-----Original Message----- From: l.pinguin@gmail.com [mailto:l.pinguin@gmail.com] On Behalf Of Remy Bohmer Sent: Thursday, July 09, 2009 1:00 PM To: Prafulla Wadaskar; Michael Trimarchi Cc: u-boot@lists.denx.de; Prabhanjan Sarnaik; Ronen Shitrit; Ashish Karkare Subject: Re: [PATCH 5/6] usb: add Marvell Kirkwood ehci host controller driver
Hello Prafulla,
As stated in other email thread too. Now all dependencies for Sheevaplug board support are in accepted state except this patch http://lists.denx.de/pipermail/u-boot/2009-June/055112.html
I know that but I cannot test it because I do not have the hardware, so I depend on others here. (it might be even you). I already mentioned that I would prefer the cleanup patch from Michael Trimarchi, if it is correct.
At least there is 1 significant difference between your patch and the patch of Michael
Your patch contains this:
- typeReq = req->request << 8 | req->requesttype;
typeReq = req->request | req->requesttype << 8;
switch (le16_to_cpu(typeReq)) {
and Michael patch contains this:
- typeReq = req->request << 8 | req->requesttype;
- typeReq = req->request | req->requesttype << 8;
- switch (le16_to_cpu(typeReq)) {
- switch (typeReq) {
Dear Remy These two lines makes no difference for me since le16_to_cpu macro doesn't change anything for arm little-endian machines (i.e. Sheevaplug) Well.. for big endian machines someone need to check since it swaps bytes.
I think we should not remove this macro since it provides arch specific abstraction.
Regards.. Prafulla . .
So, at least one of them is wrong. This difference is reason enough for me to not pull either patch in until this is clear.
Kind Regards,
Remy