[PATCH 0/1] Fix energy calcultation in RAPL MSR

[PATCH 0/1] Fix energy calcultation in RAPL MSR

[Anthony Harivel]

Hi Paolo,

The value reported by the calculation was looking very wrong to me.

This should fix it for good.

Calculated value appears now way more accurate from what is expected 
from the feature:

with "-smp 4":

[...]
# modprobe intel_rapl_msr.ko
intel_rapl_common: Found RAPL domain package
intel_rapl_common: found 1 domains on package-0
intel_rapl_common: Core CPU package-0:package energy=61035pJ, time=976us, power=125000uW
intel_rapl_common: update package-0 domain package data
# 
# 
# 
# cat /sys/class/powercap/intel-rapl:0/energy_uj 
303160
# cat /sys/class/powercap/intel-rapl:0/energy_uj 
325621
# cat /sys/class/powercap/intel-rapl:0/energy_uj 
325621
# cat /sys/class/powercap/intel-rapl:0/energy_uj 
357237
# cat /sys/class/powercap/intel-rapl:0/energy_uj 
357237
# cat /sys/class/powercap/intel-rapl:0/energy_uj 
388792
[...]

Tested with several SMP architecture.
(i.e -smp 4 / -smp 8,sockets=2,dies=1 / -smp 8,sockets=4,dies=2 )

Anthony Harivel (1):
  target/i386: Fix arguments for vmsr_read_thread_stat()

 target/i386/kvm/kvm.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

-- 
2.45.2

[PATCH 1/1] target/i386: Fix arguments for vmsr_read_thread_stat()

[Anthony Harivel]

Snapshot of the stat utime and stime for each thread, taken before and
after the pause, must be stored in separate locations

Signed-off-by: Anthony Harivel <aharivel@redhat.com>
---
 target/i386/kvm/kvm.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/target/i386/kvm/kvm.c b/target/i386/kvm/kvm.c
index 31f149c9902c..2fa88ef1e373 100644
--- a/target/i386/kvm/kvm.c
+++ b/target/i386/kvm/kvm.c
@@ -2712,8 +2712,8 @@ static void *kvm_msr_energy_thread(void *data)
             thd_stat[i].thread_id = thread_ids[i];
             vmsr_read_thread_stat(vmsr->pid,
                                   thd_stat[i].thread_id,
-                                  thd_stat[i].utime,
-                                  thd_stat[i].stime,
+                                  &thd_stat[i].utime[0],
+                                  &thd_stat[i].stime[0],
                                   &thd_stat[i].cpu_id);
             thd_stat[i].pkg_id =
                 vmsr_get_physical_package_id(thd_stat[i].cpu_id);
@@ -2777,8 +2777,8 @@ static void *kvm_msr_energy_thread(void *data)
         for (int i = 0; i < num_threads; i++) {
             vmsr_read_thread_stat(vmsr->pid,
                                   thd_stat[i].thread_id,
-                                  thd_stat[i].utime,
-                                  thd_stat[i].stime,
+                                  &thd_stat[i].utime[1],
+                                  &thd_stat[i].stime[1],
                                   &thd_stat[i].cpu_id);
 
             if (vmsr->pid < 0) {
-- 
2.45.2

Re: [PATCH 1/1] target/i386: Fix arguments for vmsr_read_thread_stat()

[Zhao Liu]

On Wed, Aug 07, 2024 at 02:43:20PM +0200, Anthony Harivel wrote:
> Date: Wed,  7 Aug 2024 14:43:20 +0200
> From: Anthony Harivel <aharivel@redhat.com>
> Subject: [PATCH 1/1] target/i386: Fix arguments for vmsr_read_thread_stat()
> 
> Snapshot of the stat utime and stime for each thread, taken before and
> after the pause, must be stored in separate locations
> 
> Signed-off-by: Anthony Harivel <aharivel@redhat.com>
> ---
>  target/i386/kvm/kvm.c | 8 ++++----
>  1 file changed, 4 insertions(+), 4 deletions(-)

Reviewed-by: Zhao Liu <zhao1.liu@intel.com>