Kernel/linux-next
1
0
Fork 0
mirror of https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git synced 2025-01-15 11:47:47 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

rcutorture: Decorate failing reader segments with CPU ID

Browse Source 
This commit adds CPU number to the "Failure/close-call rcutorture reader
segments" list printed at the end of an rcutorture run that had too-short
grace periods.  This information can help debugging interactions with
migration and CPU hotplug.

However, experience indicates that sampling the CPU number in rcutorture's
read-side code can reduce the probability of too-short bugs by a small
integer factor.  And small integer factors are crucial to RCU bug hunting,
so this commit also introduces a default-off RCU_TORTURE_TEST_LOG_CPU
Kconfig option to enable this CPU-number-logging functionality at
build time.

Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
This commit is contained in:
Paul E. McKenney 2024-10-16 09:50:40 -07:00 committed by Uladzislau Rezki (Sony)
parent 5ec090011b
commit 579a05da40
2 changed files with 22 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

15
kernel/rcu/Kconfig.debug
View File

@ -53,6 +53,21 @@ config RCU_TORTURE_TEST
Say M if you want the RCU torture tests to build as a module.
Say N if you are unsure.
config RCU_TORTURE_TEST_LOG_CPU
tristate "Log CPU for rcutorture failures"
depends on RCU_TORTURE_TEST
default n
help
This option causes rcutorture to decorate each entry of its
log of failure/close-call rcutorture reader segments with the
number of the CPU that the reader was running on at the time.
This information can be useful, but it does incur additional
overhead, overhead that can make both failures and close calls
less probable.
Say Y here if you want CPU IDs logged.
Say N if you are unsure.
config RCU_REF_SCALE_TEST
tristate "Scalability tests for read-side synchronization (RCU and others)"
depends on DEBUG_KERNEL

9
kernel/rcu/rcutorture.c
View File

@ -262,6 +262,7 @@ struct rt_read_seg {
unsigned long rt_delay_ms;
unsigned long rt_delay_us;
bool rt_preempted;
int rt_cpu;
};
static int err_segs_recorded;
static struct rt_read_seg err_segs[RCUTORTURE_RDR_MAX_SEGS];
@ -1862,6 +1863,8 @@ static void rcutorture_one_extend(int *readstate, int newstate,
WARN_ON_ONCE(idxold2 < 0);
WARN_ON_ONCE(idxold2 & ~RCUTORTURE_RDR_ALLBITS);
rtrsp->rt_readstate = newstate;
if (IS_ENABLED(CONFIG_RCU_TORTURE_TEST_LOG_CPU))
rtrsp->rt_cpu = raw_smp_processor_id();
/* First, put new protection in place to avoid critical-section gap. */
if (statesnew & RCUTORTURE_RDR_BH)
@ -3559,8 +3562,10 @@ rcu_torture_cleanup(void)
err_segs[i].rt_delay_us);
firsttime = 0;
}
pr_cont("%s\n",
err_segs[i].rt_preempted ? "preempted" : "");
pr_cont("%s", err_segs[i].rt_preempted ? "preempted" : "");
if (IS_ENABLED(CONFIG_RCU_TORTURE_TEST_LOG_CPU))
pr_cont(" CPU %d", err_segs[i].rt_cpu);
pr_cont("\n");
}
}