Hi Jean, > On Tue, 15 May 2007 12:53:18 -0600, David Hubbard wrote: > > On 5/15/07, Jean Delvare <khali at linux-fr.org> wrote: > > > This confirms that the configuration space of the W83627DHG is mapped > > > to 0x4e/0x4f. This explains the log message that you reported: the > > > driver printed it after it failed to find a device at 0x2e/0x2f. Then > > > it tried 0x4e/0x4f and succeeded, so the message was only a warning, > > > not an error. > > > > > > The w83627ehf driver should be changed to _not_ print a message when the > > > device ID reads 0xffff (or 0x0000, for that matter) as it means there > > > is _no_ chip at this address. > > > > > > On top of that, the driver should _not_ print a message by default when > > > it finds an unsupported ID. It is perfectly valid to have a W83627DHG > > > at 0x4e/0x4f and another LPC chip at 0x2e/0x2f. I think we already > > > heard of boards with two LPC chips, and we don't want to fill the log > > > with irrelevant messages in that case. So I believe that the driver > > > should only print this message when compiled with debugging enabled, as > > > the w83627hf driver is doing. > > > > > > David, can you please submit a patch doing that? > > > > Yes, I'd consider it a bug when it finds an OK chip at 0x43. I'll look > > at how w83627hf driver does it and copy that. > > Do you have a patch ready? A bug report now exists on kernel.org: > http://bugzilla.kernel.org/show_bug.cgi?id=8593 Thanks for the heads up. I'm still working on getting rid of i2c-isa, but I'll include a patch to fix this in my next update. David
