| Message ID | 20210302174443.514363-1-imbrenda@linux.ibm.com (mailing list archive) |
|---|---|
| Headers | show |
| Series | s390/kvm: fix MVPG when in VSIE | expand |
On 02.03.21 18:44, Claudio Imbrenda wrote: > The current handling of the MVPG instruction when executed in a nested > guest is wrong, and can lead to the nested guest hanging. > > This patchset fixes the behaviour to be more architecturally correct, > and fixes the hangs observed. > > v4->v5 > * split kvm_s390_logical_to_effective so it can be reused for vSIE > * fix existing comments and add some more comments > * use the new split _kvm_s390_logical_to_effective in vsie_handle_mvpg > > v3->v4 > * added PEI_ prefix to DAT_PROT and NOT_PTE macros > * added small comment to explain what they are about > > v2->v3 > * improved some comments > * improved some variable and parameter names for increased readability > * fixed missing handling of page faults in the MVPG handler > * small readability improvements > > v1->v2 > * complete rewrite queued (with small fixups) for kvms390. Still not sure if this will land in master or next. Opinions? > > Claudio Imbrenda (3): > s390/kvm: split kvm_s390_logical_to_effective > s390/kvm: extend kvm_s390_shadow_fault to return entry pointer > s390/kvm: VSIE: correctly handle MVPG when in VSIE > > arch/s390/kvm/gaccess.c | 30 ++++++++++-- > arch/s390/kvm/gaccess.h | 35 ++++++++++--- > arch/s390/kvm/vsie.c | 106 ++++++++++++++++++++++++++++++++++++---- > 3 files changed, 151 insertions(+), 20 deletions(-) >
On 3/8/21 4:19 PM, Christian Borntraeger wrote: > On 02.03.21 18:44, Claudio Imbrenda wrote: >> The current handling of the MVPG instruction when executed in a nested >> guest is wrong, and can lead to the nested guest hanging. >> >> This patchset fixes the behaviour to be more architecturally correct, >> and fixes the hangs observed. >> >> v4->v5 >> * split kvm_s390_logical_to_effective so it can be reused for vSIE >> * fix existing comments and add some more comments >> * use the new split _kvm_s390_logical_to_effective in vsie_handle_mvpg >> >> v3->v4 >> * added PEI_ prefix to DAT_PROT and NOT_PTE macros >> * added small comment to explain what they are about >> >> v2->v3 >> * improved some comments >> * improved some variable and parameter names for increased readability >> * fixed missing handling of page faults in the MVPG handler >> * small readability improvements >> >> v1->v2 >> * complete rewrite > > > queued (with small fixups) for kvms390. Still not sure if this will land in master or next. > Opinions? I'd go for the next merge window >> >> Claudio Imbrenda (3): >> s390/kvm: split kvm_s390_logical_to_effective >> s390/kvm: extend kvm_s390_shadow_fault to return entry pointer >> s390/kvm: VSIE: correctly handle MVPG when in VSIE >> >> arch/s390/kvm/gaccess.c | 30 ++++++++++-- >> arch/s390/kvm/gaccess.h | 35 ++++++++++--- >> arch/s390/kvm/vsie.c | 106 ++++++++++++++++++++++++++++++++++++---- >> 3 files changed, 151 insertions(+), 20 deletions(-) >>
On Mon, 8 Mar 2021 16:26:58 +0100 Janosch Frank <frankja@linux.ibm.com> wrote: > On 3/8/21 4:19 PM, Christian Borntraeger wrote: > > On 02.03.21 18:44, Claudio Imbrenda wrote: > >> The current handling of the MVPG instruction when executed in a > >> nested guest is wrong, and can lead to the nested guest hanging. > >> > >> This patchset fixes the behaviour to be more architecturally > >> correct, and fixes the hangs observed. > >> > >> v4->v5 > >> * split kvm_s390_logical_to_effective so it can be reused for vSIE > >> * fix existing comments and add some more comments > >> * use the new split _kvm_s390_logical_to_effective in > >> vsie_handle_mvpg > >> > >> v3->v4 > >> * added PEI_ prefix to DAT_PROT and NOT_PTE macros > >> * added small comment to explain what they are about > >> > >> v2->v3 > >> * improved some comments > >> * improved some variable and parameter names for increased > >> readability > >> * fixed missing handling of page faults in the MVPG handler > >> * small readability improvements > >> > >> v1->v2 > >> * complete rewrite > > > > > > queued (with small fixups) for kvms390. Still not sure if this will > > land in master or next. Opinions? > > I'd go for the next merge window I agree > >> > >> Claudio Imbrenda (3): > >> s390/kvm: split kvm_s390_logical_to_effective > >> s390/kvm: extend kvm_s390_shadow_fault to return entry pointer > >> s390/kvm: VSIE: correctly handle MVPG when in VSIE > >> > >> arch/s390/kvm/gaccess.c | 30 ++++++++++-- > >> arch/s390/kvm/gaccess.h | 35 ++++++++++--- > >> arch/s390/kvm/vsie.c | 106 > >> ++++++++++++++++++++++++++++++++++++---- 3 files changed, 151 > >> insertions(+), 20 deletions(-) >