perf tools fixes for v6.0: 4th batch

- Fix polling of system-wide events related to mixing per-cpu and per-thread
   events.
 
 - Do not check if /proc/modules is unchanged when copying /proc/kcore,
   that doesn't get in the way of post processing analysis.
 
 - Include program header in ELF files generated for JIT files, so that they can
   be opened by tools using elfutils libraries.
 
 - Enter namespaces when synthesizing build-ids.
 
 - Fix some bugs related to a recent cpu_map overhaul where we should be
   using an index and not the cpu number.
 
 - Fix BPF program ELF section name, using the naming expected by libbpf when
   using BPF counters in 'perf stat'.
 
 - Add a new test for perf stat cgroup BPF counter.
 
 - Adjust check on 'perf test wp' for older kernels, where the
   PERF_EVENT_IOC_MODIFY_ATTRIBUTES ioctl isn't supported.
 
 - Sync x86 cpufeatures with the kernel sources, no changes in tooling.
 
 Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
 -----BEGIN PGP SIGNATURE-----
 
 iHUEABYKAB0WIQR2GiIUctdOfX2qHhGyPKLppCJ+JwUCYyuOXgAKCRCyPKLppCJ+
 J7PsAPwMmEKNG08rBuZG64my1sS1hKNP2uAWcii871hFe0jSowEArWd2ErRw5wPw
 Y3uT9gP11NHDN+B87HxVErNkNUp3PQM=
 =S2hR
 -----END PGP SIGNATURE-----

Merge tag 'perf-tools-fixes-for-v6.0-2022-09-21' of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux

Pull perf tools fixes from Arnaldo Carvalho de Melo:

 - Fix polling of system-wide events related to mixing per-cpu and
   per-thread events.

 - Do not check if /proc/modules is unchanged when copying /proc/kcore,
   that doesn't get in the way of post processing analysis.

 - Include program header in ELF files generated for JIT files, so that
   they can be opened by tools using elfutils libraries.

 - Enter namespaces when synthesizing build-ids.

 - Fix some bugs related to a recent cpu_map overhaul where we should be
   using an index and not the cpu number.

 - Fix BPF program ELF section name, using the naming expected by libbpf
   when using BPF counters in 'perf stat'.

 - Add a new test for perf stat cgroup BPF counter.

 - Adjust check on 'perf test wp' for older kernels, where the
   PERF_EVENT_IOC_MODIFY_ATTRIBUTES ioctl isn't supported.

 - Sync x86 cpufeatures with the kernel sources, no changes in tooling.

* tag 'perf-tools-fixes-for-v6.0-2022-09-21' of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux:
  perf tools: Honor namespace when synthesizing build-ids
  tools headers cpufeatures: Sync with the kernel sources
  perf kcore_copy: Do not check /proc/modules is unchanged
  libperf evlist: Fix polling of system-wide events
  perf record: Fix cpu mask bit setting for mixed mmaps
  perf test: Skip wp modify test on old kernels
  perf jit: Include program header in ELF files
  perf test: Add a new test for perf stat cgroup BPF counter
  perf stat: Use evsel->core.cpus to iterate cpus in BPF cgroup counters
  perf stat: Fix cpu map index in bperf cgroup code
  perf stat: Fix BPF program section name

tools/arch/x86/include/asm/cpufeatures.h

@@ -457,7 +457,8 @@
 #define X86_BUG_ITLB_MULTIHIT		X86_BUG(23) /* CPU may incur MCE during certain page attribute changes */
 #define X86_BUG_SRBDS			X86_BUG(24) /* CPU may leak RNG bits if not mitigated */
 #define X86_BUG_MMIO_STALE_DATA		X86_BUG(25) /* CPU is affected by Processor MMIO Stale Data vulnerabilities */
-#define X86_BUG_RETBLEED		X86_BUG(26) /* CPU is affected by RETBleed */
-#define X86_BUG_EIBRS_PBRSB		X86_BUG(27) /* EIBRS is vulnerable to Post Barrier RSB Predictions */
+#define X86_BUG_MMIO_UNKNOWN		X86_BUG(26) /* CPU is too old and its MMIO Stale Data status is unknown */
+#define X86_BUG_RETBLEED		X86_BUG(27) /* CPU is affected by RETBleed */
+#define X86_BUG_EIBRS_PBRSB		X86_BUG(28) /* EIBRS is vulnerable to Post Barrier RSB Predictions */
 
 #endif /* _ASM_X86_CPUFEATURES_H */

tools/lib/perf/evlist.c

@@ -441,6 +441,7 @@ mmap_per_evsel(struct perf_evlist *evlist, struct perf_evlist_mmap_ops *ops,
 
 	perf_evlist__for_each_entry(evlist, evsel) {
 		bool overwrite = evsel->attr.write_backward;
+		enum fdarray_flags flgs;
 		struct perf_mmap *map;
 		int *output, fd, cpu;
 
@@ -504,8 +505,8 @@ mmap_per_evsel(struct perf_evlist *evlist, struct perf_evlist_mmap_ops *ops,
 
 		revent = !overwrite ? POLLIN : 0;
 
-		if (!evsel->system_wide &&
-		    perf_evlist__add_pollfd(evlist, fd, map, revent, fdarray_flag__default) < 0) {
+		flgs = evsel->system_wide ? fdarray_flag__nonfilterable : fdarray_flag__default;
+		if (perf_evlist__add_pollfd(evlist, fd, map, revent, flgs) < 0) {
 			perf_mmap__put(map);
 			return -1;
 		}

tools/perf/builtin-record.c

@@ -3371,6 +3371,8 @@ static int record__mmap_cpu_mask_init(struct mmap_cpu_mask *mask, struct perf_cp
 		return 0;
 
 	perf_cpu_map__for_each_cpu(cpu, idx, cpus) {
+		if (cpu.cpu == -1)
+			continue;
 		/* Return ENODEV is input cpu is greater than max cpu */
 		if ((unsigned long)cpu.cpu > mask->nbits)
 			return -ENODEV;

tools/perf/tests/shell/stat_bpf_counters_cgrp.sh

@@ -0,0 +1,83 @@
+#!/bin/sh
+# perf stat --bpf-counters --for-each-cgroup test
+# SPDX-License-Identifier: GPL-2.0
+
+set -e
+
+test_cgroups=
+if [ "$1" = "-v" ]; then
+	verbose="1"
+fi
+
+# skip if --bpf-counters --for-each-cgroup is not supported
+check_bpf_counter()
+{
+	if ! perf stat -a --bpf-counters --for-each-cgroup / true > /dev/null 2>&1; then
+		if [ "${verbose}" = "1" ]; then
+			echo "Skipping: --bpf-counters --for-each-cgroup not supported"
+			perf --no-pager stat -a --bpf-counters --for-each-cgroup / true || true
+		fi
+		exit 2
+	fi
+}
+
+# find two cgroups to measure
+find_cgroups()
+{
+	# try usual systemd slices first
+	if [ -d /sys/fs/cgroup/system.slice -a -d /sys/fs/cgroup/user.slice ]; then
+		test_cgroups="system.slice,user.slice"
+		return
+	fi
+
+	# try root and self cgroups
+	local self_cgrp=$(grep perf_event /proc/self/cgroup | cut -d: -f3)
+	if [ -z ${self_cgrp} ]; then
+		# cgroup v2 doesn't specify perf_event
+		self_cgrp=$(grep ^0: /proc/self/cgroup | cut -d: -f3)
+	fi
+
+	if [ -z ${self_cgrp} ]; then
+		test_cgroups="/"
+	else
+		test_cgroups="/,${self_cgrp}"
+	fi
+}
+
+# As cgroup events are cpu-wide, we cannot simply compare the result.
+# Just check if it runs without failure and has non-zero results.
+check_system_wide_counted()
+{
+	local output
+
+	output=$(perf stat -a --bpf-counters --for-each-cgroup ${test_cgroups} -e cpu-clock -x, sleep 1  2>&1)
+	if echo ${output} | grep -q -F "<not "; then
+		echo "Some system-wide events are not counted"
+		if [ "${verbose}" = "1" ]; then
+			echo ${output}
+		fi
+		exit 1
+	fi
+}
+
+check_cpu_list_counted()
+{
+	local output
+
+	output=$(perf stat -C 1 --bpf-counters --for-each-cgroup ${test_cgroups} -e cpu-clock -x, taskset -c 1 sleep 1  2>&1)
+	if echo ${output} | grep -q -F "<not "; then
+		echo "Some CPU events are not counted"
+		if [ "${verbose}" = "1" ]; then
+			echo ${output}
+		fi
+		exit 1
+	fi
+}
+
+check_bpf_counter
+find_cgroups
+
+check_system_wide_counted
+check_cpu_list_counted
+
+exit 0

tools/perf/tests/wp.c

@@ -2,7 +2,9 @@
 #include <stdlib.h>
 #include <string.h>
 #include <unistd.h>
+#include <errno.h>
 #include <sys/ioctl.h>
+#include <linux/compiler.h>
 #include <linux/hw_breakpoint.h>
 #include <linux/kernel.h>
 #include "tests.h"
@@ -137,8 +139,7 @@ static int test__wp_rw(struct test_suite *test __maybe_unused,
 #endif
 }
 
-static int test__wp_modify(struct test_suite *test __maybe_unused,
-			   int subtest __maybe_unused)
+static int test__wp_modify(struct test_suite *test __maybe_unused, int subtest __maybe_unused)
 {
 #if defined(__s390x__)
 	return TEST_SKIP;
@@ -160,6 +161,11 @@ static int test__wp_modify(struct test_suite *test __maybe_unused,
 	new_attr.disabled = 1;
 	ret = ioctl(fd, PERF_EVENT_IOC_MODIFY_ATTRIBUTES, &new_attr);
 	if (ret < 0) {
+		if (errno == ENOTTY) {
+			test->test_cases[subtest].skip_reason = "missing kernel support";
+			ret = TEST_SKIP;
+		}
+
 		pr_debug("ioctl(PERF_EVENT_IOC_MODIFY_ATTRIBUTES) failed\n");
 		close(fd);
 		return ret;

tools/perf/util/bpf_counter_cgroup.c

@@ -95,7 +95,7 @@ static int bperf_load_program(struct evlist *evlist)
 
 	perf_cpu_map__for_each_cpu(cpu, i, evlist->core.all_cpus) {
 		link = bpf_program__attach_perf_event(skel->progs.on_cgrp_switch,
-						      FD(cgrp_switch, cpu.cpu));
+						      FD(cgrp_switch, i));
 		if (IS_ERR(link)) {
 			pr_err("Failed to attach cgroup program\n");
 			err = PTR_ERR(link);
@@ -115,15 +115,15 @@ static int bperf_load_program(struct evlist *evlist)
 			evsel->cgrp = NULL;
 
 			/* open single copy of the events w/o cgroup */
-			err = evsel__open_per_cpu(evsel, evlist->core.all_cpus, -1);
+			err = evsel__open_per_cpu(evsel, evsel->core.cpus, -1);
 			if (err) {
 				pr_err("Failed to open first cgroup events\n");
 				goto out;
 			}
 
 			map_fd = bpf_map__fd(skel->maps.events);
-			perf_cpu_map__for_each_cpu(cpu, j, evlist->core.all_cpus) {
-				int fd = FD(evsel, cpu.cpu);
+			perf_cpu_map__for_each_cpu(cpu, j, evsel->core.cpus) {
+				int fd = FD(evsel, j);
 				__u32 idx = evsel->core.idx * total_cpus + cpu.cpu;
 
 				err = bpf_map_update_elem(map_fd, &idx, &fd,
@@ -269,7 +269,7 @@ static int bperf_cgrp__read(struct evsel *evsel)
 			goto out;
 		}
 
-		perf_cpu_map__for_each_cpu(cpu, i, evlist->core.all_cpus) {
+		perf_cpu_map__for_each_cpu(cpu, i, evsel->core.cpus) {
 			counts = perf_counts(evsel->counts, i, 0);
 			counts->val = values[cpu.cpu].counter;
 			counts->ena = values[cpu.cpu].enabled;

tools/perf/util/bpf_skel/bperf_cgroup.bpf.c

@@ -176,7 +176,7 @@ static int bperf_cgroup_count(void)
 }
 
 // This will be attached to cgroup-switches event for each cpu
-SEC("perf_events")
+SEC("perf_event")
 int BPF_PROG(on_cgrp_switch)
 {
 	return bperf_cgroup_count();

tools/perf/util/genelf.c

@@ -253,6 +253,7 @@ jit_write_elf(int fd, uint64_t load_addr, const char *sym,
 	Elf_Data *d;
 	Elf_Scn *scn;
 	Elf_Ehdr *ehdr;
+	Elf_Phdr *phdr;
 	Elf_Shdr *shdr;
 	uint64_t eh_frame_base_offset;
 	char *strsym = NULL;
@@ -287,6 +288,19 @@ jit_write_elf(int fd, uint64_t load_addr, const char *sym,
 	ehdr->e_version = EV_CURRENT;
 	ehdr->e_shstrndx= unwinding ? 4 : 2; /* shdr index for section name */
 
+	/*
+	 * setup program header
+	 */
+	phdr = elf_newphdr(e, 1);
+	phdr[0].p_type = PT_LOAD;
+	phdr[0].p_offset = 0;
+	phdr[0].p_vaddr = 0;
+	phdr[0].p_paddr = 0;
+	phdr[0].p_filesz = csize;
+	phdr[0].p_memsz = csize;
+	phdr[0].p_flags = PF_X | PF_R;
+	phdr[0].p_align = 8;
+
 	/*
 	 * setup text section
 	 */

tools/perf/util/genelf.h

@@ -53,8 +53,10 @@ int jit_add_debug_info(Elf *e, uint64_t code_addr, void *debug, int nr_debug_ent
 
 #if GEN_ELF_CLASS == ELFCLASS64
 #define elf_newehdr	elf64_newehdr
+#define elf_newphdr	elf64_newphdr
 #define elf_getshdr	elf64_getshdr
 #define Elf_Ehdr	Elf64_Ehdr
+#define Elf_Phdr	Elf64_Phdr
 #define Elf_Shdr	Elf64_Shdr
 #define Elf_Sym		Elf64_Sym
 #define ELF_ST_TYPE(a)	ELF64_ST_TYPE(a)
@@ -62,8 +64,10 @@ int jit_add_debug_info(Elf *e, uint64_t code_addr, void *debug, int nr_debug_ent
 #define ELF_ST_VIS(a)	ELF64_ST_VISIBILITY(a)
 #else
 #define elf_newehdr	elf32_newehdr
+#define elf_newphdr	elf32_newphdr
 #define elf_getshdr	elf32_getshdr
 #define Elf_Ehdr	Elf32_Ehdr
+#define Elf_Phdr	Elf32_Phdr
 #define Elf_Shdr	Elf32_Shdr
 #define Elf_Sym		Elf32_Sym
 #define ELF_ST_TYPE(a)	ELF32_ST_TYPE(a)

tools/perf/util/symbol-elf.c

@@ -2102,8 +2102,8 @@ static int kcore_copy__compare_file(const char *from_dir, const char *to_dir,
  * unusual.  One significant peculiarity is that the mapping (start -> pgoff)
  * is not the same for the kernel map and the modules map.  That happens because
  * the data is copied adjacently whereas the original kcore has gaps.  Finally,
- * kallsyms and modules files are compared with their copies to check that
- * modules have not been loaded or unloaded while the copies were taking place.
+ * kallsyms file is compared with its copy to check that modules have not been
+ * loaded or unloaded while the copies were taking place.
  *
  * Return: %0 on success, %-1 on failure.
  */
@@ -2166,9 +2166,6 @@ int kcore_copy(const char *from_dir, const char *to_dir)
 			goto out_extract_close;
 	}
 
-	if (kcore_copy__compare_file(from_dir, to_dir, "modules"))
-		goto out_extract_close;
-
 	if (kcore_copy__compare_file(from_dir, to_dir, "kallsyms"))
 		goto out_extract_close;
 

tools/perf/util/synthetic-events.c

@@ -367,13 +367,24 @@ static void perf_record_mmap2__read_build_id(struct perf_record_mmap2 *event,
 					     bool is_kernel)
 {
 	struct build_id bid;
+	struct nsinfo *nsi;
+	struct nscookie nc;
 	int rc;
 
-	if (is_kernel)
+	if (is_kernel) {
 		rc = sysfs__read_build_id("/sys/kernel/notes", &bid);
-	else
-		rc = filename__read_build_id(event->filename, &bid) > 0 ? 0 : -1;
+		goto out;
+	}
 
+	nsi = nsinfo__new(event->pid);
+	nsinfo__mountns_enter(nsi, &nc);
+
+	rc = filename__read_build_id(event->filename, &bid) > 0 ? 0 : -1;
+
+	nsinfo__mountns_exit(&nc);
+	nsinfo__put(nsi);
+
+out:
 	if (rc == 0) {
 		memcpy(event->build_id, bid.data, sizeof(bid.data));
 		event->build_id_size = (u8) bid.size;