From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mx0a-001b2d01.pphosted.com (mx0b-001b2d01.pphosted.com [148.163.158.5]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by lists.ozlabs.org (Postfix) with ESMTPS id 3yFVj93sfJzDrFp for ; Mon, 16 Oct 2017 05:43:56 +1100 (AEDT) Received: from pps.filterd (m0098419.ppops.net [127.0.0.1]) by mx0b-001b2d01.pphosted.com (8.16.0.21/8.16.0.21) with SMTP id v9FIhoO9102711 for ; Sun, 15 Oct 2017 14:43:54 -0400 Received: from e06smtp11.uk.ibm.com (e06smtp11.uk.ibm.com [195.75.94.107]) by mx0b-001b2d01.pphosted.com with ESMTP id 2dm6k9k7h3-1 (version=TLSv1.2 cipher=AES256-SHA bits=256 verify=NOT) for ; Sun, 15 Oct 2017 14:43:53 -0400 Received: from localhost by e06smtp11.uk.ibm.com with IBM ESMTP SMTP Gateway: Authorized Use Only! Violators will be prosecuted for from ; Sun, 15 Oct 2017 19:43:52 +0100 Received: from d23av05.au.ibm.com (d23av05.au.ibm.com [9.190.234.119]) by b06cxnps4076.portsmouth.uk.ibm.com (8.14.9/8.14.9/NCO v10.0) with ESMTP id v9FIhmau17563896 for ; Sun, 15 Oct 2017 18:43:49 GMT Received: from d23av05.au.ibm.com (localhost [127.0.0.1]) by d23av05.au.ibm.com (8.14.4/8.14.4/NCO v10.0 AVout) with ESMTP id v9FIhlqJ031864 for ; Mon, 16 Oct 2017 05:43:48 +1100 From: Madhavan Srinivasan To: mpe@ellerman.id.au Cc: linuxppc-dev@lists.ozlabs.org, Madhavan Srinivasan Subject: [PATCH 1/2] powerpc/perf/imc: use cpu_to_node instead topology_physical_package_id Date: Mon, 16 Oct 2017 00:13:41 +0530 Message-Id: <1508093022-11496-1-git-send-email-maddy@linux.vnet.ibm.com> List-Id: Linux on PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , ... [1069001518.000000] [c000003f95b3f770] [c0000000000b2574] init_imc_pmu+0x1f4/0xc40 [1069005374.000000] [c000003f95b3f850] [c00000000008fec8] opal_imc_counters_probe+0x2e8/0x3e0 [1069009426.000000] [c000003f95b3f950] [c0000000006153a4] platform_drv_probe+0x44/0x90 [1069012818.000000] [c000003f95b3f9c0] [c0000000006124c0] really_probe+0x290/0x370 [1069016302.000000] [c000003f95b3fa50] [c0000000006126c8] __driver_attach+0x128/0x130 [1069019564.000000] [c000003f95b3fa90] [c00000000060f38c] bus_for_each_dev+0x9c/0x110 [1069022838.000000] [c000003f95b3fae0] [c000000000611dfc] driver_attach+0x3c/0x60 [1069026104.000000] [c000003f95b3fb10] [c0000000006118d8] bus_add_driver+0x298/0x320 [1069029428.000000] [c000003f95b3fb90] [c0000000006135b8] driver_register+0xb8/0x1a0 [1069033016.000000] [c000003f95b3fc00] [c00000000061533c] __platform_driver_register+0x8c/0xb0 [1069036362.000000] [c000003f95b3fc30] [c000000000cbb0fc] opal_imc_driver_init+0x24/0x38 [1069039756.000000] [c000003f95b3fc50] [c00000000000cc70] do_one_initcall+0xd0/0x250 [1069043094.000000] [c000003f95b3fd20] [c000000000ca44d0] kernel_init_freeable+0x244/0x324 [1069046490.000000] [c000003f95b3fdc0] [c00000000000d600] kernel_init+0x30/0x1b0 [1069050348.000000] [c000003f95b3fe30] [c00000000000b268] ret_from_kernel_thread+0x5c/0x74 init_imc_pmu() use topology_physical_package_id() to detect the phy_id of the processor it is on to get local memory. But this cause crashes when node_id are not same as physicaly id. As a fix use cpu_to_node(). Reported-By: Rob Lippert Tested-By: Madhavan Srinivasan Signed-off-by: Madhavan Srinivasan --- arch/powerpc/perf/imc-pmu.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/arch/powerpc/perf/imc-pmu.c b/arch/powerpc/perf/imc-pmu.c index 88126245881b..88d1415e2547 100644 --- a/arch/powerpc/perf/imc-pmu.c +++ b/arch/powerpc/perf/imc-pmu.c @@ -467,7 +467,7 @@ static int nest_imc_event_init(struct perf_event *event) * Nest HW counter memory resides in a per-chip reserve-memory (HOMER). * Get the base memory addresss for this cpu. */ - chip_id = topology_physical_package_id(event->cpu); + chip_id = cpu_to_chip_id(event->cpu); pcni = pmu->mem_info; do { if (pcni->id == chip_id) { @@ -524,19 +524,19 @@ static int nest_imc_event_init(struct perf_event *event) */ static int core_imc_mem_init(int cpu, int size) { - int phys_id, rc = 0, core_id = (cpu / threads_per_core); + int nid, rc = 0, core_id = (cpu / threads_per_core); struct imc_mem_info *mem_info; /* * alloc_pages_node() will allocate memory for core in the * local node only. */ - phys_id = topology_physical_package_id(cpu); + nid = cpu_to_node(cpu); mem_info = &core_imc_pmu->mem_info[core_id]; mem_info->id = core_id; /* We need only vbase for core counters */ - mem_info->vbase = page_address(alloc_pages_node(phys_id, + mem_info->vbase = page_address(alloc_pages_node(nid, GFP_KERNEL | __GFP_ZERO | __GFP_THISNODE | __GFP_NOWARN, get_order(size))); if (!mem_info->vbase) @@ -783,14 +783,14 @@ static int core_imc_event_init(struct perf_event *event) static int thread_imc_mem_alloc(int cpu_id, int size) { u64 ldbar_value, *local_mem = per_cpu(thread_imc_mem, cpu_id); - int phys_id = topology_physical_package_id(cpu_id); + int nid = cpu_to_node(cpu_id); if (!local_mem) { /* * This case could happen only once at start, since we dont * free the memory in cpu offline path. */ - local_mem = page_address(alloc_pages_node(phys_id, + local_mem = page_address(alloc_pages_node(nid, GFP_KERNEL | __GFP_ZERO | __GFP_THISNODE | __GFP_NOWARN, get_order(size))); if (!local_mem) -- 2.7.4