History

powercap: intel_rapl: Fix possible recursive lock warning

With the RAPL PMU addition, there is a recursive locking when CPU online callback function calls rapl_package_add_pmu(). Here cpu_hotplug_lock is already acquired by cpuhp_thread_fun() and rapl_package_add_pmu() tries to acquire again. <4>[ 8.197433] ============================================ <4>[ 8.197437] WARNING: possible recursive locking detected <4>[ 8.197440] 6.19.0-rc1-lgci-xe-xe-4242-05b7c58b3367dca84+ #1 Not tainted <4>[ 8.197444] -------------------------------------------- <4>[ 8.197447] cpuhp/0/20 is trying to acquire lock: <4>[ 8.197450] ffffffff83487870 (cpu_hotplug_lock){++++}-{0:0}, at: rapl_package_add_pmu+0x37/0x370 [intel_rapl_common] <4>[ 8.197463] but task is already holding lock: <4>[ 8.197466] ffffffff83487870 (cpu_hotplug_lock){++++}-{0:0}, at: cpuhp_thread_fun+0x6d/0x290 <4>[ 8.197477] other info that might help us debug this: <4>[ 8.197480] Possible unsafe locking scenario: <4>[ 8.197483] CPU0 <4>[ 8.197485] ---- <4>[ 8.197487] lock(cpu_hotplug_lock); <4>[ 8.197490] lock(cpu_hotplug_lock); <4>[ 8.197493] *** DEADLOCK *** .. .. <4>[ 8.197542] __lock_acquire+0x146e/0x2790 <4>[ 8.197548] lock_acquire+0xc4/0x2c0 <4>[ 8.197550] ? rapl_package_add_pmu+0x37/0x370 [intel_rapl_common] <4>[ 8.197556] cpus_read_lock+0x41/0x110 <4>[ 8.197558] ? rapl_package_add_pmu+0x37/0x370 [intel_rapl_common] <4>[ 8.197561] rapl_package_add_pmu+0x37/0x370 [intel_rapl_common] <4>[ 8.197565] rapl_cpu_online+0x85/0x87 [intel_rapl_msr] <4>[ 8.197568] ? __pfx_rapl_cpu_online+0x10/0x10 [intel_rapl_msr] <4>[ 8.197570] cpuhp_invoke_callback+0x41f/0x6c0 <4>[ 8.197573] ? cpuhp_thread_fun+0x6d/0x290 <4>[ 8.197575] cpuhp_thread_fun+0x1e2/0x290 <4>[ 8.197578] ? smpboot_thread_fn+0x26/0x290 <4>[ 8.197581] smpboot_thread_fn+0x12f/0x290 <4>[ 8.197584] ? __pfx_smpboot_thread_fn+0x10/0x10 <4>[ 8.197586] kthread+0x11f/0x250 <4>[ 8.197589] ? __pfx_kthread+0x10/0x10 <4>[ 8.197592] ret_from_fork+0x344/0x3a0 <4>[ 8.197595] ? __pfx_kthread+0x10/0x10 <4>[ 8.197597] ret_from_fork_asm+0x1a/0x30 <4>[ 8.197604] </TASK> Fix this issue in the same way as rapl powercap package domain is added from the same CPU online callback by introducing another interface which doesn't call cpus_read_lock(). Add rapl_package_add_pmu_locked() and rapl_package_remove_pmu_locked() which don't call cpus_read_lock().

Srinivas Pandruvada committed 11 months ago

Commit dcd0b621 parent efc4c353 files · +24 −8Browse files

Unified · Split

3 files
  1. drivers/powercap/intel_rapl_common.c +18 −6
  2. drivers/powercap/intel_rapl_msr.c +2 −2
  3. include/linux/intel_rapl.h +4 −0
drivers/powercap/intel_rapl_common.c+18−6
@@ -2032,7 +2032,7 @@ end: · 2031 unmodified lines
2032 return ret;2032 return ret;
2033}2033}
2034 2034
2035int rapl_package_add_pmu(struct rapl_package *rp)2035int rapl_package_add_pmu_locked(struct rapl_package *rp)
2036{2036{
2037 struct rapl_package_pmu_data *data = &rp->pmu_data;2037 struct rapl_package_pmu_data *data = &rp->pmu_data;
2038 int idx;2038 int idx;
@@ -2040,8 +2040,6 @@ int rapl_package_add_pmu(struct rapl_package *rp) · 1 unmodified line
2040 if (rp->has_pmu)2040 if (rp->has_pmu)
2041 return -EEXIST;2041 return -EEXIST;
2042 2042
2043 guard(cpus_read_lock)();
2044
2045 for (idx = 0; idx < rp->nr_domains; idx++) {2043 for (idx = 0; idx < rp->nr_domains; idx++) {
2046 struct rapl_domain *rd = &rp->domains[idx];2044 struct rapl_domain *rd = &rp->domains[idx];
2047 int domain = rd->id;2045 int domain = rd->id;
@@ -2091,17 +2089,23 @@ int rapl_package_add_pmu(struct rapl_package *rp) · 43 unmodified lines
2091 2089
2092 return rapl_pmu_update(rp);2090 return rapl_pmu_update(rp);
2093}2091}
2092EXPORT_SYMBOL_GPL(rapl_package_add_pmu_locked);
2093
2094int rapl_package_add_pmu(struct rapl_package *rp)
2095{
2096 guard(cpus_read_lock)();
2097
2098 return rapl_package_add_pmu_locked(rp);
2099}
2094EXPORT_SYMBOL_GPL(rapl_package_add_pmu);2100EXPORT_SYMBOL_GPL(rapl_package_add_pmu);
2095 2101
2096void rapl_package_remove_pmu(struct rapl_package *rp)2102void rapl_package_remove_pmu_locked(struct rapl_package *rp)
2097{2103{
2098 struct rapl_package *pos;2104 struct rapl_package *pos;
2099 2105
2100 if (!rp->has_pmu)2106 if (!rp->has_pmu)
2101 return;2107 return;
2102 2108
2103 guard(cpus_read_lock)();
2104
2105 list_for_each_entry(pos, &rapl_packages, plist) {2109 list_for_each_entry(pos, &rapl_packages, plist) {
2106 /* PMU is still needed */2110 /* PMU is still needed */
2107 if (pos->has_pmu && pos != rp)2111 if (pos->has_pmu && pos != rp)
@@ -2111,6 +2115,14 @@ void rapl_package_remove_pmu(struct rapl_package *rp) · 3 unmodified lines
2111 perf_pmu_unregister(&rapl_pmu.pmu);2115 perf_pmu_unregister(&rapl_pmu.pmu);
2112 memset(&rapl_pmu, 0, sizeof(struct rapl_pmu));2116 memset(&rapl_pmu, 0, sizeof(struct rapl_pmu));
2113}2117}
2118EXPORT_SYMBOL_GPL(rapl_package_remove_pmu_locked);
2119
2120void rapl_package_remove_pmu(struct rapl_package *rp)
2121{
2122 guard(cpus_read_lock)();
2123
2124 rapl_package_remove_pmu_locked(rp);
2125}
2114EXPORT_SYMBOL_GPL(rapl_package_remove_pmu);2126EXPORT_SYMBOL_GPL(rapl_package_remove_pmu);
2115#endif2127#endif
2116 2128
drivers/powercap/intel_rapl_msr.c+2−2
@@ -82,7 +82,7 @@ static int rapl_cpu_online(unsigned int cpu) · 81 unmodified lines
82 if (IS_ERR(rp))82 if (IS_ERR(rp))
83 return PTR_ERR(rp);83 return PTR_ERR(rp);
84 if (rapl_msr_pmu)84 if (rapl_msr_pmu)
85 rapl_package_add_pmu(rp);85 rapl_package_add_pmu_locked(rp);
86 }86 }
87 cpumask_set_cpu(cpu, &rp->cpumask);87 cpumask_set_cpu(cpu, &rp->cpumask);
88 return 0;88 return 0;
@@ -101,7 +101,7 @@ static int rapl_cpu_down_prep(unsigned int cpu) · 12 unmodified lines
101 lead_cpu = cpumask_first(&rp->cpumask);101 lead_cpu = cpumask_first(&rp->cpumask);
102 if (lead_cpu >= nr_cpu_ids) {102 if (lead_cpu >= nr_cpu_ids) {
103 if (rapl_msr_pmu)103 if (rapl_msr_pmu)
104 rapl_package_remove_pmu(rp);104 rapl_package_remove_pmu_locked(rp);
105 rapl_remove_package_cpuslocked(rp);105 rapl_remove_package_cpuslocked(rp);
106 } else if (rp->lead_cpu == cpu) {106 } else if (rp->lead_cpu == cpu) {
107 rp->lead_cpu = lead_cpu;107 rp->lead_cpu = lead_cpu;
include/linux/intel_rapl.h+4−0
@@ -214,10 +214,14 @@ void rapl_remove_package(struct rapl_package *rp); · 213 unmodified lines
214 214
215#ifdef CONFIG_PERF_EVENTS215#ifdef CONFIG_PERF_EVENTS
216int rapl_package_add_pmu(struct rapl_package *rp);216int rapl_package_add_pmu(struct rapl_package *rp);
217int rapl_package_add_pmu_locked(struct rapl_package *rp);
217void rapl_package_remove_pmu(struct rapl_package *rp);218void rapl_package_remove_pmu(struct rapl_package *rp);
219void rapl_package_remove_pmu_locked(struct rapl_package *rp);
218#else220#else
219static inline int rapl_package_add_pmu(struct rapl_package *rp) { return 0; }221static inline int rapl_package_add_pmu(struct rapl_package *rp) { return 0; }
222static inline int rapl_package_add_pmu_locked(struct rapl_package *rp) { return 0; }
220static inline void rapl_package_remove_pmu(struct rapl_package *rp) { }223static inline void rapl_package_remove_pmu(struct rapl_package *rp) { }
224static inline void rapl_package_remove_pmu_locked(struct rapl_package *rp) { }
221#endif225#endif
222 226
223#endif /* __INTEL_RAPL_H__ */227#endif /* __INTEL_RAPL_H__ */