Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Mon, 12 Nov 2001 16:59:13 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Mon, 12 Nov 2001 16:58:54 -0500 Received: from mail.pha.ha-vel.cz ([195.39.72.3]:34833 "HELO mail.pha.ha-vel.cz") by vger.kernel.org with SMTP id ; Mon, 12 Nov 2001 16:58:49 -0500 Date: Mon, 12 Nov 2001 22:58:32 +0100 From: Vojtech Pavlik To: Alan Cox Cc: Neale Banks , Jeronimo Pellegrini , Nils Philippsen , linux-kernel@vger.kernel.org Subject: Re: [PATCH] VIA timer fix was removed? Message-ID: <20011112225832.A19107@suse.cz> In-Reply-To:

Mime-Version: 1.0 Content-Type: multipart/mixed; boundary="u3/rZRmxL6MmkK24" Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: ; from alan@lxorguk.ukuu.org.uk on Mon, Nov 12, 2001 at 09:31:35PM +0000 Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org --u3/rZRmxL6MmkK24 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline On Mon, Nov 12, 2001 at 09:31:35PM +0000, Alan Cox wrote: > > Attached (untested) patch against 21.2.20 (which still has the $SUBJECT > > code) should implement timer=no-via686a option to disable this. Hopefully > > I'll get it tested in the next day or two. > > This isnt the real problem - we are seeing it triggered by cases we dont > underatand that seem to be software. We need to find those I don't think it's software. At least it's definitely not locking. It's happening on machines where none of the drivers missing the locks are used (only ftape and analog joystick). I think it's the old Neptune bug biting us again. I'd like to verify this theory - do you have any list of people who've seen the VIA bugfix triggered on non-VIA hardware? Some might be willing to test experimental patches ... ... like the one attached. It doesn't add anything that already hasn't been there, but: 1) Should be safe for machines that have a different bug than VIA. It'll print a message but won't reset the timer if the > LATCH reading isn't persistent. 2) Should printk enough data to shed some light on what is triggering the VIA check. Patch against 2.4.15-pre4. -- Vojtech Pavlik SuSE Labs --u3/rZRmxL6MmkK24 Content-Type: text/plain; charset=us-ascii Content-Disposition: attachment; filename="new-via-check.diff" diff -urN linux-2.4.15-pre4/arch/i386/kernel/time.c linux/arch/i386/kernel/time.c --- linux-2.4.15-pre4/arch/i386/kernel/time.c Mon Nov 12 22:31:52 2001 +++ linux/arch/i386/kernel/time.c Mon Nov 12 22:52:25 2001 @@ -112,6 +112,50 @@ return delay_at_last_interrupt + edx; } +/* + * VIA hardware bug workaround with check if it is really needed and + * a printk that could tell us what's exactly happening on machines which + * trigger the check, but are not VIA-based. + * + * Must be called with the i8253_spinlock held. + */ + +static void via_reset_and_whine(int *count) +{ + static unsigned long last_whine = 0; + unsigned long new_whine; + int count2; + + new_whine = last_whine; + + outb_p(0x00, 0x43); /* Re-read the timer */ + count2 = inb_p(0x40); + count2 |= inb(0x40) << 8; + + if (time_after(jiffies, last_whine)) { + printk(KERN_WARNING "timer.c: VIA bug check triggered. " + "Value read %d [%#x], re-read %d [%#x]\n", + *count, *count, count2, count2); + new_whine = jiffies + HZ; + } + + *count = count2; + + if (count2 > LATCH) { /* Still bad */ + if (time_after(jiffies, last_whine)) { + printk(KERN_WARNING "timer.c VIA bug really present. " + "Resetting PIT timer.\n"); + new_whine = jiffies + HZ; + } + outb_p(0x34, 0x43); + outb_p(LATCH & 0xff, 0x40); + outb(LATCH >> 8, 0x40); + *count = LATCH - 1; + } + + last_whine = new_whine; +} + #define TICK_SIZE tick spinlock_t i8253_lock = SPIN_LOCK_UNLOCKED; @@ -180,12 +224,8 @@ count |= inb_p(0x40) << 8; /* VIA686a test code... reset the latch if count > max + 1 */ - if (count > LATCH) { - outb_p(0x34, 0x43); - outb_p(LATCH & 0xff, 0x40); - outb(LATCH >> 8, 0x40); - count = LATCH - 1; - } + if (count > LATCH) + via_reset_and_whine(&count); spin_unlock(&i8253_lock); @@ -501,6 +541,10 @@ count = inb_p(0x40); /* read the latched count */ count |= inb(0x40) << 8; + + if (count > LATCH) + via_reset_and_whine(&count); + spin_unlock(&i8253_lock); count = ((LATCH-1) - count) * TICK_SIZE; --u3/rZRmxL6MmkK24-- - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/