On Wed, 8 Apr 2009, Sascha Hauer wrote: > Hi Guennadi, > > > Caution: CC-ing the ARM kernel ML, which is moderated for non-subscribers, > > beware when replying. > > > > Ths patch is RFC because it actually doesn't work with the current next, > > but it does work if I revert the clkdev support for i.MX31... > > Note that the driver was only working due to a bug *before* clkdev > support which gets fixed by accident with this patch. The code without > clkdev did a > > clk_disable(&usb_pll_clk); > > during initialization. This was without effect because the use count of > this clock was 0 and thus the actual disable function was not called. > The result was that the clock was still running from the bootloader > (it's reset default). > > Try to turn this clock on even for the ULPI case, this at least helped > me with the host driver. IIRC the Freescale code turned this clock on > aswell, but disabled it for the ULPI case after initialization. Yes, indeed! See, not knowing that the clock implementation before your patch had that specific bug I had (well, almost) no chance to find out why it suddenly broke after your patch:-) Indeed, it works now. So, we shall get it ready for 2.6.31. Expect an updated patch soon. Thanks Guennadi --- Guennadi Liakhovetski, Ph.D. DENX Software Engineering GmbH, MD: Wolfgang Denk & Detlev Zundel HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany Phone: +49-8142-66989-0 Fax: +49-8142-66989-80 Email: office@xxxxxxx -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
