| Message ID | 1458047717-25052-1-git-send-email-lvivier@redhat.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
Missing in the subject: it's a kvm-unit-tests patch. Laurent On 15/03/2016 14:15, Laurent Vivier wrote: > When they are started, processors are in 32-bit mode, > as we are testing ppc64 processors, enable the 64bit mode > on starting. > > Signed-off-by: Laurent Vivier <lvivier@redhat.com> > --- > powerpc/cstart64.S | 9 +++++++++ > 1 file changed, 9 insertions(+) > > diff --git a/powerpc/cstart64.S b/powerpc/cstart64.S > index c87e3d6..634b854 100644 > --- a/powerpc/cstart64.S > +++ b/powerpc/cstart64.S > @@ -18,6 +18,15 @@ > .globl start > start: > FIXUP_ENDIAN > + > + /* enable 64-bit mode */ > + mfmsr r11 > + li r12,-1 > + rldicr r12,r12,0,0 > + or r11,r11,r12 > + mtmsrd r11 > + isync > + > /* > * We were loaded at QEMU's kernel load address, but we're not > * allowed to link there due to how QEMU deals with linker VMAs, > -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On 15.03.2016 14:15, Laurent Vivier wrote: > When they are started, processors are in 32-bit mode, > as we are testing ppc64 processors, enable the 64bit mode > on starting. > > Signed-off-by: Laurent Vivier <lvivier@redhat.com> > --- > powerpc/cstart64.S | 9 +++++++++ > 1 file changed, 9 insertions(+) > > diff --git a/powerpc/cstart64.S b/powerpc/cstart64.S > index c87e3d6..634b854 100644 > --- a/powerpc/cstart64.S > +++ b/powerpc/cstart64.S > @@ -18,6 +18,15 @@ > .globl start > start: > FIXUP_ENDIAN > + > + /* enable 64-bit mode */ > + mfmsr r11 > + li r12,-1 > + rldicr r12,r12,0,0 Nice idea to set the highest bit, I didn't know that one yet :-) > + or r11,r11,r12 > + mtmsrd r11 > + isync Looks good! Reviewed-by: Thomas Huth <thuth@redhat.com> PS: As far as I can see, the C code has already been compiled for 64 bits ... how did that work if it has been run in 32-bit mode so far?? -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On 15.03.16 19:43, Thomas Huth wrote: > On 15.03.2016 14:15, Laurent Vivier wrote: >> When they are started, processors are in 32-bit mode, >> as we are testing ppc64 processors, enable the 64bit mode >> on starting. >> >> Signed-off-by: Laurent Vivier <lvivier@redhat.com> >> --- >> powerpc/cstart64.S | 9 +++++++++ >> 1 file changed, 9 insertions(+) >> >> diff --git a/powerpc/cstart64.S b/powerpc/cstart64.S >> index c87e3d6..634b854 100644 >> --- a/powerpc/cstart64.S >> +++ b/powerpc/cstart64.S >> @@ -18,6 +18,15 @@ >> .globl start >> start: >> FIXUP_ENDIAN >> + >> + /* enable 64-bit mode */ >> + mfmsr r11 >> + li r12,-1 >> + rldicr r12,r12,0,0 > > Nice idea to set the highest bit, I didn't know that one yet :-) > >> + or r11,r11,r12 >> + mtmsrd r11 >> + isync > > Looks good! > > Reviewed-by: Thomas Huth <thuth@redhat.com> > > > PS: As far as I can see, the C code has already been compiled for 64 > bits ... how did that work if it has been run in 32-bit mode so far?? And why would we enter a 64bit CPU without MSR_SF set? Alex -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On 15.03.2016 19:44, Alexander Graf wrote: > > > On 15.03.16 19:43, Thomas Huth wrote: >> On 15.03.2016 14:15, Laurent Vivier wrote: >>> When they are started, processors are in 32-bit mode, >>> as we are testing ppc64 processors, enable the 64bit mode >>> on starting. >>> >>> Signed-off-by: Laurent Vivier <lvivier@redhat.com> >>> --- >>> powerpc/cstart64.S | 9 +++++++++ >>> 1 file changed, 9 insertions(+) >>> >>> diff --git a/powerpc/cstart64.S b/powerpc/cstart64.S >>> index c87e3d6..634b854 100644 >>> --- a/powerpc/cstart64.S >>> +++ b/powerpc/cstart64.S >>> @@ -18,6 +18,15 @@ >>> .globl start >>> start: >>> FIXUP_ENDIAN >>> + >>> + /* enable 64-bit mode */ >>> + mfmsr r11 >>> + li r12,-1 >>> + rldicr r12,r12,0,0 >> >> Nice idea to set the highest bit, I didn't know that one yet :-) >> >>> + or r11,r11,r12 >>> + mtmsrd r11 >>> + isync >> >> Looks good! >> >> Reviewed-by: Thomas Huth <thuth@redhat.com> >> >> >> PS: As far as I can see, the C code has already been compiled for 64 >> bits ... how did that work if it has been run in 32-bit mode so far?? > > And why would we enter a 64bit CPU without MSR_SF set? Ah, I see, 64-bit mode is already enabled in translate_init.c for 64-bit CPUs! I somehow thought that 64-bit PPCs would start in 32-bit mode (for backwards compatibility), and thus I expected that the VMs would also be started in 32-bit mode. But seems like this is not the case... and I learned again a new detail about PPCs :-) Thomas -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On 15/03/2016 19:44, Alexander Graf wrote: > > > On 15.03.16 19:43, Thomas Huth wrote: >> On 15.03.2016 14:15, Laurent Vivier wrote: >>> When they are started, processors are in 32-bit mode, >>> as we are testing ppc64 processors, enable the 64bit mode >>> on starting. >>> >>> Signed-off-by: Laurent Vivier <lvivier@redhat.com> >>> --- >>> powerpc/cstart64.S | 9 +++++++++ >>> 1 file changed, 9 insertions(+) >>> >>> diff --git a/powerpc/cstart64.S b/powerpc/cstart64.S >>> index c87e3d6..634b854 100644 >>> --- a/powerpc/cstart64.S >>> +++ b/powerpc/cstart64.S >>> @@ -18,6 +18,15 @@ >>> .globl start >>> start: >>> FIXUP_ENDIAN >>> + >>> + /* enable 64-bit mode */ >>> + mfmsr r11 >>> + li r12,-1 >>> + rldicr r12,r12,0,0 >> >> Nice idea to set the highest bit, I didn't know that one yet :-) >> >>> + or r11,r11,r12 >>> + mtmsrd r11 >>> + isync >> >> Looks good! >> >> Reviewed-by: Thomas Huth <thuth@redhat.com> >> >> >> PS: As far as I can see, the C code has already been compiled for 64 >> bits ... how did that work if it has been run in 32-bit mode so far?? > > And why would we enter a 64bit CPU without MSR_SF set? I don't know, but: When I start "qemu-system-ppc64 -machine pseries -s -S" and check the MSR with gdb, the MSR is always set to 0 (or "info registers" in monitor). It's why I've added this initialization. So, perhaps there is a bug in QEMU? [Perhaps "env->msr |= (1ULL << MSR_SF)" is overwritten by the following hreg_store_msr(env, msr, 1);"?] Laurent -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On 15/03/2016 19:44, Alexander Graf wrote: > > > On 15.03.16 19:43, Thomas Huth wrote: >> On 15.03.2016 14:15, Laurent Vivier wrote: >>> When they are started, processors are in 32-bit mode, >>> as we are testing ppc64 processors, enable the 64bit mode >>> on starting. >>> >>> Signed-off-by: Laurent Vivier <lvivier@redhat.com> >>> --- >>> powerpc/cstart64.S | 9 +++++++++ >>> 1 file changed, 9 insertions(+) >>> >>> diff --git a/powerpc/cstart64.S b/powerpc/cstart64.S >>> index c87e3d6..634b854 100644 >>> --- a/powerpc/cstart64.S >>> +++ b/powerpc/cstart64.S >>> @@ -18,6 +18,15 @@ >>> .globl start >>> start: >>> FIXUP_ENDIAN >>> + >>> + /* enable 64-bit mode */ >>> + mfmsr r11 >>> + li r12,-1 >>> + rldicr r12,r12,0,0 >> >> Nice idea to set the highest bit, I didn't know that one yet :-) >> >>> + or r11,r11,r12 >>> + mtmsrd r11 >>> + isync >> >> Looks good! >> >> Reviewed-by: Thomas Huth <thuth@redhat.com> >> >> >> PS: As far as I can see, the C code has already been compiled for 64 >> bits ... how did that work if it has been run in 32-bit mode so far?? > > And why would we enter a 64bit CPU without MSR_SF set? OK, it seems to be a bug in QEMU. I've sent a patch to QEMU ML trying to fix that. Laurent -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/powerpc/cstart64.S b/powerpc/cstart64.S index c87e3d6..634b854 100644 --- a/powerpc/cstart64.S +++ b/powerpc/cstart64.S @@ -18,6 +18,15 @@ .globl start start: FIXUP_ENDIAN + + /* enable 64-bit mode */ + mfmsr r11 + li r12,-1 + rldicr r12,r12,0,0 + or r11,r11,r12 + mtmsrd r11 + isync + /* * We were loaded at QEMU's kernel load address, but we're not * allowed to link there due to how QEMU deals with linker VMAs,
When they are started, processors are in 32-bit mode, as we are testing ppc64 processors, enable the 64bit mode on starting. Signed-off-by: Laurent Vivier <lvivier@redhat.com> --- powerpc/cstart64.S | 9 +++++++++ 1 file changed, 9 insertions(+)