| Message ID | 20190426111003.21246-11-borntraeger@de.ibm.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | s390x: new guest features | expand |
On 26.04.19 13:10, Christian Borntraeger wrote: > 8561 and 8562 will be gen15 machines. There is no name yet, lets us use > the cpu id as base name. Later on we can provide aliases with the proper > name. > > Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com> > --- > target/s390x/cpu_models.c | 2 ++ > 1 file changed, 2 insertions(+) > > diff --git a/target/s390x/cpu_models.c b/target/s390x/cpu_models.c > index d683635eb5..dd6415103f 100644 > --- a/target/s390x/cpu_models.c > +++ b/target/s390x/cpu_models.c > @@ -83,6 +83,8 @@ static S390CPUDef s390_cpu_defs[] = { > CPUDEF_INIT(0x3906, 14, 1, 47, 0x08000000U, "z14", "IBM z14 GA1"), > CPUDEF_INIT(0x3906, 14, 2, 47, 0x08000000U, "z14.2", "IBM z14 GA2"), > CPUDEF_INIT(0x3907, 14, 1, 47, 0x08000000U, "z14ZR1", "IBM z14 Model ZR1 GA1"), > + CPUDEF_INIT(0x8561, 15, 1, 47, 0x08000000U, "8561", "IBM 8561 GA1"), > + CPUDEF_INIT(0x8562, 15, 1, 47, 0x08000000U, "8562", "IBM 8562 GA1"), > }; > > #define QEMU_MAX_CPU_TYPE 0x2827 > Can you fixup the comment at the beginning of the definitions, stating that base features of new HW are no longer a superset?
On 26.04.19 13:21, David Hildenbrand wrote: > On 26.04.19 13:10, Christian Borntraeger wrote: >> 8561 and 8562 will be gen15 machines. There is no name yet, lets us use >> the cpu id as base name. Later on we can provide aliases with the proper >> name. >> >> Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com> >> --- >> target/s390x/cpu_models.c | 2 ++ >> 1 file changed, 2 insertions(+) >> >> diff --git a/target/s390x/cpu_models.c b/target/s390x/cpu_models.c >> index d683635eb5..dd6415103f 100644 >> --- a/target/s390x/cpu_models.c >> +++ b/target/s390x/cpu_models.c >> @@ -83,6 +83,8 @@ static S390CPUDef s390_cpu_defs[] = { >> CPUDEF_INIT(0x3906, 14, 1, 47, 0x08000000U, "z14", "IBM z14 GA1"), >> CPUDEF_INIT(0x3906, 14, 2, 47, 0x08000000U, "z14.2", "IBM z14 GA2"), >> CPUDEF_INIT(0x3907, 14, 1, 47, 0x08000000U, "z14ZR1", "IBM z14 Model ZR1 GA1"), >> + CPUDEF_INIT(0x8561, 15, 1, 47, 0x08000000U, "8561", "IBM 8561 GA1"), >> + CPUDEF_INIT(0x8562, 15, 1, 47, 0x08000000U, "8562", "IBM 8562 GA1"), >> }; >> >> #define QEMU_MAX_CPU_TYPE 0x2827 >> > > Can you fixup the comment at the beginning of the definitions, stating > that base features of new HW are no longer a superset? Will do.
On 26.04.19 13:21, David Hildenbrand wrote: > On 26.04.19 13:10, Christian Borntraeger wrote: >> 8561 and 8562 will be gen15 machines. There is no name yet, lets us use >> the cpu id as base name. Later on we can provide aliases with the proper >> name. >> >> Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com> >> --- >> target/s390x/cpu_models.c | 2 ++ >> 1 file changed, 2 insertions(+) >> >> diff --git a/target/s390x/cpu_models.c b/target/s390x/cpu_models.c >> index d683635eb5..dd6415103f 100644 >> --- a/target/s390x/cpu_models.c >> +++ b/target/s390x/cpu_models.c >> @@ -83,6 +83,8 @@ static S390CPUDef s390_cpu_defs[] = { >> CPUDEF_INIT(0x3906, 14, 1, 47, 0x08000000U, "z14", "IBM z14 GA1"), >> CPUDEF_INIT(0x3906, 14, 2, 47, 0x08000000U, "z14.2", "IBM z14 GA2"), >> CPUDEF_INIT(0x3907, 14, 1, 47, 0x08000000U, "z14ZR1", "IBM z14 Model ZR1 GA1"), >> + CPUDEF_INIT(0x8561, 15, 1, 47, 0x08000000U, "8561", "IBM 8561 GA1"), >> + CPUDEF_INIT(0x8562, 15, 1, 47, 0x08000000U, "8562", "IBM 8562 GA1"), >> }; >> >> #define QEMU_MAX_CPU_TYPE 0x2827 >> > > Can you fixup the comment at the beginning of the definitions, stating > that base features of new HW are no longer a superset? > Something like /* * CPU definition list in order of release. Up to generation 14 base features * of a following release have been a superset of the previous release. With * generation 15 two features have been deprecated. */
On 26.04.19 13:10, Christian Borntraeger wrote: > 8561 and 8562 will be gen15 machines. There is no name yet, lets us use > the cpu id as base name. Later on we can provide aliases with the proper > name. > > Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com> > --- > target/s390x/cpu_models.c | 2 ++ > 1 file changed, 2 insertions(+) > > diff --git a/target/s390x/cpu_models.c b/target/s390x/cpu_models.c > index d683635eb5..dd6415103f 100644 > --- a/target/s390x/cpu_models.c > +++ b/target/s390x/cpu_models.c > @@ -83,6 +83,8 @@ static S390CPUDef s390_cpu_defs[] = { > CPUDEF_INIT(0x3906, 14, 1, 47, 0x08000000U, "z14", "IBM z14 GA1"), > CPUDEF_INIT(0x3906, 14, 2, 47, 0x08000000U, "z14.2", "IBM z14 GA2"), > CPUDEF_INIT(0x3907, 14, 1, 47, 0x08000000U, "z14ZR1", "IBM z14 Model ZR1 GA1"), > + CPUDEF_INIT(0x8561, 15, 1, 47, 0x08000000U, "8561", "IBM 8561 GA1"), > + CPUDEF_INIT(0x8562, 15, 1, 47, 0x08000000U, "8562", "IBM 8562 GA1"), > }; Thinking out loud, I know that official names are not published yet. Looking at the recent history (z13, z14), my educated guess would be z15. I guess pretty much everybody would guess that. So I wonder if using "z15" and e.g. "z15ZR1" or "z15s" would be easier to have, although in the end maybe not correct (whatever names smart people are able to come up with, history told us that consistent naming does not seem to be one of the core strengths of CPU marketing people in general). So if we need aliases either way, and z15 is easier to understand than "8561" for any user, what about going forward with that and introducing actual alias later on? We can tag the names to be inofficial. CPUDEF_INIT(0x8561, 15, 1, 47, 0x08000000U, "z15", "Inofficial name for gen15 model 1 GA1 (8561)"), CPUDEF_INIT(0x8562, 15, 1, 47, 0x08000000U, "z15ZR1", "Inofficial name for gen15 model 2 GA1 (8562)"), (please don't eat me alive)
On 26.04.19 13:24, Christian Borntraeger wrote: > > > On 26.04.19 13:21, David Hildenbrand wrote: >> On 26.04.19 13:10, Christian Borntraeger wrote: >>> 8561 and 8562 will be gen15 machines. There is no name yet, lets us use >>> the cpu id as base name. Later on we can provide aliases with the proper >>> name. >>> >>> Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com> >>> --- >>> target/s390x/cpu_models.c | 2 ++ >>> 1 file changed, 2 insertions(+) >>> >>> diff --git a/target/s390x/cpu_models.c b/target/s390x/cpu_models.c >>> index d683635eb5..dd6415103f 100644 >>> --- a/target/s390x/cpu_models.c >>> +++ b/target/s390x/cpu_models.c >>> @@ -83,6 +83,8 @@ static S390CPUDef s390_cpu_defs[] = { >>> CPUDEF_INIT(0x3906, 14, 1, 47, 0x08000000U, "z14", "IBM z14 GA1"), >>> CPUDEF_INIT(0x3906, 14, 2, 47, 0x08000000U, "z14.2", "IBM z14 GA2"), >>> CPUDEF_INIT(0x3907, 14, 1, 47, 0x08000000U, "z14ZR1", "IBM z14 Model ZR1 GA1"), >>> + CPUDEF_INIT(0x8561, 15, 1, 47, 0x08000000U, "8561", "IBM 8561 GA1"), >>> + CPUDEF_INIT(0x8562, 15, 1, 47, 0x08000000U, "8562", "IBM 8562 GA1"), >>> }; >>> >>> #define QEMU_MAX_CPU_TYPE 0x2827 >>> >> >> Can you fixup the comment at the beginning of the definitions, stating >> that base features of new HW are no longer a superset? >> > > Something like > > /* > * CPU definition list in order of release. Up to generation 14 base features> * of a following release have been a superset of the previous release. With > * generation 15 two features have been deprecated. > */ > "one base feature and one optional feature" ?
On 26.04.19 13:32, David Hildenbrand wrote: > On 26.04.19 13:10, Christian Borntraeger wrote: >> 8561 and 8562 will be gen15 machines. There is no name yet, lets us use >> the cpu id as base name. Later on we can provide aliases with the proper >> name. >> >> Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com> >> --- >> target/s390x/cpu_models.c | 2 ++ >> 1 file changed, 2 insertions(+) >> >> diff --git a/target/s390x/cpu_models.c b/target/s390x/cpu_models.c >> index d683635eb5..dd6415103f 100644 >> --- a/target/s390x/cpu_models.c >> +++ b/target/s390x/cpu_models.c >> @@ -83,6 +83,8 @@ static S390CPUDef s390_cpu_defs[] = { >> CPUDEF_INIT(0x3906, 14, 1, 47, 0x08000000U, "z14", "IBM z14 GA1"), >> CPUDEF_INIT(0x3906, 14, 2, 47, 0x08000000U, "z14.2", "IBM z14 GA2"), >> CPUDEF_INIT(0x3907, 14, 1, 47, 0x08000000U, "z14ZR1", "IBM z14 Model ZR1 GA1"), >> + CPUDEF_INIT(0x8561, 15, 1, 47, 0x08000000U, "8561", "IBM 8561 GA1"), >> + CPUDEF_INIT(0x8562, 15, 1, 47, 0x08000000U, "8562", "IBM 8562 GA1"), >> }; > > Thinking out loud, I know that official names are not published yet. > Looking at the recent history (z13, z14), my educated guess would be > z15. I guess pretty much everybody would guess that. Not sure about trademark aspects - especially if this really becomes z15. The smaller machine has no real history (ZR1 vs. s vs BC). So I think I would rather have a correct number than a partially correct name. > > So I wonder if using "z15" and e.g. "z15ZR1" or "z15s" would be easier > to have, although in the end maybe not correct (whatever names smart > people are able to come up with, history told us that consistent naming > does not seem to be one of the core strengths of CPU marketing people in > general). > > So if we need aliases either way, and z15 is easier to understand than > "8561" for any user, what about going forward with that and introducing > actual alias later on? We can tag the names to be inofficial. > > CPUDEF_INIT(0x8561, 15, 1, 47, 0x08000000U, "z15", "Inofficial name for > gen15 model 1 GA1 (8561)"), > CPUDEF_INIT(0x8562, 15, 1, 47, 0x08000000U, "z15ZR1", "Inofficial name > for gen15 model 2 GA1 (8562)"), > > (please don't eat me alive) >
On 26.04.19 13:36, Christian Borntraeger wrote: > > > On 26.04.19 13:32, David Hildenbrand wrote: >> On 26.04.19 13:10, Christian Borntraeger wrote: >>> 8561 and 8562 will be gen15 machines. There is no name yet, lets us use >>> the cpu id as base name. Later on we can provide aliases with the proper >>> name. >>> >>> Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com> >>> --- >>> target/s390x/cpu_models.c | 2 ++ >>> 1 file changed, 2 insertions(+) >>> >>> diff --git a/target/s390x/cpu_models.c b/target/s390x/cpu_models.c >>> index d683635eb5..dd6415103f 100644 >>> --- a/target/s390x/cpu_models.c >>> +++ b/target/s390x/cpu_models.c >>> @@ -83,6 +83,8 @@ static S390CPUDef s390_cpu_defs[] = { >>> CPUDEF_INIT(0x3906, 14, 1, 47, 0x08000000U, "z14", "IBM z14 GA1"), >>> CPUDEF_INIT(0x3906, 14, 2, 47, 0x08000000U, "z14.2", "IBM z14 GA2"), >>> CPUDEF_INIT(0x3907, 14, 1, 47, 0x08000000U, "z14ZR1", "IBM z14 Model ZR1 GA1"), >>> + CPUDEF_INIT(0x8561, 15, 1, 47, 0x08000000U, "8561", "IBM 8561 GA1"), >>> + CPUDEF_INIT(0x8562, 15, 1, 47, 0x08000000U, "8562", "IBM 8562 GA1"), >>> }; >> >> Thinking out loud, I know that official names are not published yet. >> Looking at the recent history (z13, z14), my educated guess would be >> z15. I guess pretty much everybody would guess that. > > Not sure about trademark aspects - especially if this really becomes z15. The smaller > machine has no real history (ZR1 vs. s vs BC). So I think I would rather have a correct > number than a partially correct name. We could also use "gen15a" and "gen15b", still better to get than magic numbers. (yeah well, they are not magic) If you want to stick with numbers, be aware that cpu numbers are not injective, so at some point we will need e.g. "8561.2", just so you're aware of the implications.
On 26.04.19 13:52, David Hildenbrand wrote: > On 26.04.19 13:36, Christian Borntraeger wrote: >> >> >> On 26.04.19 13:32, David Hildenbrand wrote: >>> On 26.04.19 13:10, Christian Borntraeger wrote: >>>> 8561 and 8562 will be gen15 machines. There is no name yet, lets us use >>>> the cpu id as base name. Later on we can provide aliases with the proper >>>> name. >>>> >>>> Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com> >>>> --- >>>> target/s390x/cpu_models.c | 2 ++ >>>> 1 file changed, 2 insertions(+) >>>> >>>> diff --git a/target/s390x/cpu_models.c b/target/s390x/cpu_models.c >>>> index d683635eb5..dd6415103f 100644 >>>> --- a/target/s390x/cpu_models.c >>>> +++ b/target/s390x/cpu_models.c >>>> @@ -83,6 +83,8 @@ static S390CPUDef s390_cpu_defs[] = { >>>> CPUDEF_INIT(0x3906, 14, 1, 47, 0x08000000U, "z14", "IBM z14 GA1"), >>>> CPUDEF_INIT(0x3906, 14, 2, 47, 0x08000000U, "z14.2", "IBM z14 GA2"), >>>> CPUDEF_INIT(0x3907, 14, 1, 47, 0x08000000U, "z14ZR1", "IBM z14 Model ZR1 GA1"), >>>> + CPUDEF_INIT(0x8561, 15, 1, 47, 0x08000000U, "8561", "IBM 8561 GA1"), >>>> + CPUDEF_INIT(0x8562, 15, 1, 47, 0x08000000U, "8562", "IBM 8562 GA1"), >>>> }; >>> >>> Thinking out loud, I know that official names are not published yet. >>> Looking at the recent history (z13, z14), my educated guess would be >>> z15. I guess pretty much everybody would guess that. >> >> Not sure about trademark aspects - especially if this really becomes z15. The smaller >> machine has no real history (ZR1 vs. s vs BC). So I think I would rather have a correct >> number than a partially correct name. > > We could also use "gen15a" and "gen15b", still better to get than magic > numbers. (yeah well, they are not magic) > > If you want to stick with numbers, be aware that cpu numbers are not > injective, so at some point we will need e.g. "8561.2", just so you're > aware of the implications. I think whatever we have here is only used internally for expansion (host-model) and the user will use later the real name when available. (custom). So probably this does not matter for a long time. But I might be wrong. I tend to prefer gen15 over z15 but 856x has also its charm.
On 26.04.19 14:01, Christian Borntraeger wrote: > > > On 26.04.19 13:52, David Hildenbrand wrote: >> On 26.04.19 13:36, Christian Borntraeger wrote: >>> >>> >>> On 26.04.19 13:32, David Hildenbrand wrote: >>>> On 26.04.19 13:10, Christian Borntraeger wrote: >>>>> 8561 and 8562 will be gen15 machines. There is no name yet, lets us use >>>>> the cpu id as base name. Later on we can provide aliases with the proper >>>>> name. >>>>> >>>>> Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com> >>>>> --- >>>>> target/s390x/cpu_models.c | 2 ++ >>>>> 1 file changed, 2 insertions(+) >>>>> >>>>> diff --git a/target/s390x/cpu_models.c b/target/s390x/cpu_models.c >>>>> index d683635eb5..dd6415103f 100644 >>>>> --- a/target/s390x/cpu_models.c >>>>> +++ b/target/s390x/cpu_models.c >>>>> @@ -83,6 +83,8 @@ static S390CPUDef s390_cpu_defs[] = { >>>>> CPUDEF_INIT(0x3906, 14, 1, 47, 0x08000000U, "z14", "IBM z14 GA1"), >>>>> CPUDEF_INIT(0x3906, 14, 2, 47, 0x08000000U, "z14.2", "IBM z14 GA2"), >>>>> CPUDEF_INIT(0x3907, 14, 1, 47, 0x08000000U, "z14ZR1", "IBM z14 Model ZR1 GA1"), >>>>> + CPUDEF_INIT(0x8561, 15, 1, 47, 0x08000000U, "8561", "IBM 8561 GA1"), >>>>> + CPUDEF_INIT(0x8562, 15, 1, 47, 0x08000000U, "8562", "IBM 8562 GA1"), >>>>> }; >>>> >>>> Thinking out loud, I know that official names are not published yet. >>>> Looking at the recent history (z13, z14), my educated guess would be >>>> z15. I guess pretty much everybody would guess that. >>> >>> Not sure about trademark aspects - especially if this really becomes z15. The smaller >>> machine has no real history (ZR1 vs. s vs BC). So I think I would rather have a correct >>> number than a partially correct name. >> >> We could also use "gen15a" and "gen15b", still better to get than magic >> numbers. (yeah well, they are not magic) >> >> If you want to stick with numbers, be aware that cpu numbers are not >> injective, so at some point we will need e.g. "8561.2", just so you're >> aware of the implications. > > I think whatever we have here is only used internally for expansion (host-model) > and the user will use later the real name when available. (custom). So probably this > does not matter for a long time. But I might be wrong. > I tend to prefer gen15 over z15 but 856x has also its charm. > Another question would be, can we rename before the next QEMU release, so it will never be officially part of QEMU? Then we don't need aliases after all. Of course, distros have to take care of that as well, but that shouldn't be upstream QEMUs business.
On Fri, 26 Apr 2019 14:05:30 +0200 David Hildenbrand <david@redhat.com> wrote: > On 26.04.19 14:01, Christian Borntraeger wrote: > > > > > > On 26.04.19 13:52, David Hildenbrand wrote: > >> On 26.04.19 13:36, Christian Borntraeger wrote: > >>> > >>> > >>> On 26.04.19 13:32, David Hildenbrand wrote: > >>>> On 26.04.19 13:10, Christian Borntraeger wrote: > >>>>> 8561 and 8562 will be gen15 machines. There is no name yet, lets us use > >>>>> the cpu id as base name. Later on we can provide aliases with the proper > >>>>> name. > >>>>> > >>>>> Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com> > >>>>> --- > >>>>> target/s390x/cpu_models.c | 2 ++ > >>>>> 1 file changed, 2 insertions(+) > >>>>> > >>>>> diff --git a/target/s390x/cpu_models.c b/target/s390x/cpu_models.c > >>>>> index d683635eb5..dd6415103f 100644 > >>>>> --- a/target/s390x/cpu_models.c > >>>>> +++ b/target/s390x/cpu_models.c > >>>>> @@ -83,6 +83,8 @@ static S390CPUDef s390_cpu_defs[] = { > >>>>> CPUDEF_INIT(0x3906, 14, 1, 47, 0x08000000U, "z14", "IBM z14 GA1"), > >>>>> CPUDEF_INIT(0x3906, 14, 2, 47, 0x08000000U, "z14.2", "IBM z14 GA2"), > >>>>> CPUDEF_INIT(0x3907, 14, 1, 47, 0x08000000U, "z14ZR1", "IBM z14 Model ZR1 GA1"), > >>>>> + CPUDEF_INIT(0x8561, 15, 1, 47, 0x08000000U, "8561", "IBM 8561 GA1"), > >>>>> + CPUDEF_INIT(0x8562, 15, 1, 47, 0x08000000U, "8562", "IBM 8562 GA1"), > >>>>> }; > >>>> > >>>> Thinking out loud, I know that official names are not published yet. > >>>> Looking at the recent history (z13, z14), my educated guess would be > >>>> z15. I guess pretty much everybody would guess that. > >>> > >>> Not sure about trademark aspects - especially if this really becomes z15. The smaller > >>> machine has no real history (ZR1 vs. s vs BC). So I think I would rather have a correct > >>> number than a partially correct name. > >> > >> We could also use "gen15a" and "gen15b", still better to get than magic > >> numbers. (yeah well, they are not magic) > >> > >> If you want to stick with numbers, be aware that cpu numbers are not > >> injective, so at some point we will need e.g. "8561.2", just so you're > >> aware of the implications. > > > > I think whatever we have here is only used internally for expansion (host-model) > > and the user will use later the real name when available. (custom). So probably this > > does not matter for a long time. But I might be wrong. > > I tend to prefer gen15 over z15 but 856x has also its charm. FWIW, I'd prefer gen15 over 856x, but... > > > > Another question would be, can we rename before the next QEMU release, > so it will never be officially part of QEMU? Then we don't need aliases > after all. Of course, distros have to take care of that as well, but > that shouldn't be upstream QEMUs business. ...if we could do that, I'd like that even better.
On 26.04.19 14:55, Cornelia Huck wrote: > On Fri, 26 Apr 2019 14:05:30 +0200 > David Hildenbrand <david@redhat.com> wrote: > >> On 26.04.19 14:01, Christian Borntraeger wrote: >>> >>> >>> On 26.04.19 13:52, David Hildenbrand wrote: >>>> On 26.04.19 13:36, Christian Borntraeger wrote: >>>>> >>>>> >>>>> On 26.04.19 13:32, David Hildenbrand wrote: >>>>>> On 26.04.19 13:10, Christian Borntraeger wrote: >>>>>>> 8561 and 8562 will be gen15 machines. There is no name yet, lets us use >>>>>>> the cpu id as base name. Later on we can provide aliases with the proper >>>>>>> name. >>>>>>> >>>>>>> Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com> >>>>>>> --- >>>>>>> target/s390x/cpu_models.c | 2 ++ >>>>>>> 1 file changed, 2 insertions(+) >>>>>>> >>>>>>> diff --git a/target/s390x/cpu_models.c b/target/s390x/cpu_models.c >>>>>>> index d683635eb5..dd6415103f 100644 >>>>>>> --- a/target/s390x/cpu_models.c >>>>>>> +++ b/target/s390x/cpu_models.c >>>>>>> @@ -83,6 +83,8 @@ static S390CPUDef s390_cpu_defs[] = { >>>>>>> CPUDEF_INIT(0x3906, 14, 1, 47, 0x08000000U, "z14", "IBM z14 GA1"), >>>>>>> CPUDEF_INIT(0x3906, 14, 2, 47, 0x08000000U, "z14.2", "IBM z14 GA2"), >>>>>>> CPUDEF_INIT(0x3907, 14, 1, 47, 0x08000000U, "z14ZR1", "IBM z14 Model ZR1 GA1"), >>>>>>> + CPUDEF_INIT(0x8561, 15, 1, 47, 0x08000000U, "8561", "IBM 8561 GA1"), >>>>>>> + CPUDEF_INIT(0x8562, 15, 1, 47, 0x08000000U, "8562", "IBM 8562 GA1"), >>>>>>> }; >>>>>> >>>>>> Thinking out loud, I know that official names are not published yet. >>>>>> Looking at the recent history (z13, z14), my educated guess would be >>>>>> z15. I guess pretty much everybody would guess that. >>>>> >>>>> Not sure about trademark aspects - especially if this really becomes z15. The smaller >>>>> machine has no real history (ZR1 vs. s vs BC). So I think I would rather have a correct >>>>> number than a partially correct name. >>>> >>>> We could also use "gen15a" and "gen15b", still better to get than magic >>>> numbers. (yeah well, they are not magic) >>>> >>>> If you want to stick with numbers, be aware that cpu numbers are not >>>> injective, so at some point we will need e.g. "8561.2", just so you're >>>> aware of the implications. >>> >>> I think whatever we have here is only used internally for expansion (host-model) >>> and the user will use later the real name when available. (custom). So probably this >>> does not matter for a long time. But I might be wrong. >>> I tend to prefer gen15 over z15 but 856x has also its charm. > > FWIW, I'd prefer gen15 over 856x, but... I actually think that the cpu id would be a nice name for expansion because it is guarenteed to stay and it is unique and it allows to have a different content (if that would be necessary) for 2 gen15 machines. > >>> >> >> Another question would be, can we rename before the next QEMU release, >> so it will never be officially part of QEMU? Then we don't need aliases >> after all. Of course, distros have to take care of that as well, but >> that shouldn't be upstream QEMUs business. > > ...if we could do that, I'd like that even better. I dont think that I know the name in time before the next release. So lets go with gen15a/gen15b or 8561/8562?
On Mon, 29 Apr 2019 09:51:39 +0200 Christian Borntraeger <borntraeger@de.ibm.com> wrote: > On 26.04.19 14:55, Cornelia Huck wrote: > > On Fri, 26 Apr 2019 14:05:30 +0200 > > David Hildenbrand <david@redhat.com> wrote: > > > >> On 26.04.19 14:01, Christian Borntraeger wrote: > >>> > >>> > >>> On 26.04.19 13:52, David Hildenbrand wrote: > >>>> On 26.04.19 13:36, Christian Borntraeger wrote: > >>>>> > >>>>> > >>>>> On 26.04.19 13:32, David Hildenbrand wrote: > >>>>>> On 26.04.19 13:10, Christian Borntraeger wrote: > >>>>>>> 8561 and 8562 will be gen15 machines. There is no name yet, lets us use > >>>>>>> the cpu id as base name. Later on we can provide aliases with the proper > >>>>>>> name. > >>>>>>> > >>>>>>> Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com> > >>>>>>> --- > >>>>>>> target/s390x/cpu_models.c | 2 ++ > >>>>>>> 1 file changed, 2 insertions(+) > >>>>>>> > >>>>>>> diff --git a/target/s390x/cpu_models.c b/target/s390x/cpu_models.c > >>>>>>> index d683635eb5..dd6415103f 100644 > >>>>>>> --- a/target/s390x/cpu_models.c > >>>>>>> +++ b/target/s390x/cpu_models.c > >>>>>>> @@ -83,6 +83,8 @@ static S390CPUDef s390_cpu_defs[] = { > >>>>>>> CPUDEF_INIT(0x3906, 14, 1, 47, 0x08000000U, "z14", "IBM z14 GA1"), > >>>>>>> CPUDEF_INIT(0x3906, 14, 2, 47, 0x08000000U, "z14.2", "IBM z14 GA2"), > >>>>>>> CPUDEF_INIT(0x3907, 14, 1, 47, 0x08000000U, "z14ZR1", "IBM z14 Model ZR1 GA1"), > >>>>>>> + CPUDEF_INIT(0x8561, 15, 1, 47, 0x08000000U, "8561", "IBM 8561 GA1"), > >>>>>>> + CPUDEF_INIT(0x8562, 15, 1, 47, 0x08000000U, "8562", "IBM 8562 GA1"), > >>>>>>> }; > >>>>>> > >>>>>> Thinking out loud, I know that official names are not published yet. > >>>>>> Looking at the recent history (z13, z14), my educated guess would be > >>>>>> z15. I guess pretty much everybody would guess that. > >>>>> > >>>>> Not sure about trademark aspects - especially if this really becomes z15. The smaller > >>>>> machine has no real history (ZR1 vs. s vs BC). So I think I would rather have a correct > >>>>> number than a partially correct name. > >>>> > >>>> We could also use "gen15a" and "gen15b", still better to get than magic > >>>> numbers. (yeah well, they are not magic) > >>>> > >>>> If you want to stick with numbers, be aware that cpu numbers are not > >>>> injective, so at some point we will need e.g. "8561.2", just so you're > >>>> aware of the implications. > >>> > >>> I think whatever we have here is only used internally for expansion (host-model) > >>> and the user will use later the real name when available. (custom). So probably this > >>> does not matter for a long time. But I might be wrong. > >>> I tend to prefer gen15 over z15 but 856x has also its charm. > > > > FWIW, I'd prefer gen15 over 856x, but... > > I actually think that the cpu id would be a nice name for expansion because it is > guarenteed to stay and it is unique and it allows to have a different content (if > that would be necessary) for 2 gen15 machines. > > > >>> > >> > >> Another question would be, can we rename before the next QEMU release, > >> so it will never be officially part of QEMU? Then we don't need aliases > >> after all. Of course, distros have to take care of that as well, but > >> that shouldn't be upstream QEMUs business. > > > > ...if we could do that, I'd like that even better. > > I dont think that I know the name in time before the next release. > So lets go with gen15a/gen15b or 8561/8562? I always have trouble remembering number combinations, so I'd vote for gen15a/gen15b if the official names are not an option.
On 29.04.19 10:53, Cornelia Huck wrote: > On Mon, 29 Apr 2019 09:51:39 +0200 > Christian Borntraeger <borntraeger@de.ibm.com> wrote: > >> On 26.04.19 14:55, Cornelia Huck wrote: >>> On Fri, 26 Apr 2019 14:05:30 +0200 >>> David Hildenbrand <david@redhat.com> wrote: >>> >>>> On 26.04.19 14:01, Christian Borntraeger wrote: >>>>> >>>>> >>>>> On 26.04.19 13:52, David Hildenbrand wrote: >>>>>> On 26.04.19 13:36, Christian Borntraeger wrote: >>>>>>> >>>>>>> >>>>>>> On 26.04.19 13:32, David Hildenbrand wrote: >>>>>>>> On 26.04.19 13:10, Christian Borntraeger wrote: >>>>>>>>> 8561 and 8562 will be gen15 machines. There is no name yet, lets us use >>>>>>>>> the cpu id as base name. Later on we can provide aliases with the proper >>>>>>>>> name. >>>>>>>>> >>>>>>>>> Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com> >>>>>>>>> --- >>>>>>>>> target/s390x/cpu_models.c | 2 ++ >>>>>>>>> 1 file changed, 2 insertions(+) >>>>>>>>> >>>>>>>>> diff --git a/target/s390x/cpu_models.c b/target/s390x/cpu_models.c >>>>>>>>> index d683635eb5..dd6415103f 100644 >>>>>>>>> --- a/target/s390x/cpu_models.c >>>>>>>>> +++ b/target/s390x/cpu_models.c >>>>>>>>> @@ -83,6 +83,8 @@ static S390CPUDef s390_cpu_defs[] = { >>>>>>>>> CPUDEF_INIT(0x3906, 14, 1, 47, 0x08000000U, "z14", "IBM z14 GA1"), >>>>>>>>> CPUDEF_INIT(0x3906, 14, 2, 47, 0x08000000U, "z14.2", "IBM z14 GA2"), >>>>>>>>> CPUDEF_INIT(0x3907, 14, 1, 47, 0x08000000U, "z14ZR1", "IBM z14 Model ZR1 GA1"), >>>>>>>>> + CPUDEF_INIT(0x8561, 15, 1, 47, 0x08000000U, "8561", "IBM 8561 GA1"), >>>>>>>>> + CPUDEF_INIT(0x8562, 15, 1, 47, 0x08000000U, "8562", "IBM 8562 GA1"), >>>>>>>>> }; >>>>>>>> >>>>>>>> Thinking out loud, I know that official names are not published yet. >>>>>>>> Looking at the recent history (z13, z14), my educated guess would be >>>>>>>> z15. I guess pretty much everybody would guess that. >>>>>>> >>>>>>> Not sure about trademark aspects - especially if this really becomes z15. The smaller >>>>>>> machine has no real history (ZR1 vs. s vs BC). So I think I would rather have a correct >>>>>>> number than a partially correct name. >>>>>> >>>>>> We could also use "gen15a" and "gen15b", still better to get than magic >>>>>> numbers. (yeah well, they are not magic) >>>>>> >>>>>> If you want to stick with numbers, be aware that cpu numbers are not >>>>>> injective, so at some point we will need e.g. "8561.2", just so you're >>>>>> aware of the implications. >>>>> >>>>> I think whatever we have here is only used internally for expansion (host-model) >>>>> and the user will use later the real name when available. (custom). So probably this >>>>> does not matter for a long time. But I might be wrong. >>>>> I tend to prefer gen15 over z15 but 856x has also its charm. >>> >>> FWIW, I'd prefer gen15 over 856x, but... >> >> I actually think that the cpu id would be a nice name for expansion because it is >> guarenteed to stay and it is unique and it allows to have a different content (if >> that would be necessary) for 2 gen15 machines. >>> >>>>> >>>> >>>> Another question would be, can we rename before the next QEMU release, >>>> so it will never be officially part of QEMU? Then we don't need aliases >>>> after all. Of course, distros have to take care of that as well, but >>>> that shouldn't be upstream QEMUs business. >>> >>> ...if we could do that, I'd like that even better. >> >> I dont think that I know the name in time before the next release. >> So lets go with gen15a/gen15b or 8561/8562? > > I always have trouble remembering number combinations, so I'd vote for > gen15a/gen15b if the official names are not an option. OK, I will respin with gen15a/gen15b.
> I dont think that I know the name in time before the next release. > So lets go with gen15a/gen15b or 8561/8562? > We can still fixup if the names happen to be known earlier.
diff --git a/target/s390x/cpu_models.c b/target/s390x/cpu_models.c index d683635eb5..dd6415103f 100644 --- a/target/s390x/cpu_models.c +++ b/target/s390x/cpu_models.c @@ -83,6 +83,8 @@ static S390CPUDef s390_cpu_defs[] = { CPUDEF_INIT(0x3906, 14, 1, 47, 0x08000000U, "z14", "IBM z14 GA1"), CPUDEF_INIT(0x3906, 14, 2, 47, 0x08000000U, "z14.2", "IBM z14 GA2"), CPUDEF_INIT(0x3907, 14, 1, 47, 0x08000000U, "z14ZR1", "IBM z14 Model ZR1 GA1"), + CPUDEF_INIT(0x8561, 15, 1, 47, 0x08000000U, "8561", "IBM 8561 GA1"), + CPUDEF_INIT(0x8562, 15, 1, 47, 0x08000000U, "8562", "IBM 8562 GA1"), }; #define QEMU_MAX_CPU_TYPE 0x2827
8561 and 8562 will be gen15 machines. There is no name yet, lets us use the cpu id as base name. Later on we can provide aliases with the proper name. Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com> --- target/s390x/cpu_models.c | 2 ++ 1 file changed, 2 insertions(+)