[ltp] Re: Timeout with IBM minipci cvs atheros and 2.6
Alexander Gran
linux-thinkpad@linux-thinkpad.org
Mon, 26 Apr 2004 10:27:45 +0200
=2D----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Am Sonntag, 25. April 2004 22:41 schrieb Loic Minier:
> Indeed.
It helps a bit, if you modify the driver not to reinit, but to just reset t=
he=20
card:
change=20
static void
ath_tx_timeout(struct net_device *dev)
{
struct ath_softc *sc =3D dev->priv;
sc->sc_stats.ast_watchdog++;
#ifdef AR_DEBUG
if (ath_debug)
ath_hal_dumpstate(sc->sc_ah);
#endif
ath_init(dev);
}
to=20
static void
ath_tx_timeout(struct net_device *dev)
{
struct ath_softc *sc =3D dev->priv;
sc->sc_stats.ast_watchdog++;
#ifdef AR_DEBUG
if (ath_debug)
ath_hal_dumpstate(sc->sc_ah);
#endif
ath_reset(dev);
}
in driver/if_ath.c
This takes "just" 3 sec.
> Right now, I've turned off acpi, cpufreq and anything related to
> system/cpu states that could confuse ath drivers, result: did not help
> at all.
Same here. even switched to apm. No chance.
> The ath0 interface works fine as long as I don't traffic too much, if I
> do:
strange, relly. But at least it's the same bug i see.
> =3D> 4th interrupt position if you consider cascading.
I even got:
CPU0
0: 532251 XT-PIC timer
1: 2065 XT-PIC i8042
2: 0 XT-PIC cascade
3: 3 XT-PIC serial
7: 0 XT-PIC parport0
8: 4 XT-PIC rtc
9: 25032 XT-PIC yenta, uhci_hcd, radeon@PCI:1:0:0
10: 10304 XT-PIC yenta, uhci_hcd, ehci_hcd, Intel=20
82801DB-ICH4, Intel 82801DB-ICH4 Modem
11: 5423 XT-PIC uhci_hcd, ath0
12: 19360 XT-PIC i8042
14: 14243 XT-PIC ide0
15: 62 XT-PIC ide1
but not better.
Alex
=2D --=20
Encrypted Mails welcome.
PGP-Key at http://zodiac.dnsalias.org/misc/pgpkey.asc | Key-ID: 0x6D7DD291
=2D----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)
iD8DBQFAjMgB/aHb+2190pERAh+QAJ9pzfh8DKl8hXWRK1nt/lgiW2J7hwCglNV4
G5SkvaUdvyjbzkOQdKQIRog=3D
=3DRkeC
=2D----END PGP SIGNATURE-----