From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Greaves Subject: Re: kernel boot hang, SATA_VIA compiled without APIC_IO Date: Sun, 12 Dec 2004 20:43:27 +0000 Message-ID: <41BCAD6F.50007@dgreaves.com> References: <200412030345.45282.eldad@stoneshaft.ath.cx> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Return-path: Received: from s2.ukfsn.org ([217.158.120.143]:13540 "EHLO mail.ukfsn.org") by vger.kernel.org with ESMTP id S262102AbULLUnc (ORCPT ); Sun, 12 Dec 2004 15:43:32 -0500 In-Reply-To: <200412030345.45282.eldad@stoneshaft.ath.cx> Sender: linux-ide-owner@vger.kernel.org List-Id: linux-ide@vger.kernel.org To: Eldad Zack Cc: linux-kernel@vger.kernel.org, linux-ide@vger.kernel.org, jgarzik@pobox.com Thanks for posting this Eldad, I had the same problem and this saved lots of trial and error!! As an FYI all I did was exchange my ASUS A7V600 with an ASUS A7V600-X. Maybe the BIOS changed version, maybe the chipset is different - not easy to tell. Since I boot from my SATA drive this was quite troubling. Luckily I still had an old 2.6.6 kernel. I had to revert back to a 2.6.6 kernel, 2.6.7 wouldn't work. Jeff (or whoever) I too would be happy to help debug if there's anything I can do. I can't check the exact failure point for a week (away from the machine) but here's the relevant bit of the dmesg (admitedley from a good booting kernel with APIC_IO set) showing (simulated) the point where the failure occured: libata version 1.02 loaded. sata_via version 0.20 ACPI: PCI interrupt 0000:00:0f.0[B] -> GSI 20 (level, low) -> IRQ 20 sata_via(0000:00:0f.0): routed to hard irq line 0 ata1: SATA max UDMA/133 cmd 0xB400 ctl 0xB002 bmdma 0xA000 irq 20 ata2: SATA max UDMA/133 cmd 0xA800 ctl 0xA402 bmdma 0xA008 irq 20 ata1: dev 0 After I get back on 19th Dec I'll be happy to try additional tests, provide additional info etc if anyone asks. HTH David Eldad Zack wrote: >Hello, > >I've recently got a SATA capable machine (Via chipset) and I've exprienced a >nasty hang at boottime, using kernel 2.6.9. >After some recompiling different parameters it boiled down to APIC_IO being >not selected (this is a UP machine). > >Without APIC_IO selected the system would hang while loading SATA. > >I've only tried 2.6.5 to notice it would not hang but would emit messeges as >"hde: lost interrupt", and eventually go on with the boot, the sata being >unusable. > >Out of curiousity, I'd like to know if APIC_IO is absolutly needed when >dealing with SATA, and also, I'd like to help debug this problem so that a >kernel compiled without APIC_IO would at the very least not hang... > > > >