Re: [PATCH] powerpc/powernv: Enable CPU idle state detection for POWER11

[Christophe Leroy <christophe.leroy@csgroup.eu>]

Le 21/04/2025 à 09:07, Aboorva Devarajan a écrit :
> [Vous ne recevez pas souvent de courriers de aboorvad@linux.ibm.com. Découvrez pourquoi ceci est important à https://aka.ms/LearnAboutSenderIdentification ]
> 
> Extend idle state detection to POWER11 by updating the PVR check.
> This ensures POWER11 correctly recognizes supported stop states,
> similar to POWER9 and POWER10.
> 
> Reviewed-by: Mahesh Salgaonkar <mahesh@linux.ibm.com>
> Signed-off-by: Aboorva Devarajan <aboorvad@linux.ibm.com>
> 
> ---
> 
> Without Patch: (Power11 - PowerNV systems)
> 
> CPUidle driver: powernv_idle
> CPUidle governor: menu
> analyzing CPU 927:
> 
> Number of idle states: 1
> Available idle states: snooze
> snooze:
> Flags/Description: snooze
> Latency: 0
> Usage: 251631
> Duration: 207497715900
> 
> With Patch: (Power11 - PowerNV systems)
> 
> CPUidle driver: powernv_idle
> CPUidle governor: menu
> analyzing CPU 959:
> 
> Number of idle states: 4
> Available idle states: snooze stop0_lite stop0 stop3
> snooze:
> Flags/Description: snooze
> Latency: 0
> Usage: 2
> Duration: 33
> stop0_lite:
> Flags/Description: stop0_lite
> Latency: 1
> Usage: 1
> Duration: 52
> stop0:
> Flags/Description: stop0
> Latency: 10
> Usage: 13
> Duration: 1920
> stop3:
> Flags/Description: stop3
> Latency: 45
> Usage: 381
> Duration: 21638478
> 
> 
>   arch/powerpc/platforms/powernv/idle.c | 8 ++++----
>   1 file changed, 4 insertions(+), 4 deletions(-)
> 
> diff --git a/arch/powerpc/platforms/powernv/idle.c b/arch/powerpc/platforms/powernv/idle.c
> index d98b933e4984..10f4d7d5eb44 100644
> --- a/arch/powerpc/platforms/powernv/idle.c
> +++ b/arch/powerpc/platforms/powernv/idle.c
> @@ -1171,8 +1171,8 @@ static void __init pnv_arch300_idle_init(void)
>          u64 max_residency_ns = 0;
>          int i;
> 
> -       /* stop is not really architected, we only have p9,p10 drivers */
> -       if (!pvr_version_is(PVR_POWER10) && !pvr_version_is(PVR_POWER9))
> +       /* stop is not really architected, we only have p9,p10 and p11 drivers */
> +       if (!(PVR_VER(mfspr(SPRN_PVR)) >= PVR_POWER9))

Isn't this test too permissive ?

#define PVR_POWER9	0x004E
#define PVR_POWER10	0x0080
#define PVR_POWER11	0x0082
#define PVR_BE		0x0070
#define PVR_PA6T	0x0090

#define PVR_VER_E500V1	0x8020
#define PVR_VER_E500V2	0x8021
#define PVR_VER_E500MC	0x8023
#define PVR_VER_E5500	0x8024
#define PVR_VER_E6500	0x8040
#define PVR_VER_7450	0x8000
#define PVR_VER_7455	0x8001
#define PVR_VER_7447	0x8002
#define PVR_VER_7447A	0x8003
#define PVR_VER_7448	0x8004

>                  return;
> 
>          /*
> @@ -1189,8 +1189,8 @@ static void __init pnv_arch300_idle_init(void)
>                  struct pnv_idle_states_t *state = &pnv_idle_states[i];
>                  u64 psscr_rl = state->psscr_val & PSSCR_RL_MASK;
> 
> -               /* No deep loss driver implemented for POWER10 yet */
> -               if (pvr_version_is(PVR_POWER10) &&
> +               /* No deep loss driver implemented for POWER10 and POWER11 yet */
> +               if ((PVR_VER(mfspr(SPRN_PVR)) >= PVR_POWER10) &&
>                                  state->flags & (OPAL_PM_TIMEBASE_STOP|OPAL_PM_LOSE_FULL_CONTEXT))
>                          continue;
> 
> --
> 2.49.0
>