[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [PATCH v9 09/10] target/ppc: PMU Event-Based exception support
From: |
Cédric Le Goater |
Subject: |
Re: [PATCH v9 09/10] target/ppc: PMU Event-Based exception support |
Date: |
Fri, 3 Dec 2021 10:27:44 +0100 |
User-agent: |
Mozilla/5.0 (X11; Linux x86_64; rv:91.0) Gecko/20100101 Thunderbird/91.3.0 |
Hello,
On 12/1/21 16:17, Daniel Henrique Barboza wrote:
From: Gustavo Romero <gromero@linux.ibm.com>
Following up the rfebb implementation, this patch adds the EBB exception
support that are triggered by Performance Monitor alerts. This exception
occurs when an enabled PMU condition or event happens and both MMCR0_EBE
and BESCR_PME are set.
The supported PM alerts will consist of counter negative conditions of
the PMU counters. This will be achieved by a timer mechanism that will
predict when a counter becomes negative. The PMU timer callback will set
the appropriate bits in MMCR0 and fire a PMC interrupt. The EBB
exception code will then set the appropriate BESCR bits, set the next
instruction pointer to the address pointed by the return register
(SPR_EBBRR), and redirect execution to the handler (pointed by
SPR_EBBHR).
CC: Gustavo Romero <gustavo.romero@linaro.org>
Signed-off-by: Gustavo Romero <gromero@linux.ibm.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
---
target/ppc/cpu.h | 5 ++++-
target/ppc/excp_helper.c | 29 +++++++++++++++++++++++++++++
target/ppc/power8-pmu.c | 40 ++++++++++++++++++++++++++++++++++++++--
3 files changed, 71 insertions(+), 3 deletions(-)
diff --git a/target/ppc/cpu.h b/target/ppc/cpu.h
index 741b8baf4c..8e0e6319ee 100644
--- a/target/ppc/cpu.h
+++ b/target/ppc/cpu.h
@@ -129,8 +129,10 @@ enum {
/* ISA 3.00 additions */
POWERPC_EXCP_HVIRT = 101,
POWERPC_EXCP_SYSCALL_VECTORED = 102, /* scv exception
*/
+ POWERPC_EXCP_EBB = 103, /* Event-based branch exception */
+
/* EOL
*/
- POWERPC_EXCP_NB = 103,
+ POWERPC_EXCP_NB = 104,
/* QEMU exceptions: special cases we want to stop translation
*/
POWERPC_EXCP_SYSCALL_USER = 0x203, /* System call in user mode only
*/
};
@@ -2452,6 +2454,7 @@ enum {
PPC_INTERRUPT_HMI, /* Hypervisor Maintenance interrupt */
PPC_INTERRUPT_HDOORBELL, /* Hypervisor Doorbell interrupt */
PPC_INTERRUPT_HVIRT, /* Hypervisor virtualization interrupt */
+ PPC_INTERRUPT_PMC, /* PMU interrupt */
This is referred as a Performance Monitor event-based exception in ISA.
PPC_INTERRUPT_PM_EBB would be a better name I think.
We also have :
Load Monitored event-based exceptions
External event-based exceptions
};
/* Processor Compatibility mask (PCR) */
diff --git a/target/ppc/excp_helper.c b/target/ppc/excp_helper.c
index 7ead32279c..a26d266fe6 100644
--- a/target/ppc/excp_helper.c
+++ b/target/ppc/excp_helper.c
@@ -799,6 +799,23 @@ static inline void powerpc_excp(PowerPCCPU *cpu, int
excp_model, int excp)
cpu_abort(cs, "Non maskable external exception "
"is not implemented yet !\n");
break;
+ case POWERPC_EXCP_EBB: /* Event-based branch exception */
+ if ((env->spr[SPR_FSCR] & (1ull << FSCR_EBB)) &&
+ (env->spr[SPR_BESCR] & BESCR_GE) &&
+ (env->spr[SPR_BESCR] & BESCR_PME)) {
Shouldn't we check these conditions to gate the EBB before sending it ?
What about the HFSCR_EBB case ?
+ target_ulong nip;
+
+ env->spr[SPR_BESCR] &= ~BESCR_GE; /* Clear GE */
+ env->spr[SPR_BESCR] |= BESCR_PMEO; /* Set PMEO */
+ env->spr[SPR_EBBRR] = env->nip; /* Save NIP for rfebb insn */
+ nip = env->spr[SPR_EBBHR]; /* EBB handler */
+ powerpc_set_excp_state(cpu, nip, env->msr);
+ }
I am in favor of not adding anything more under the powerpc_excp() routine.
Can we add a helper to isolate the EBB case ? I have the feeling it could
become big.
+ /*
+ * This interrupt is handled by userspace. No need
+ * to proceed.
+ */
+ return;
default:
excp_invalid:
cpu_abort(cs, "Invalid PowerPC exception %d. Aborting\n", excp);
@@ -1046,6 +1063,18 @@ static void ppc_hw_interrupt(CPUPPCState *env)
powerpc_excp(cpu, env->excp_model, POWERPC_EXCP_THERM);
return;
}
+ /* PMC -> Event-based branch exception */
+ if (env->pending_interrupts & (1 << PPC_INTERRUPT_PMC)) {
+ /*
+ * Performance Monitor event-based exception can only
+ * occur in problem state.
+ */
Yes. is the PM EBB exception gated by EE also ?
Thanks,
C.
+ if (msr_pr == 1) {
+ env->pending_interrupts &= ~(1 << PPC_INTERRUPT_PMC);
+ powerpc_excp(cpu, env->excp_model, POWERPC_EXCP_EBB);
+ return;
+ }
+ }
}
if (env->resume_as_sreset) {
diff --git a/target/ppc/power8-pmu.c b/target/ppc/power8-pmu.c
index 08d1902cd5..279b824c3f 100644
--- a/target/ppc/power8-pmu.c
+++ b/target/ppc/power8-pmu.c
@@ -297,6 +297,20 @@ void helper_store_pmc(CPUPPCState *env, uint32_t sprn,
uint64_t value)
pmc_update_overflow_timer(env, sprn);
}
+static void pmu_delete_timers(CPUPPCState *env)
+{
+ QEMUTimer *pmc_overflow_timer;
+ int sprn;
+
+ for (sprn = SPR_POWER_PMC1; sprn <= SPR_POWER_PMC6; sprn++) {
+ pmc_overflow_timer = get_cyc_overflow_timer(env, sprn);
+
+ if (pmc_overflow_timer) {
+ timer_del(pmc_overflow_timer);
+ }
+ }
+}
+
static void fire_PMC_interrupt(PowerPCCPU *cpu)
{
CPUPPCState *env = &cpu->env;
@@ -305,8 +319,30 @@ static void fire_PMC_interrupt(PowerPCCPU *cpu)
return;
}
- /* PMC interrupt not implemented yet */
- return;
+ pmu_update_cycles(env);
+
+ if (env->spr[SPR_POWER_MMCR0] & MMCR0_FCECE) {
+ env->spr[SPR_POWER_MMCR0] &= ~MMCR0_FCECE;
+ env->spr[SPR_POWER_MMCR0] |= MMCR0_FC;
+
+ /* Changing MMCR0_FC demands a new hflags compute */
+ hreg_compute_hflags(env);
+
+ /*
+ * Delete all pending timers if we need to freeze
+ * the PMC. We'll restart them when the PMC starts
+ * running again.
+ */
+ pmu_delete_timers(env);
+ }
+
+ if (env->spr[SPR_POWER_MMCR0] & MMCR0_PMAE) {
+ env->spr[SPR_POWER_MMCR0] &= ~MMCR0_PMAE;
+ env->spr[SPR_POWER_MMCR0] |= MMCR0_PMAO;
+ }
+
+ /* Fire the PMC hardware exception */
+ ppc_set_irq(cpu, PPC_INTERRUPT_PMC, 1);
}
/* This helper assumes that the PMC is running. */
- [PATCH v9 01/10] target/ppc: introduce PMUEventType and PMU overflow timers, (continued)
- [PATCH v9 01/10] target/ppc: introduce PMUEventType and PMU overflow timers, Daniel Henrique Barboza, 2021/12/01
- [PATCH v9 04/10] target/ppc: PMU: update counters on MMCR1 write, Daniel Henrique Barboza, 2021/12/01
- [PATCH v9 03/10] target/ppc: PMU: update counters on PMCs r/w, Daniel Henrique Barboza, 2021/12/01
- [PATCH v9 02/10] target/ppc: PMU basic cycle count for pseries TCG, Daniel Henrique Barboza, 2021/12/01
- [PATCH v9 05/10] target/ppc: enable PMU counter overflow with cycle events, Daniel Henrique Barboza, 2021/12/01
- [PATCH v9 06/10] target/ppc: enable PMU instruction count, Daniel Henrique Barboza, 2021/12/01
- [PATCH v9 07/10] target/ppc/power8-pmu.c: add PM_RUN_INST_CMPL (0xFA) event, Daniel Henrique Barboza, 2021/12/01
- [PATCH v9 08/10] PPC64/TCG: Implement 'rfebb' instruction, Daniel Henrique Barboza, 2021/12/01
- [PATCH v9 09/10] target/ppc: PMU Event-Based exception support, Daniel Henrique Barboza, 2021/12/01
- [PATCH v9 10/10] target/ppc/excp_helper.c: EBB handling adjustments, Daniel Henrique Barboza, 2021/12/01
- Re: [PATCH v9 00/10] PMU-EBB support for PPC64 TCG, Cédric Le Goater, 2021/12/03
- Re: [PATCH v9 00/10] PMU-EBB support for PPC64 TCG, Cédric Le Goater, 2021/12/15