On Thu, Feb 14, 2013 at 09:25:56PM +0100, Ivo Van Doorn wrote: > This is incorrect, all other debugfs files are read/write, so there is > no point in a extra commit file. The other eeprom debugfs files only modify the in-memory copy of the eeprom data; they do not actually write it back out to the eeprom. Furthermore, the rt2800usb driver often "validates" (ie modifies) the in-memory copy of the eeprom data, so I needed to have a way to re-read the raw data so I would only modify what was absolutely necessary. (IIRC the PCI-based modules expose a standard I2C eeprom which can be accessed/updated independently) > rt2800 specific code doesn't belong in the generic code. You should > make your solution completely driver independent. I see your point. This should come via bolting on generic eeprom read/write functions to rt2x00lib_ops, and calling those from the debugfs hooks, correct? Alternatively this could be bolted up to ethtool's eeprom manipulation hooks instead of the debugfs hook I hacked in there. Let me know what you'd prefer. > Have you checked if this is specific to rt2800 or would other Ralink > USB devices be able to work with this as well? I only have access to RT2870 modules at the moment, so I don't have any way of verifying this applies to non-RT2800 modules. I suppose I could wade through ralink's posted source code, but the comments in rt2x00usb.h seem to indicate eeprom write support is generic. - Solomon -- Solomon Peachy pizza at shaftnet dot org Melbourne, FL ^^ (mail/jabber/gtalk) ^^ Quidquid latine dictum sit, altum viditur.
Attachment:
pgpqbRJFDmotS.pgp
Description: PGP signature
