Date: Sun, 30 Oct 2005 01:25:05 -0500 From: "Matthew D. Fuller" <fullermd@over-yonder.net> To: John Baldwin <jhb@freebsd.org> Cc: freebsd-current@freebsd.org Subject: Re: The case of the missing USB controllers Message-ID: <20051030062505.GA1367@over-yonder.net> In-Reply-To: <200510261321.50623.jhb@freebsd.org> References: <20051023044115.7050216A421@hub.freebsd.org> <200510251521.03036.jhb@freebsd.org> <20051025224508.GB74484@over-yonder.net> <200510261321.50623.jhb@freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Wed, Oct 26, 2005 at 01:21:49PM -0400 I heard the voice of John Baldwin, and lo! it spake thus: > > > > 2) Does it work ok if you disable APIC? > > > > > > 3) Do you have a verbose dmesg and mptable output available > > > somewhere? > > > > mptable -verbose attached. The dmesg will probably have to wait > > for the weekend; I really hate losing my state :| > > Ok. I kind of need both of them to make sense of things, but I can > wait, no rush. OK. Here we go. I updated another box to this afternoon's RELENG_4, and tried both a GENERIC[0] kernel, and a GENERIC+SMP+APIC[1]. I also booted my current -CURRENT kernel (~Oct 1) with APIC disabled[2] and normal[3]. Note the minor difference that for the RELENG_4 boots I had to enable the ATA controller, which is normally disabled. With the non-APIC/SMP, on both 4 and 7, the [onboard] USB worked. It took interrupts, the mouse responded. Unplugging and replugging it worked just fine. With APIC/SMP, on both 4 and 7, it failed to work and never took an interrupt. Note in the 4.smp dmesg there's a "No debugger in kernel" line where I Ctrl-Alt-Esc'd, right between the CPU1 launch and the uhub0 SET_ADDR_FAILED. It froze there for a long time (minute+). That same message comes up in 7 when I try to hotplug stuff; it shuts off power to that port, and so it won't probe anything plugged in after boot. Stuff plugged in at boot stays powered up, though (so at least I get to play with the LED on the mouse ;). But on 4 it didn't last that long. In none of the 4 combinations did the USB2 card (Adaptec AUA-4000, the package says) probe at all, other than perhaps in the bus enumeration (vendor 0x9004==Adaptec? There's 2 2940's in other PCI slots, so that sounds right). So, without the APIC, it seems to work well enough. Unfortunately, with this staggering CPU muscle, I can't really afford to lose half my processing power, even to get away from a terrifyingly nasty RS232 serial mouse... [0] <http://www.over-yonder.net/~fullermd/dl/dm/dmesg.4.generic>; [1] <http://www.over-yonder.net/~fullermd/dl/dm/dmesg.4.smp>; [2] <http://www.over-yonder.net/~fullermd/dl/dm/dmesg.7.noapic>; [3] <http://www.over-yonder.net/~fullermd/dl/dm/dmesg.7.normal>; -- Matthew Fuller (MF4839) | fullermd@over-yonder.net Systems/Network Administrator | http://www.over-yonder.net/~fullermd/ On the Internet, nobody can hear you scream.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20051030062505.GA1367>