I've got a patch that adds OHCI support for a new SoC. I'd say it is all good except that it requires a work-around for accessing OHCI registers. The code needs to hijack readl() so that it can perform the read twice.
What it does is define OHCI_readl() and then replace uses of readl() with that macro. For all other implementations the OHCI_readl() call translates to readl (). I did it this way so that there would be no doubt that OHCI_readl () might not be the standard macro.
The questions are these. Is the applied method acceptable? Is there another way you'd prefer?
If that's the chip bug, then wrapping readl() is necessary. Can you rename to ohci_readl() (no Mixed_CASE_sYmbols!) and move that to the end of ohci.h? Also: the comments assuming PCI should go; so should the dbg() macros (use pr_debug or ohci_dbg); didn't apply on 2.6.5-mm.
I like the way this uses platform_bus instead of a chip-specific bus. In fact, only the clock start/stop looks particularly specific to the LH7A404 -- so far!
- Dave
------------------------------------------------------- This SF.Net email is sponsored by: IBM Linux Tutorials Free Linux tutorial presented by Daniel Robbins, President and CEO of GenToo technologies. Learn everything from fundamentals to system administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click _______________________________________________ [EMAIL PROTECTED] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel