Re: Fix fallout from tg3_readphy() value is zero on error

["David S. Miller" <davem@xxxxxxxxxxxxx>]

On Fri, 21 Jan 2005 12:52:31 +1100
Peter Chubb <peterc@xxxxxxxxxxxxxxxxxx> wrote:

> After DaveM's patch to make the `default' value returned by
> tg3_readphy() to be zero, the attached patch fixes the few cases where
> a zero changes the previous behaviur, and deletes a couple of
> initialisation to zero that are no longer necessary.

The setting to zero was to eliminate a gcc warning because
it's flow analysis in earlier versions can't tell that we
do always set it to some value.

I think it's wiser to just make sure every case checks the
tg3_read_phy() return value.  So I'll work on that.

So is your 16-way system working now with this patch on
top of the current driver?