[libvirt] [PATCH V1 0/2] cpu: Handle only high order 16 bits of PVR for IBM
Handle only high order 16 bits of PVR for IBM Power processors Pradipta Kr. Banerjee (2): cpu: Handle only high order 16 bits of PVR for IBM Power processors cpu_map.xml: Add base POWER8 processor details src/cpu/cpu_map.xml | 5 +++++ src/cpu/cpu_powerpc.c | 7 ++++++- 2 files changed, 11 insertions(+), 1 deletion(-) -- 1.9.3
IBM Power processors encode PVR as CPU family in higher 16 bits and a CPU version in lower 16 bits. Since there is no significant change in behavior between versions, there is no point to add every single CPU version in cpu_map.xml and check for the same in cpu_powerpc.c Signed-off-by: Pradipta Kr. Banerjee <bpradip@in.ibm.com> --- src/cpu/cpu_powerpc.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) diff --git a/src/cpu/cpu_powerpc.c b/src/cpu/cpu_powerpc.c index 67cb9ff..18dbc99 100644 --- a/src/cpu/cpu_powerpc.c +++ b/src/cpu/cpu_powerpc.c @@ -93,7 +93,12 @@ ppcModelFindPVR(const struct ppc_map *map, model = map->models; while (model != NULL) { - if (model->data.pvr == pvr) + /*IBM PowerPC processors encode PVR as CPU family in higher 16 bits and + *a CPU version in lower 16 bits. Since there is no significant change + *in behavior between versions, there is no point to add every single + *CPU version in cpu_map.xml + */ + if ((model->data.pvr & 0xFFFF0000) == (pvr & 0xFFFF0000)) return model; model = model->next; -- 1.9.3
On Tue, Aug 19, 2014 at 11:00:12PM +0530, Pradipta Kr. Banerjee wrote:
IBM Power processors encode PVR as CPU family in higher 16 bits and a CPU version in lower 16 bits. Since there is no significant change in behavior between versions, there is no point to add every single CPU version in cpu_map.xml and check for the same in cpu_powerpc.c
Just an idea, but what if we select the model based on the whole pvr and only in case none is found, we will fallback to the first one that has the same high 16 bits? When we have already power8_v1.0 and you're adding power8, we will never select the second one in the way you have implemented it.
Signed-off-by: Pradipta Kr. Banerjee <bpradip@in.ibm.com> --- src/cpu/cpu_powerpc.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-)
diff --git a/src/cpu/cpu_powerpc.c b/src/cpu/cpu_powerpc.c index 67cb9ff..18dbc99 100644 --- a/src/cpu/cpu_powerpc.c +++ b/src/cpu/cpu_powerpc.c @@ -93,7 +93,12 @@ ppcModelFindPVR(const struct ppc_map *map,
model = map->models; while (model != NULL) { - if (model->data.pvr == pvr) + /*IBM PowerPC processors encode PVR as CPU family in higher 16 bits and + *a CPU version in lower 16 bits. Since there is no significant change + *in behavior between versions, there is no point to add every single + *CPU version in cpu_map.xml + */ + if ((model->data.pvr & 0xFFFF0000) == (pvr & 0xFFFF0000)) return model;
model = model->next; -- 1.9.3
-- libvir-list mailing list libvir-list@redhat.com https://www.redhat.com/mailman/listinfo/libvir-list
On 08/20/2014 09:02 AM, Martin Kletzander wrote:
On Tue, Aug 19, 2014 at 11:00:12PM +0530, Pradipta Kr. Banerjee wrote:
IBM Power processors encode PVR as CPU family in higher 16 bits and a CPU version in lower 16 bits. Since there is no significant change in behavior between versions, there is no point to add every single CPU version in cpu_map.xml and check for the same in cpu_powerpc.c
Just an idea, but what if we select the model based on the whole pvr and only in case none is found, we will fallback to the first one that has the same high 16 bits? When we have already power8_v1.0 and you're adding power8, we will never select the second one in the way you have implemented it. Actually none of the model_vX.X are required. Only base processor versions are required. Should I send a V2 with those changes ?
Additionally PowerKVM doesn't support specifying cpu model for guests (qemu cmdline: -cpu <MODEL>). The only way possible is to use qemu compat mode (qemu cmdline: -cpu host,compat=model) support of which is yet to be added in libvirt. The only supported models in compat modes are power6,power7,power8. There was an RFC posted on the same - https://www.redhat.com/archives/libvir-list/2014-June/msg01338.html Please let me know your thoughts.
Signed-off-by: Pradipta Kr. Banerjee <bpradip@in.ibm.com> --- src/cpu/cpu_powerpc.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-)
diff --git a/src/cpu/cpu_powerpc.c b/src/cpu/cpu_powerpc.c index 67cb9ff..18dbc99 100644 --- a/src/cpu/cpu_powerpc.c +++ b/src/cpu/cpu_powerpc.c @@ -93,7 +93,12 @@ ppcModelFindPVR(const struct ppc_map *map,
model = map->models; while (model != NULL) { - if (model->data.pvr == pvr) + /*IBM PowerPC processors encode PVR as CPU family in higher 16 bits and + *a CPU version in lower 16 bits. Since there is no significant change + *in behavior between versions, there is no point to add every single + *CPU version in cpu_map.xml + */ + if ((model->data.pvr & 0xFFFF0000) == (pvr & 0xFFFF0000)) return model;
model = model->next; -- 1.9.3
-- libvir-list mailing list libvir-list@redhat.com https://www.redhat.com/mailman/listinfo/libvir-list
-- libvir-list mailing list libvir-list@redhat.com https://www.redhat.com/mailman/listinfo/libvir-list
-- Regards, Pradipta Kumar B(bpradip@in.ibm.com) IBM Systems & Technology Labs, India.
Add base POWER8 processor details Signed-off-by: Pradipta Kr. Banerjee <bpradip@in.ibm.com> --- src/cpu/cpu_map.xml | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/src/cpu/cpu_map.xml b/src/cpu/cpu_map.xml index 12987a0..d75a92e 100644 --- a/src/cpu/cpu_map.xml +++ b/src/cpu/cpu_map.xml @@ -622,6 +622,11 @@ <pvr value='0x004a0201'/> </model> + <model name='POWER8'> + <vendor name='IBM'/> + <pvr value='0x004b0000'/> + </model> + <model name='POWER8_v1.0'> <vendor name='IBM'/> <pvr value='0x004b0100'/> -- 1.9.3
participants (3)
-
Martin Kletzander -
Pradipta Kr. Banerjee -
Pradipta Kumar Banerjee