
TD overlay state.

BR,
Ales

usb_patch_131212_0.diff
Description: Text Data

signature.asc
Description: OpenPGP digital signature