Re: [RFC/T] b43: to few loop tries in do_dummy_tx

[Larry Finger <Larry.Finger@lwfinger.net>]

Stefanik G=E1bor wrote:
>=20
> It would be good to know whether the Windows driver also suffers from
> PHY transmission errors... The "hybrid" Linux driver likely does have
> them, possibly even more than us, judging from its poor performance.
>=20

We didn't see the PHY transmission errors in b43 until fairly
recently. Unfortunately, they are too intermittent to tell if it is a
regression, and more importantly, it would be a bitch to bisect the
code as one would need many weeks of testing before saying that a
particular kernel was good.

When I first got 64-bit dma working on a BCM4311/2, there was always
one such error. That was removed when Michael changed the dma code.
Now I do not get any on most runs, but every once in a while (3 times
since July!), I get a storm of these.

Unlike most of you, I have seen the Broadcom code. There is almost no
error checking in their routines, which is good for us - it makes the
code smaller and the decompilation easier. I will never RE the binary
blob of the hybrid driver for a number of reasons; however, I doubt
that their culture has changed. As to why their driver has such poor
performance, your guess is as good as mine.

Larry
--
To unsubscribe from this list: send the line "unsubscribe linux-wireles=
s" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html