That's 9k-geoff, not 9legacy :)
Vào 08:48, T.6, 20 Th3, 2026 adi <[email protected]> đã viết:
> I think managed to adapt it to 9k:
>
> diff -c /n/dump/2026/0318/sys/src/9k/k10/trap.c /sys/src/9k/k10/trap.c
> /n/dump/2026/0318/sys/src/9k/k10/trap.c:622,628 -
> /sys/src/9k/k10/trap.c:622,628
> if(vec->irq == IdtPIC+IrqCLOCK || vec->irq ==
> IdtTIMER)
> clockintr = 1;
> else
> - if(up)
> + if(up && m->ilockdepth == 0)
> preempted();
> }
> } else if (user)
> /n/dump/2026/0318/sys/src/9k/k10/trap.c:632,638 -
> /sys/src/9k/k10/trap.c:632,638
> splhi();
>
> /* delaysched set because we held a lock or because our quantum
> ended */
> - if(up && up->delaysched && clockintr){
> + if(up && up->delaysched && clockintr && m->ilockdepth == 0){
> sched();
> splhi();
> }
> diff -c /n/dump/2026/0318/sys/src/9k/k10/x86watchdog.c
> /sys/src/9k/k10/x86watchdog.c
> /n/dump/2026/0318/sys/src/9k/k10/x86watchdog.c:140,145 -
> /sys/src/9k/k10/x86watchdog.c:140,146
> r = rdmsr(Msrmiscen);
> if(!(r & 0x80LL)) {
> print("x86watchdog: p4 can't enable\n");
> + iunlock(wd);
> return;
> }
> zeromsrs(0x300, 18); /* perfctr */
>
> On Tue, Mar 17, 2026 at 12:05:26PM +0700, hahahahacker2009 wrote:
> > Vào Thứ 3, 17 thg 3, 2026 vào lúc 09:14 David du Colombier
> > <[email protected]> đã viết:
> > >
> > > Please try this patch
> http://9legacy.org/9legacy/patch/pc-x86watchdog-ilock.diff
> > >
> > > It should fix the issue you encountered with KVM.
> > >
> >
> > This patch is for the 9/pc kernel, but I'm seeing the issue on 9k/k10
> kernel
> > > --
> > > David du Colombier
------------------------------------------
9fans: 9fans
Permalink:
https://9fans.topicbox.com/groups/9fans/Ta43f70982c6bc008-M5c16f598655d74fac4315781
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription
