tree: https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git master head: 320eb81df4f6c1a1814fd02ebb4ba41eb80a3c7e commit: dc836adde581f0d755d58926912bf489bda7e4f5 [1552/4667] rcu: Summarize RCU CPU stall warnings during CSD-lock stalls config: x86_64-randconfig-013-20240814 (https://download.01.org/0day-ci/archive/20240814/202408142217.iUA6mot8-lkp@xxxxxxxxx/config) compiler: gcc-11 (Debian 11.3.0-12) 11.3.0 reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20240814/202408142217.iUA6mot8-lkp@xxxxxxxxx/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@xxxxxxxxx> | Closes: https://lore.kernel.org/oe-kbuild-all/202408142217.iUA6mot8-lkp@xxxxxxxxx/ All errors (new ones prefixed by >>): ld: vmlinux.o: in function `check_cpu_stall': >> kernel/rcu/tree_stall.h:798: undefined reference to `csd_lock_is_stuck' vim +798 kernel/rcu/tree_stall.h 725 726 static void check_cpu_stall(struct rcu_data *rdp) 727 { 728 bool self_detected; 729 unsigned long gs1; 730 unsigned long gs2; 731 unsigned long gps; 732 unsigned long j; 733 unsigned long jn; 734 unsigned long js; 735 struct rcu_node *rnp; 736 737 lockdep_assert_irqs_disabled(); 738 if ((rcu_stall_is_suppressed() && !READ_ONCE(rcu_kick_kthreads)) || 739 !rcu_gp_in_progress()) 740 return; 741 rcu_stall_kick_kthreads(); 742 743 /* 744 * Check if it was requested (via rcu_cpu_stall_reset()) that the FQS 745 * loop has to set jiffies to ensure a non-stale jiffies value. This 746 * is required to have good jiffies value after coming out of long 747 * breaks of jiffies updates. Not doing so can cause false positives. 748 */ 749 if (READ_ONCE(rcu_state.nr_fqs_jiffies_stall) > 0) 750 return; 751 752 j = jiffies; 753 754 /* 755 * Lots of memory barriers to reject false positives. 756 * 757 * The idea is to pick up rcu_state.gp_seq, then 758 * rcu_state.jiffies_stall, then rcu_state.gp_start, and finally 759 * another copy of rcu_state.gp_seq. These values are updated in 760 * the opposite order with memory barriers (or equivalent) during 761 * grace-period initialization and cleanup. Now, a false positive 762 * can occur if we get an new value of rcu_state.gp_start and a old 763 * value of rcu_state.jiffies_stall. But given the memory barriers, 764 * the only way that this can happen is if one grace period ends 765 * and another starts between these two fetches. This is detected 766 * by comparing the second fetch of rcu_state.gp_seq with the 767 * previous fetch from rcu_state.gp_seq. 768 * 769 * Given this check, comparisons of jiffies, rcu_state.jiffies_stall, 770 * and rcu_state.gp_start suffice to forestall false positives. 771 */ 772 gs1 = READ_ONCE(rcu_state.gp_seq); 773 smp_rmb(); /* Pick up ->gp_seq first... */ 774 js = READ_ONCE(rcu_state.jiffies_stall); 775 smp_rmb(); /* ...then ->jiffies_stall before the rest... */ 776 gps = READ_ONCE(rcu_state.gp_start); 777 smp_rmb(); /* ...and finally ->gp_start before ->gp_seq again. */ 778 gs2 = READ_ONCE(rcu_state.gp_seq); 779 if (gs1 != gs2 || 780 ULONG_CMP_LT(j, js) || 781 ULONG_CMP_GE(gps, js)) 782 return; /* No stall or GP completed since entering function. */ 783 rnp = rdp->mynode; 784 jn = jiffies + ULONG_MAX / 2; 785 self_detected = READ_ONCE(rnp->qsmask) & rdp->grpmask; 786 if (rcu_gp_in_progress() && 787 (self_detected || ULONG_CMP_GE(j, js + RCU_STALL_RAT_DELAY)) && 788 cmpxchg(&rcu_state.jiffies_stall, js, jn) == js) { 789 /* 790 * If a virtual machine is stopped by the host it can look to 791 * the watchdog like an RCU stall. Check to see if the host 792 * stopped the vm. 793 */ 794 if (kvm_check_and_clear_guest_paused()) 795 return; 796 797 rcu_stall_notifier_call_chain(RCU_STALL_NOTIFY_NORM, (void *)j - gps); > 798 if (READ_ONCE(csd_lock_suppress_rcu_stall) && csd_lock_is_stuck()) { 799 pr_err("INFO: %s detected stall, but suppressed full report due to a stuck CSD-lock.\n", rcu_state.name); 800 } else if (self_detected) { 801 /* We haven't checked in, so go dump stack. */ 802 print_cpu_stall(gps); 803 } else { 804 /* They had a few time units to dump stack, so complain. */ 805 print_other_cpu_stall(gs2, gps); 806 } 807 808 if (READ_ONCE(rcu_cpu_stall_ftrace_dump)) 809 rcu_ftrace_dump(DUMP_ALL); 810 811 if (READ_ONCE(rcu_state.jiffies_stall) == jn) { 812 jn = jiffies + 3 * rcu_jiffies_till_stall_check() + 3; 813 WRITE_ONCE(rcu_state.jiffies_stall, jn); 814 } 815 } 816 } 817 -- 0-DAY CI Kernel Test Service https://github.com/intel/lkp-tests/wiki
