[PATCH v2 1/4] powerpc/perf/core-book3s: Avoid loading platform pmu driver during dump kernel

[PATCH v2 1/4] powerpc/perf/core-book3s: Avoid loading platform pmu driver during dump kernel

[Madhavan Srinivasan]

Platform pmu driver are built-in and these does not
have a commandline parameter to avoid loading them
during the kdump kernel. Add check for kdump and fadump
kernel to avoids loading them.

Signed-off-by: Madhavan Srinivasan <maddy@linux.ibm.com>
---
Changelog v1:
- Added more details to commmit message

 arch/powerpc/perf/core-book3s.c | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/arch/powerpc/perf/core-book3s.c b/arch/powerpc/perf/core-book3s.c
index 2b79171ee185..9a009dff63b3 100644
--- a/arch/powerpc/perf/core-book3s.c
+++ b/arch/powerpc/perf/core-book3s.c
@@ -22,6 +22,8 @@
 
 #ifdef CONFIG_PPC64
 #include "internal.h"
+#include <asm/fadump.h>
+#include <asm/kexec.h>
 #endif
 
 #define BHRB_MAX_ENTRIES	32
@@ -2571,6 +2573,12 @@ static int __init init_ppc64_pmu(void)
 		return 0;
 	}
 
+	/*
+	 * If the dump kernel is active, skip loading these drivers
+	 */
+	if (is_kdump_kernel() || is_fadump_active())
+		return 0;
+
 	/* run through all the pmu drivers one at a time */
 	if (!init_power5_pmu())
 		return 0;
-- 
2.47.0

[PATCH v2 2/4] powerpc/perf/hv-24x7: Avoid loading hv-24x7 during dump kernel

[Madhavan Srinivasan]

hv-24x7 pmu driver are intended to get system-wide resourse
metrics and these are built-in. hv-24x7 pmu does not have
commandline option to disable them. Add check for kdump and
fadump kernel to avoids loading them.

Signed-off-by: Madhavan Srinivasan <maddy@linux.ibm.com>
---
Changelog v1:
- Added more details to commmit message

 arch/powerpc/perf/hv-24x7.c   | 3 +++
 arch/powerpc/perf/hv-common.h | 2 ++
 2 files changed, 5 insertions(+)

diff --git a/arch/powerpc/perf/hv-24x7.c b/arch/powerpc/perf/hv-24x7.c
index d400fa391c27..3a626cd8cf54 100644
--- a/arch/powerpc/perf/hv-24x7.c
+++ b/arch/powerpc/perf/hv-24x7.c
@@ -1698,6 +1698,9 @@ static int hv_24x7_init(void)
 	unsigned int pvr = mfspr(SPRN_PVR);
 	struct hv_perf_caps caps;
 
+	if (is_kdump_kernel() || is_fadump_active())
+		return 0;
+
 	if (!firmware_has_feature(FW_FEATURE_LPAR)) {
 		pr_debug("not a virtualized system, not enabling\n");
 		return -ENODEV;
diff --git a/arch/powerpc/perf/hv-common.h b/arch/powerpc/perf/hv-common.h
index 2cce17bc321c..a4c062d2264e 100644
--- a/arch/powerpc/perf/hv-common.h
+++ b/arch/powerpc/perf/hv-common.h
@@ -4,6 +4,8 @@
 
 #include <linux/perf_event.h>
 #include <linux/types.h>
+#include <asm/fadump.h>
+#include <asm/kexec.h>
 
 struct hv_perf_caps {
 	u16 version;
-- 
2.47.0

[PATCH v2 3/4] powerpc/perf/hv-gpci: Avoid loading hv-gpci pmu during dump kernel

[Madhavan Srinivasan]

hv-gpci pmu driver is intended to get powervm hypervisor
system-wide metrics and these are built-in. hv-gpci pmu
not have commandline option to disable them. Add check
for kdump and fadump kernel to avoids loading them.

Signed-off-by: Madhavan Srinivasan <maddy@linux.ibm.com>
---
Changelog v1:
- Added more details to commmit message

 arch/powerpc/perf/hv-gpci.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/arch/powerpc/perf/hv-gpci.c b/arch/powerpc/perf/hv-gpci.c
index 241551d1282f..e0c3df0a048f 100644
--- a/arch/powerpc/perf/hv-gpci.c
+++ b/arch/powerpc/perf/hv-gpci.c
@@ -989,6 +989,9 @@ static int hv_gpci_init(void)
 	struct hv_perf_caps caps;
 	struct hv_gpci_request_buffer *arg;
 
+	if (is_kdump_kernel() || is_fadump_active())
+		return 0;
+
 	hv_gpci_assert_offsets_correct();
 
 	if (!firmware_has_feature(FW_FEATURE_LPAR)) {
-- 
2.47.0

[PATCH v2 4/4] powerpc/perf/vpa-pmu: Avoid loading vpa-pmu driver during dump kernel

[Madhavan Srinivasan]

vpa-pmu driver is used to collect latency metrics for host to guest
or guest to host context switches in a PowerVM KVM guest scenario.
These are enabled using CONFIG_VPA_PMU and when these are compiled
as built-in, add check for kdump and fadump kernel to avoids
loading them, since there are no commandline option to disable them.

Signed-off-by: Madhavan Srinivasan <maddy@linux.ibm.com>
---
Changelog v1:
- Added more details to commmit message

 arch/powerpc/perf/vpa-pmu.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/arch/powerpc/perf/vpa-pmu.c b/arch/powerpc/perf/vpa-pmu.c
index 6a5bfd2a13b5..f7f72b976c1d 100644
--- a/arch/powerpc/perf/vpa-pmu.c
+++ b/arch/powerpc/perf/vpa-pmu.c
@@ -10,6 +10,8 @@
 #include <linux/perf_event.h>
 #include <asm/kvm_ppc.h>
 #include <asm/kvm_book3s_64.h>
+#include <asm/fadump.h>
+#include <asm/kexec.h>
 
 #define MODULE_VERS "1.0"
 #define MODULE_NAME "pseries_vpa_pmu"
@@ -183,6 +185,9 @@ static int __init pseries_vpa_pmu_init(void)
 	if (!firmware_has_feature(FW_FEATURE_LPAR) || is_kvm_guest())
 		return -ENODEV;
 
+	if (is_kdump_kernel() || is_fadump_active())
+		return 0;
+
 	perf_pmu_register(&vpa_pmu, vpa_pmu.name, -1);
 	pr_info("Virtual Processor Area PMU registered.\n");
 
-- 
2.47.0

Re: [PATCH v2 1/4] powerpc/perf/core-book3s: Avoid loading platform pmu driver during dump kernel

[kernel test robot]

Hi Madhavan,

kernel test robot noticed the following build errors:

[auto build test ERROR on powerpc/next]
[also build test ERROR on powerpc/fixes linus/master v6.14-rc4 next-20250228]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch#_base_tree_information]

url:    https://github.com/intel-lab-lkp/linux/commits/Madhavan-Srinivasan/powerpc-perf-hv-24x7-Avoid-loading-hv-24x7-during-dump-kernel/20250302-022531
base:   https://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux.git next
patch link:    https://lore.kernel.org/r/20250301182310.6832-1-maddy%40linux.ibm.com
patch subject: [PATCH v2 1/4] powerpc/perf/core-book3s: Avoid loading platform pmu driver during dump kernel
config: powerpc64-randconfig-001-20250302 (https://download.01.org/0day-ci/archive/20250302/202503021712.eVdjymsT-lkp@intel.com/config)
compiler: powerpc64-linux-gcc (GCC) 14.2.0
reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20250302/202503021712.eVdjymsT-lkp@intel.com/reproduce)

If you fix the issue in a separate patch/commit (i.e. not just a new version of
the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <lkp@intel.com>
| Closes: https://lore.kernel.org/oe-kbuild-all/202503021712.eVdjymsT-lkp@intel.com/

All errors (new ones prefixed by >>):

   arch/powerpc/perf/core-book3s.c: In function 'init_ppc64_pmu':
>> arch/powerpc/perf/core-book3s.c:2599:13: error: implicit declaration of function 'is_kdump_kernel' [-Wimplicit-function-declaration]
    2599 |         if (is_kdump_kernel() || is_fadump_active())
         |             ^~~~~~~~~~~~~~~


vim +/is_kdump_kernel +2599 arch/powerpc/perf/core-book3s.c

  2587	
  2588	static int __init init_ppc64_pmu(void)
  2589	{
  2590		if (cpu_has_feature(CPU_FTR_HVMODE) && pmu_override) {
  2591			pr_warn("disabling perf due to pmu_override= command line option.\n");
  2592			on_each_cpu(do_pmu_override, NULL, 1);
  2593			return 0;
  2594		}
  2595	
  2596		/*
  2597		 * If the dump kernel is active, skip loading these drivers
  2598		 */
> 2599		if (is_kdump_kernel() || is_fadump_active())
  2600			return 0;
  2601	
  2602		/* run through all the pmu drivers one at a time */
  2603		if (!init_power5_pmu())
  2604			return 0;
  2605		else if (!init_power5p_pmu())
  2606			return 0;
  2607		else if (!init_power6_pmu())
  2608			return 0;
  2609		else if (!init_power7_pmu())
  2610			return 0;
  2611		else if (!init_power8_pmu())
  2612			return 0;
  2613		else if (!init_power9_pmu())
  2614			return 0;
  2615		else if (!init_power10_pmu())
  2616			return 0;
  2617		else if (!init_power11_pmu())
  2618			return 0;
  2619		else if (!init_ppc970_pmu())
  2620			return 0;
  2621		else
  2622			return init_generic_compat_pmu();
  2623	}
  2624	early_initcall(init_ppc64_pmu);
  2625	

-- 
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests/wiki

Re: [PATCH v2 2/4] powerpc/perf/hv-24x7: Avoid loading hv-24x7 during dump kernel

[kernel test robot]

Hi Madhavan,

kernel test robot noticed the following build errors:

[auto build test ERROR on powerpc/next]
[also build test ERROR on powerpc/fixes linus/master v6.14-rc4 next-20250228]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch#_base_tree_information]

url:    https://github.com/intel-lab-lkp/linux/commits/Madhavan-Srinivasan/powerpc-perf-hv-24x7-Avoid-loading-hv-24x7-during-dump-kernel/20250302-022531
base:   https://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux.git next
patch link:    https://lore.kernel.org/r/20250301182310.6832-2-maddy%40linux.ibm.com
patch subject: [PATCH v2 2/4] powerpc/perf/hv-24x7: Avoid loading hv-24x7 during dump kernel
config: powerpc64-randconfig-001-20250302 (https://download.01.org/0day-ci/archive/20250302/202503021908.Ed1gO0Gi-lkp@intel.com/config)
compiler: powerpc64-linux-gcc (GCC) 14.2.0
reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20250302/202503021908.Ed1gO0Gi-lkp@intel.com/reproduce)

If you fix the issue in a separate patch/commit (i.e. not just a new version of
the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <lkp@intel.com>
| Closes: https://lore.kernel.org/oe-kbuild-all/202503021908.Ed1gO0Gi-lkp@intel.com/

All errors (new ones prefixed by >>):

   arch/powerpc/perf/hv-24x7.c: In function 'hv_24x7_init':
>> arch/powerpc/perf/hv-24x7.c:1701:13: error: implicit declaration of function 'is_kdump_kernel' [-Wimplicit-function-declaration]
    1701 |         if (is_kdump_kernel() || is_fadump_active())
         |             ^~~~~~~~~~~~~~~


vim +/is_kdump_kernel +1701 arch/powerpc/perf/hv-24x7.c

  1693	
  1694	static int hv_24x7_init(void)
  1695	{
  1696		int r;
  1697		unsigned long hret;
  1698		unsigned int pvr = mfspr(SPRN_PVR);
  1699		struct hv_perf_caps caps;
  1700	
> 1701		if (is_kdump_kernel() || is_fadump_active())
  1702			return 0;
  1703	
  1704		if (!firmware_has_feature(FW_FEATURE_LPAR)) {
  1705			pr_debug("not a virtualized system, not enabling\n");
  1706			return -ENODEV;
  1707		}
  1708	
  1709		/* POWER8 only supports v1, while POWER9 only supports v2. */
  1710		if (PVR_VER(pvr) == PVR_POWER8 || PVR_VER(pvr) == PVR_POWER8E ||
  1711		    PVR_VER(pvr) == PVR_POWER8NVL)
  1712			interface_version = 1;
  1713		else {
  1714			interface_version = 2;
  1715	
  1716			/* SMT8 in POWER9 needs to aggregate result elements. */
  1717			if (threads_per_core == 8)
  1718				aggregate_result_elements = true;
  1719		}
  1720	
  1721		hret = hv_perf_caps_get(&caps);
  1722		if (hret) {
  1723			pr_debug("could not obtain capabilities, not enabling, rc=%ld\n",
  1724					hret);
  1725			return -ENODEV;
  1726		}
  1727	
  1728		hv_page_cache = kmem_cache_create("hv-page-4096", 4096, 4096, 0, NULL);
  1729		if (!hv_page_cache)
  1730			return -ENOMEM;
  1731	
  1732		/* sampling not supported */
  1733		h_24x7_pmu.capabilities |= PERF_PMU_CAP_NO_INTERRUPT;
  1734	
  1735		r = create_events_from_catalog(&event_group.attrs,
  1736					   &event_desc_group.attrs,
  1737					   &event_long_desc_group.attrs);
  1738	
  1739		if (r)
  1740			return r;
  1741	
  1742		/* init cpuhotplug */
  1743		r = hv_24x7_cpu_hotplug_init();
  1744		if (r)
  1745			return r;
  1746	
  1747		r = perf_pmu_register(&h_24x7_pmu, h_24x7_pmu.name, -1);
  1748		if (r)
  1749			return r;
  1750	
  1751		read_24x7_sys_info();
  1752	
  1753		return 0;
  1754	}
  1755	

-- 
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests/wiki

Re: [PATCH v2 1/4] powerpc/perf/core-book3s: Avoid loading platform pmu driver during dump kernel

[kernel test robot]

Hi Madhavan,

kernel test robot noticed the following build errors:

[auto build test ERROR on powerpc/next]
[also build test ERROR on powerpc/fixes linus/master v6.14-rc5 next-20250305]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch#_base_tree_information]

url:    https://github.com/intel-lab-lkp/linux/commits/Madhavan-Srinivasan/powerpc-perf-hv-24x7-Avoid-loading-hv-24x7-during-dump-kernel/20250302-022531
base:   https://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux.git next
patch link:    https://lore.kernel.org/r/20250301182310.6832-1-maddy%40linux.ibm.com
patch subject: [PATCH v2 1/4] powerpc/perf/core-book3s: Avoid loading platform pmu driver during dump kernel
config: powerpc64-randconfig-001-20250305 (https://download.01.org/0day-ci/archive/20250305/202503052346.eTbppObo-lkp@intel.com/config)
compiler: clang version 19.1.7 (https://github.com/llvm/llvm-project cd708029e0b2869e80abe31ddb175f7c35361f90)
reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20250305/202503052346.eTbppObo-lkp@intel.com/reproduce)

If you fix the issue in a separate patch/commit (i.e. not just a new version of
the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <lkp@intel.com>
| Closes: https://lore.kernel.org/oe-kbuild-all/202503052346.eTbppObo-lkp@intel.com/

All errors (new ones prefixed by >>):

>> arch/powerpc/perf/core-book3s.c:2599:6: error: call to undeclared function 'is_kdump_kernel'; ISO C99 and later do not support implicit function declarations [-Wimplicit-function-declaration]
    2599 |         if (is_kdump_kernel() || is_fadump_active())
         |             ^
   1 error generated.


vim +/is_kdump_kernel +2599 arch/powerpc/perf/core-book3s.c

  2587	
  2588	static int __init init_ppc64_pmu(void)
  2589	{
  2590		if (cpu_has_feature(CPU_FTR_HVMODE) && pmu_override) {
  2591			pr_warn("disabling perf due to pmu_override= command line option.\n");
  2592			on_each_cpu(do_pmu_override, NULL, 1);
  2593			return 0;
  2594		}
  2595	
  2596		/*
  2597		 * If the dump kernel is active, skip loading these drivers
  2598		 */
> 2599		if (is_kdump_kernel() || is_fadump_active())
  2600			return 0;
  2601	
  2602		/* run through all the pmu drivers one at a time */
  2603		if (!init_power5_pmu())
  2604			return 0;
  2605		else if (!init_power5p_pmu())
  2606			return 0;
  2607		else if (!init_power6_pmu())
  2608			return 0;
  2609		else if (!init_power7_pmu())
  2610			return 0;
  2611		else if (!init_power8_pmu())
  2612			return 0;
  2613		else if (!init_power9_pmu())
  2614			return 0;
  2615		else if (!init_power10_pmu())
  2616			return 0;
  2617		else if (!init_power11_pmu())
  2618			return 0;
  2619		else if (!init_ppc970_pmu())
  2620			return 0;
  2621		else
  2622			return init_generic_compat_pmu();
  2623	}
  2624	early_initcall(init_ppc64_pmu);
  2625	

-- 
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests/wiki