Re: [PATCH v2 0/5] Re-write PPC64 PMU instruction count using TCG Ops
On 1/3/22 03:46, Cédric Le Goater wrote: On 12/23/21 21:18, Daniel Henrique Barboza wrote: Hi, In this version the tedious repetition was taken away from the code by using a helper that increments the PMCs based on specified conditions. As far as Avocado test goes, the performance is the same as the previous version. All PMU-EBB kernel selftests are also passing. Basically we have the same benefits from v1 but 150+ lines shorter. changes from v1: - former patches 2-6: removed - new patch 2: * added inc_spr_if_cond() helper - new patch 3: * add insn count for PMCs 1-4 - patch 4 (former 7): * use a loop to reduce code repetition when checking for counter overflows - v1 link: https://lists.gnu.org/archive/html/qemu-devel/2021-12/msg03871.html I understand that you are going to rebase on top of Richard's patchset. and so this series is now obsolete ? Yes it is. I'll post a new version of Richard's patchset shortly. Daniel Thanks, C.
Re: [PATCH v2 0/5] Re-write PPC64 PMU instruction count using TCG Ops
On 12/23/21 21:18, Daniel Henrique Barboza wrote: Hi, In this version the tedious repetition was taken away from the code by using a helper that increments the PMCs based on specified conditions. As far as Avocado test goes, the performance is the same as the previous version. All PMU-EBB kernel selftests are also passing. Basically we have the same benefits from v1 but 150+ lines shorter. changes from v1: - former patches 2-6: removed - new patch 2: * added inc_spr_if_cond() helper - new patch 3: * add insn count for PMCs 1-4 - patch 4 (former 7): * use a loop to reduce code repetition when checking for counter overflows - v1 link: https://lists.gnu.org/archive/html/qemu-devel/2021-12/msg03871.html I understand that you are going to rebase on top of Richard's patchset. and so this series is now obsolete ? Thanks, C.
[PATCH v2 0/5] Re-write PPC64 PMU instruction count using TCG Ops
Hi, In this version the tedious repetition was taken away from the code by using a helper that increments the PMCs based on specified conditions. As far as Avocado test goes, the performance is the same as the previous version. All PMU-EBB kernel selftests are also passing. Basically we have the same benefits from v1 but 150+ lines shorter. changes from v1: - former patches 2-6: removed - new patch 2: * added inc_spr_if_cond() helper - new patch 3: * add insn count for PMCs 1-4 - patch 4 (former 7): * use a loop to reduce code repetition when checking for counter overflows - v1 link: https://lists.gnu.org/archive/html/qemu-devel/2021-12/msg03871.html Daniel Henrique Barboza (5): target/ppc: introduce power8-pmu-insn-cnt.c.inc target/ppc/power8-pmu-insn-cnt: introduce inc_spr_if_cond() target/ppc/power8-pmu-insn-cnt: add PMCs1-4 insn count target/ppc/power8-pmu-insn-cnt: add pmu_check_overflow() target/ppc/power8-pmu.c: remove helper_insns_inc() target/ppc/helper.h | 2 +- target/ppc/power8-pmu-insn-cnt.c.inc | 213 +++ target/ppc/power8-pmu.c | 60 +--- target/ppc/translate.c | 44 +- 4 files changed, 220 insertions(+), 99 deletions(-) create mode 100644 target/ppc/power8-pmu-insn-cnt.c.inc -- 2.33.1
