Migration of pseries is broken with TCG because
QEMU tries to restore KVM MMU state unconditionally.
The result is a SIGSEGV in kvm_vm_ioctl():
#0 kvm_vm_ioctl (s=0x0, type=-2146390353)
at qemu/accel/kvm/kvm-all.c:2032
#1 0x00000001003e3e2c in kvmppc_configure_v3_mmu (cpu=<optimized out>,
radix=<optimized out>, gtse=<optimized out>, proc_tbl=<optimized out>)
at qemu/target/ppc/kvm.c:396
#2 0x00000001002f8b88 in spapr_post_load (opaque=0x1019103c0,
version_id=<optimized out>) at qemu/hw/ppc/spapr.c:1578
#3 0x000000010059e4cc in vmstate_load_state (f=0x106230000,
vmsd=0x1009479e0 <vmstate_spapr>, opaque=0x1019103c0,
version_id=<optimized out>) at qemu/migration/vmstate.c:165
#4 0x00000001005987e0 in vmstate_load (f=<optimized out>, se=<optimized out>)
at qemu/migration/savevm.c:748
This patch fixes the problem by not calling the KVM function with the
TCG mode.
Fixes: d39c90f5f3 ("spapr: Fix migration of Radix guests")
Signed-off-by: Laurent Vivier <lvivier@redhat.com>
---
v2: fix the comment to keep GDB backtrace lines starting with '#'
hw/ppc/spapr.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/hw/ppc/spapr.c b/hw/ppc/spapr.c
index 9efddeaee5..a471de6cab 100644
--- a/hw/ppc/spapr.c
+++ b/hw/ppc/spapr.c
@@ -1570,7 +1570,7 @@ static int spapr_post_load(void *opaque, int version_id)
err = spapr_rtc_import_offset(&spapr->rtc, spapr->rtc_offset);
}
- if (spapr->patb_entry) {
+ if (kvm_enabled() && spapr->patb_entry) {
PowerPCCPU *cpu = POWERPC_CPU(first_cpu);
bool radix = !!(spapr->patb_entry & PATBE1_GR);
bool gtse = !!(cpu->env.spr[SPR_LPCR] & LPCR_GTSE);
--
2.13.6
On 11/28/2017 03:43 PM, Laurent Vivier wrote: > Migration of pseries is broken with TCG because > QEMU tries to restore KVM MMU state unconditionally. > > The result is a SIGSEGV in kvm_vm_ioctl(): > > #0 kvm_vm_ioctl (s=0x0, type=-2146390353) > at qemu/accel/kvm/kvm-all.c:2032 > #1 0x00000001003e3e2c in kvmppc_configure_v3_mmu (cpu=<optimized out>, > radix=<optimized out>, gtse=<optimized out>, proc_tbl=<optimized out>) > at qemu/target/ppc/kvm.c:396 > #2 0x00000001002f8b88 in spapr_post_load (opaque=0x1019103c0, > version_id=<optimized out>) at qemu/hw/ppc/spapr.c:1578 > #3 0x000000010059e4cc in vmstate_load_state (f=0x106230000, > vmsd=0x1009479e0 <vmstate_spapr>, opaque=0x1019103c0, > version_id=<optimized out>) at qemu/migration/vmstate.c:165 > #4 0x00000001005987e0 in vmstate_load (f=<optimized out>, se=<optimized out>) > at qemu/migration/savevm.c:748 > > This patch fixes the problem by not calling the KVM function with the > TCG mode. > > Fixes: d39c90f5f3 ("spapr: Fix migration of Radix guests") > Signed-off-by: Laurent Vivier <lvivier@redhat.com> > --- > v2: fix the comment to keep GDB backtrace lines starting with '#' > > hw/ppc/spapr.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/hw/ppc/spapr.c b/hw/ppc/spapr.c > index 9efddeaee5..a471de6cab 100644 > --- a/hw/ppc/spapr.c > +++ b/hw/ppc/spapr.c > @@ -1570,7 +1570,7 @@ static int spapr_post_load(void *opaque, int version_id) > err = spapr_rtc_import_offset(&spapr->rtc, spapr->rtc_offset); > } > > - if (spapr->patb_entry) { > + if (kvm_enabled() && spapr->patb_entry) { As an alternative, I believe you can also do: if (!!spapr->patb_entry) { because the !! will return false if patb_entry is either undefined (as in TCG) or zero (as set in spapr_reallocate_hpt). This is the same logic used in spapr_patb_entry_needed too. But I am fine with what you did and it's also clearer, so Reviewed-by: Daniel Henrique Barboza <danielhb@linux.vnet.ibm.com> > PowerPCCPU *cpu = POWERPC_CPU(first_cpu); > bool radix = !!(spapr->patb_entry & PATBE1_GR); > bool gtse = !!(cpu->env.spr[SPR_LPCR] & LPCR_GTSE);
On Tue, 2017-11-28 at 17:26 -0200, Daniel Henrique Barboza wrote: > > On 11/28/2017 03:43 PM, Laurent Vivier wrote: > > Migration of pseries is broken with TCG because > > QEMU tries to restore KVM MMU state unconditionally. > > > > The result is a SIGSEGV in kvm_vm_ioctl(): > > > > #0 kvm_vm_ioctl (s=0x0, type=-2146390353) > > at qemu/accel/kvm/kvm-all.c:2032 > > #1 0x00000001003e3e2c in kvmppc_configure_v3_mmu > > (cpu=<optimized out>, > > radix=<optimized out>, gtse=<optimized out>, > > proc_tbl=<optimized out>) > > at qemu/target/ppc/kvm.c:396 > > #2 0x00000001002f8b88 in spapr_post_load (opaque=0x1019103c0, > > version_id=<optimized out>) at qemu/hw/ppc/spapr.c:1578 > > #3 0x000000010059e4cc in vmstate_load_state (f=0x106230000, > > vmsd=0x1009479e0 <vmstate_spapr>, opaque=0x1019103c0, > > version_id=<optimized out>) at qemu/migration/vmstate.c:165 > > #4 0x00000001005987e0 in vmstate_load (f=<optimized out>, > > se=<optimized out>) > > at qemu/migration/savevm.c:748 > > > > This patch fixes the problem by not calling the KVM function with > > the > > TCG mode. > > > > Fixes: d39c90f5f3 ("spapr: Fix migration of Radix guests") > > Signed-off-by: Laurent Vivier <lvivier@redhat.com> > > --- > > v2: fix the comment to keep GDB backtrace lines starting with '#' > > > > hw/ppc/spapr.c | 2 +- > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > diff --git a/hw/ppc/spapr.c b/hw/ppc/spapr.c > > index 9efddeaee5..a471de6cab 100644 > > --- a/hw/ppc/spapr.c > > +++ b/hw/ppc/spapr.c > > @@ -1570,7 +1570,7 @@ static int spapr_post_load(void *opaque, int > > version_id) > > err = spapr_rtc_import_offset(&spapr->rtc, spapr- > > >rtc_offset); > > } > > > > - if (spapr->patb_entry) { > > + if (kvm_enabled() && spapr->patb_entry) { > > As an alternative, I believe you can also do: > > if (!!spapr->patb_entry) { > > because the !! will return false if patb_entry is either undefined > (as in Hmmm, no. patb_entry is definately used in tcg > TCG) or zero (as set in spapr_reallocate_hpt). This is the same logic > used > in spapr_patb_entry_needed too. > > But I am fine with what you did and it's also clearer, so > > > Reviewed-by: Daniel Henrique Barboza <danielhb@linux.vnet.ibm.com> > > > > > PowerPCCPU *cpu = POWERPC_CPU(first_cpu); > > bool radix = !!(spapr->patb_entry & PATBE1_GR); > > bool gtse = !!(cpu->env.spr[SPR_LPCR] & LPCR_GTSE); > >
On Tue, 2017-11-28 at 18:43 +0100, Laurent Vivier wrote: > Migration of pseries is broken with TCG because > QEMU tries to restore KVM MMU state unconditionally. > > The result is a SIGSEGV in kvm_vm_ioctl(): > > #0 kvm_vm_ioctl (s=0x0, type=-2146390353) > at qemu/accel/kvm/kvm-all.c:2032 > #1 0x00000001003e3e2c in kvmppc_configure_v3_mmu (cpu=<optimized > out>, > radix=<optimized out>, gtse=<optimized out>, > proc_tbl=<optimized out>) > at qemu/target/ppc/kvm.c:396 > #2 0x00000001002f8b88 in spapr_post_load (opaque=0x1019103c0, > version_id=<optimized out>) at qemu/hw/ppc/spapr.c:1578 > #3 0x000000010059e4cc in vmstate_load_state (f=0x106230000, > vmsd=0x1009479e0 <vmstate_spapr>, opaque=0x1019103c0, > version_id=<optimized out>) at qemu/migration/vmstate.c:165 > #4 0x00000001005987e0 in vmstate_load (f=<optimized out>, > se=<optimized out>) > at qemu/migration/savevm.c:748 > > This patch fixes the problem by not calling the KVM function with the > TCG mode. > > Fixes: d39c90f5f3 ("spapr: Fix migration of Radix guests") > Signed-off-by: Laurent Vivier <lvivier@redhat.com> Ah, guess I never hit this because I never tested tcg migration with a qemu compiled on a kvm enabled system. Nice catch :) Reviewed-by: Suraj Jitindar Singh <sjitindarsingh@gmail.com> > --- > v2: fix the comment to keep GDB backtrace lines starting with '#' > > hw/ppc/spapr.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/hw/ppc/spapr.c b/hw/ppc/spapr.c > index 9efddeaee5..a471de6cab 100644 > --- a/hw/ppc/spapr.c > +++ b/hw/ppc/spapr.c > @@ -1570,7 +1570,7 @@ static int spapr_post_load(void *opaque, int > version_id) > err = spapr_rtc_import_offset(&spapr->rtc, spapr- > >rtc_offset); > } > > - if (spapr->patb_entry) { > + if (kvm_enabled() && spapr->patb_entry) { > PowerPCCPU *cpu = POWERPC_CPU(first_cpu); > bool radix = !!(spapr->patb_entry & PATBE1_GR); > bool gtse = !!(cpu->env.spr[SPR_LPCR] & LPCR_GTSE);
On Wed, Nov 29, 2017 at 10:58:29AM +1100, Suraj Jitindar Singh wrote: > On Tue, 2017-11-28 at 18:43 +0100, Laurent Vivier wrote: > > Migration of pseries is broken with TCG because > > QEMU tries to restore KVM MMU state unconditionally. > > > > The result is a SIGSEGV in kvm_vm_ioctl(): > > > > #0 kvm_vm_ioctl (s=0x0, type=-2146390353) > > at qemu/accel/kvm/kvm-all.c:2032 > > #1 0x00000001003e3e2c in kvmppc_configure_v3_mmu (cpu=<optimized > > out>, > > radix=<optimized out>, gtse=<optimized out>, > > proc_tbl=<optimized out>) > > at qemu/target/ppc/kvm.c:396 > > #2 0x00000001002f8b88 in spapr_post_load (opaque=0x1019103c0, > > version_id=<optimized out>) at qemu/hw/ppc/spapr.c:1578 > > #3 0x000000010059e4cc in vmstate_load_state (f=0x106230000, > > vmsd=0x1009479e0 <vmstate_spapr>, opaque=0x1019103c0, > > version_id=<optimized out>) at qemu/migration/vmstate.c:165 > > #4 0x00000001005987e0 in vmstate_load (f=<optimized out>, > > se=<optimized out>) > > at qemu/migration/savevm.c:748 > > > > This patch fixes the problem by not calling the KVM function with the > > TCG mode. > > > > Fixes: d39c90f5f3 ("spapr: Fix migration of Radix guests") > > Signed-off-by: Laurent Vivier <lvivier@redhat.com> > > Ah, guess I never hit this because I never tested tcg migration with a > qemu compiled on a kvm enabled system. Nice catch :) Same here. In fact I had the kvm_enabled() check in the initial verions but removed later as kvmppc_configure_v3_mmu() was defined separately as nop for !CONFIG_KVM. But obviously the above combination wasn't covered. Regards, Bharata.
On Tue, Nov 28, 2017 at 06:43:10PM +0100, Laurent Vivier wrote: > Migration of pseries is broken with TCG because > QEMU tries to restore KVM MMU state unconditionally. > > The result is a SIGSEGV in kvm_vm_ioctl(): > > #0 kvm_vm_ioctl (s=0x0, type=-2146390353) > at qemu/accel/kvm/kvm-all.c:2032 > #1 0x00000001003e3e2c in kvmppc_configure_v3_mmu (cpu=<optimized out>, > radix=<optimized out>, gtse=<optimized out>, proc_tbl=<optimized out>) > at qemu/target/ppc/kvm.c:396 > #2 0x00000001002f8b88 in spapr_post_load (opaque=0x1019103c0, > version_id=<optimized out>) at qemu/hw/ppc/spapr.c:1578 > #3 0x000000010059e4cc in vmstate_load_state (f=0x106230000, > vmsd=0x1009479e0 <vmstate_spapr>, opaque=0x1019103c0, > version_id=<optimized out>) at qemu/migration/vmstate.c:165 > #4 0x00000001005987e0 in vmstate_load (f=<optimized out>, se=<optimized out>) > at qemu/migration/savevm.c:748 > > This patch fixes the problem by not calling the KVM function with the > TCG mode. > > Fixes: d39c90f5f3 ("spapr: Fix migration of Radix guests") > Signed-off-by: Laurent Vivier <lvivier@redhat.com> > --- > v2: fix the comment to keep GDB backtrace lines starting with '#' Applied to ppc-for-2.11. > > hw/ppc/spapr.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/hw/ppc/spapr.c b/hw/ppc/spapr.c > index 9efddeaee5..a471de6cab 100644 > --- a/hw/ppc/spapr.c > +++ b/hw/ppc/spapr.c > @@ -1570,7 +1570,7 @@ static int spapr_post_load(void *opaque, int version_id) > err = spapr_rtc_import_offset(&spapr->rtc, spapr->rtc_offset); > } > > - if (spapr->patb_entry) { > + if (kvm_enabled() && spapr->patb_entry) { > PowerPCCPU *cpu = POWERPC_CPU(first_cpu); > bool radix = !!(spapr->patb_entry & PATBE1_GR); > bool gtse = !!(cpu->env.spr[SPR_LPCR] & LPCR_GTSE); -- David Gibson | I'll have my music baroque, and my code david AT gibson.dropbear.id.au | minimalist, thank you. NOT _the_ _other_ | _way_ _around_! http://www.ozlabs.org/~dgibson
© 2016 - 2024 Red Hat, Inc.