similar to: Hang on resume - PGRAPH TLB flush idle timeout fail

Try booting with `nouveau.nofbaccel=1` and see if that helps. Some many kernels ago, acceleration of the console on fb resume got fixed, and is now causing issues for you and a few others. This will turn off fbcon accel, which should hopefully sidestep the issue. On Sun, Jan 4, 2015 at 5:32 AM, Lukasz Stelmach <stlman at poczta.fm> wrote: > Hello, > > I'd like to bring to your

Hello, I'd like to bring to your attention a bug report[1] I have filed on the Gentoo bugzilla about a regression (quite an old one) in the nouveau driver. If there is anything missing in the report that could help fixing the problem, please feel free to ask. [1] https://bugs.gentoo.org/show_bug.cgi?id=534038 -- Było mi bardzo miło. Twoje oczy lubią mnie >Łukasz<

https://bugs.freedesktop.org/show_bug.cgi?id=69465 Priority: medium Bug ID: 69465 Assignee: nouveau at lists.freedesktop.org Summary: Kernel freeze with NVA3 kernel 3.11.1 / PGRAPH][0000:01:00.0] PGRAPH TLB flush idle timeout fail QA Contact: xorg-team at lists.x.org Severity: critical

https://bugs.freedesktop.org/show_bug.cgi?id=109397 Bug ID: 109397 Summary: PGRAPH TLB flush idle timeout fail: X hangs, ssh login possible Product: Mesa Version: unspecified Hardware: x86 (IA32) OS: Linux (All) Status: NEW Severity: major Priority: medium Component:

Hi everyone, I was wondering what to do to dig into this problem further.The kernel is several weeks old (nouveau tree), it's at commit 000463f13fba6b2f94a5bfcb0d615751ae9c34a0. As you can see from the mesages below the problem is reproducable to the point of getting exactly the same error. Nov 25 17:57:51 madman kernel: [548360.773743] nouveau E[ PGRAPH][0000:01:00.0] PGRAPH TLB flush idle

Now it outputs: nouveau E[ PGRAPH][0000:02:00.0] PGRAPH TLB flush idle timeout fail nouveau E[ PGRAPH][0000:02:00.0] PGRAPH_STATUS: BUSY DISPATCH VFETCH CCACHE_UNK4 STRMOUT_GSCHED_UNK5 UNK14XX UNK1CXX CLIPID ZCULL ENG2D UNK34XX TPRAST TPROP ROP (0x011fde03) nouveau E[ PGRAPH][0000:02:00.0] PGRAPH_VSTATUS_0: CCACHE (0x00145b4d) nouveau E[ PGRAPH][0000:02:00.0] PGRAPH_VSTATUS_1: (0x0000002d)

https://bugs.freedesktop.org/show_bug.cgi?id=110572 Bug ID: 110572 Summary: System Crash: nouveau 0000:08:00.0: gr: PGRAPH TLB flush idle timeout fail and nouveau 0000:08:00.0: mmu: ce0 mmu invalidate timeout Product: Mesa Version: 19.0 Hardware: x86-64 (AMD64) OS: Linux (All)

Now it outputs: nouveau E[ PGRAPH][0000:02:00.0] PGRAPH TLB flush idle timeout fail nouveau E[ PGRAPH][0000:02:00.0] PGRAPH_STATUS: BUSY DISPATCH VFETCH CCACHE_UNK4 STRMOUT_GSCHED_UNK5 UNK14XX UNK1CXX CLIPID ZCULL ENG2D UNK34XX TPRAST TPROP ROP (0x011fde03) nouveau E[ PGRAPH][0000:02:00.0] PGRAPH_VSTATUS: CCACHE (0x00145b4d) (0x0000002d) ENG2D ROP (0x0034db40) instead of: [drm] nouveau

blktap_poll is calling tlb_flush_all() in its main ring buffer polling loop. This seems to be superfluous: the hypervisor should be performing any necessary tlb flushes on grant table operations performed by the back-end. Even a simple memory barrier is unnecessary here as the RING_PUSH_REQUESTS() call performs a wmb() anyway. And tlb_flush_all() is not exported to modules, so this call

On 27/05/19 11:47, Peter Zijlstra wrote: > On Sat, May 25, 2019 at 10:54:50AM +0200, Juergen Gross wrote: >> On 25/05/2019 10:22, Nadav Amit wrote: > >>> diff --git a/arch/x86/include/asm/paravirt_types.h b/arch/x86/include/asm/paravirt_types.h >>> index 946f8f1f1efc..3a156e63c57d 100644 >>> --- a/arch/x86/include/asm/paravirt_types.h >>> +++

On Mon, May 27, 2019 at 12:21:59PM +0200, Paolo Bonzini wrote: > On 27/05/19 11:47, Peter Zijlstra wrote: > > --- a/arch/x86/kernel/kvm.c > > +++ b/arch/x86/kernel/kvm.c > > @@ -580,7 +580,7 @@ static void __init kvm_apf_trap_init(voi > > > > static DEFINE_PER_CPU(cpumask_var_t, __pv_tlb_mask); > > > > -static void kvm_flush_tlb_others(const

On 6/25/19 7:35 PM, Nadav Amit wrote: >>> const struct flush_tlb_info *f = info; >>> + enum tlb_flush_reason reason; >>> + >>> + reason = (f->mm == NULL) ? TLB_LOCAL_SHOOTDOWN : TLB_LOCAL_MM_SHOOTDOWN; >> >> Should we just add the "reason" to flush_tlb_info? It's OK-ish to imply >> it like this, but seems like it would be

On Wed, Jun 12, 2019 at 11:49 PM Nadav Amit <namit at vmware.com> wrote: > > To improve TLB shootdown performance, flush the remote and local TLBs > concurrently. Introduce flush_tlb_multi() that does so. The current > flush_tlb_others() interface is kept, since paravirtual interfaces need > to be adapted first before it can be removed. This is left for future > work. In

> On Jun 25, 2019, at 8:36 PM, Andy Lutomirski <luto at kernel.org> wrote: > > On Wed, Jun 12, 2019 at 11:49 PM Nadav Amit <namit at vmware.com> wrote: >> To improve TLB shootdown performance, flush the remote and local TLBs >> concurrently. Introduce flush_tlb_multi() that does so. The current >> flush_tlb_others() interface is kept, since paravirtual

> On Jul 22, 2019, at 12:14 PM, Peter Zijlstra <peterz at infradead.org> wrote: > > On Thu, Jul 18, 2019 at 05:58:32PM -0700, Nadav Amit wrote: >> @@ -709,8 +716,9 @@ void native_flush_tlb_others(const struct cpumask *cpumask, >> * doing a speculative memory access. >> */ >> if (info->freed_tables) { >> - smp_call_function_many(cpumask,

Hi, Patch below is needed to make my system work stable in PAE mode. Havn''t seen problems without PAE, not sure whenever thats just pure luck or whenever there is a bug in my PAE xenlinux kernel. To me it looks like a generic bug though. I''ve actually trapped into problems with unpin only: A process exits, somewhere in exit_mm() the page tables are unpinned, shortly thereafter

On 25/05/2019 10:22, Nadav Amit wrote: > To improve TLB shootdown performance, flush the remote and local TLBs > concurrently. Introduce flush_tlb_multi() that does so. The current > flush_tlb_others() interface is kept, since paravirtual interfaces need > to be adapted first before it can be removed. This is left for future > work. In such PV environments, TLB flushes are not

On 31/05/2019 08:36, Nadav Amit wrote: > To improve TLB shootdown performance, flush the remote and local TLBs > concurrently. Introduce flush_tlb_multi() that does so. The current > flush_tlb_others() interface is kept, since paravirtual interfaces need > to be adapted first before it can be removed. This is left for future > work. In such PV environments, TLB flushes are not

> On Jul 3, 2019, at 7:04 AM, Juergen Gross <jgross at suse.com> wrote: > > On 03.07.19 01:51, Nadav Amit wrote: >> To improve TLB shootdown performance, flush the remote and local TLBs >> concurrently. Introduce flush_tlb_multi() that does so. Introduce >> paravirtual versions of flush_tlb_multi() for KVM, Xen and hyper-v (Xen >> and hyper-v are only

On Thu, Jul 18, 2019 at 05:58:32PM -0700, Nadav Amit wrote: > @@ -709,8 +716,9 @@ void native_flush_tlb_others(const struct cpumask *cpumask, > * doing a speculative memory access. > */ > if (info->freed_tables) { > - smp_call_function_many(cpumask, flush_tlb_func_remote, > - (void *)info, 1); > + __smp_call_function_many(cpumask, flush_tlb_func_remote,