Date: Sat, 22 Feb 2003 12:46:55 +0900 From: Katsushi Kobayashi <ikob@koganei.wide.ad.jp> To: Marcel Moolenaar <marcel@xcllnt.net> Cc: firewire@FreeBSD.ORG Subject: Re: Bad news: bus resets not fixed yet Message-ID: <49532804-4618-11D7-8263-0030657D3460@koganei.wide.ad.jp> In-Reply-To: <20030222002547.GA1247@dhcp01.pn.xcllnt.net>
next in thread | previous in thread | raw e-mail | index | archive | help
Hello,
Bus-rest on firewire will be happen, when
- Network topology is changed (e.g. new device is attached)
- Unstable network topology probably due to cable or connector failure.
- A device initiated BUS reset some reason (e.g. the device want to get
some "master"
feature)
- Turn on a device
.....
Anyway, Bus-rest could be initiated by any connected device and lower
(uncontrollable from OS as PHY chipset). And the node identifier of the
firewire is dynamically assigned at every bus reset. OS can ignore the
interrupt happen at the Bus-rest. However, any communication action
after the bus-rest is ensured, since the node identifier could be
already
changed.
Is the above is answer for your question ?
On 2003.Feb.22, at 09:25 AM, Marcel Moolenaar wrote:
> Gang,
>
> I rebooted after Hidetoshi-san's patch has been committed and I
> also made sure to include ken's cd(4) fixes, but unfortunately
> the bus resets have returned. We may have timing problems that
> were hidden by the additional overhead of emitting the debug info.
> Unfortunately I don't have much time to experiment with that, but
> do want to pitch the following question(s):
> o Aren't bus resets expensive operations in general/principle?
> o My logs show 6 resets per second (unoptimized :-). Isn't this
> high no matter if there's a bug or not?
> o Are we not supposed to give devices some time to react upon the
> or otherwise build in some safety?
>
> \begin{/var/log/messages snippet}
> Feb 21 12:22:26 dhcp01 kernel: fwohci0: BUS reset
> Feb 21 12:22:27 dhcp01 kernel: fwohci0: node_id = 0xc000ffc1,
> CYCLEMASTER mode
> Feb 21 12:22:27 dhcp01 kernel: firewire0: 2 nodes, maxhop <= 1, cable
> IRM = 1 (me)
> Feb 21 12:22:27 dhcp01 kernel: fwohci0: BUS reset
> Feb 21 12:22:27 dhcp01 kernel: fwohci0: node_id = 0xc000ffc1,
> CYCLEMASTER mode
> Feb 21 12:22:27 dhcp01 kernel: firewire0: 2 nodes, maxhop <= 1, cable
> IRM = 1 (me)
> Feb 21 12:22:27 dhcp01 kernel: fwohci0: BUS reset
> Feb 21 12:22:27 dhcp01 kernel: fwohci0: node_id = 0xc000ffc1,
> CYCLEMASTER mode
> Feb 21 12:22:27 dhcp01 kernel: firewire0: 2 nodes, maxhop <= 1, cable
> IRM = 1 (me)
> Feb 21 12:22:27 dhcp01 kernel: fwohci0: BUS reset
> Feb 21 12:22:27 dhcp01 kernel: fwohci0: node_id = 0xc000ffc1,
> CYCLEMASTER mode
> Feb 21 12:22:27 dhcp01 kernel: firewire0: 2 nodes, maxhop <= 1, cable
> IRM = 1 (me)
> Feb 21 12:22:27 dhcp01 kernel: fwohci0: BUS reset
> Feb 21 12:22:27 dhcp01 kernel: fwohci0: node_id = 0xc000ffc1,
> CYCLEMASTER mode
> Feb 21 12:22:27 dhcp01 kernel: firewire0: 2 nodes, maxhop <= 1, cable
> IRM = 1 (me)
> Feb 21 12:22:27 dhcp01 kernel: fwohci0: BUS reset
> Feb 21 12:22:27 dhcp01 kernel: fwohci0: node_id = 0xc000ffc1,
> CYCLEMASTER mode
> Feb 21 12:22:27 dhcp01 kernel: firewire0: 2 nodes, maxhop <= 1, cable
> IRM = 1 (me)
> Feb 21 12:22:27 dhcp01 kernel: fwohci0: BUS reset
> Feb 21 12:22:28 dhcp01 kernel: fwohci0: node_id = 0xc000ffc0,
> CYCLEMASTER mode
> Feb 21 12:22:28 dhcp01 kernel: firewire0: 1 nodes, maxhop <= 0, cable
> IRM = 0 (me)
> \end{/var/log/messages snippet}
>
> --
> Marcel Moolenaar USPA: A-39004 marcel@xcllnt.net
>
> To Unsubscribe: send mail to majordomo@FreeBSD.org
> with "unsubscribe freebsd-firewire" in the body of the message
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-firewire" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?49532804-4618-11D7-8263-0030657D3460>