SRCU updates for v5.9
Subsystems (2)
Paul E. McKenney
maintainercover SRCU updates for v5.9
Hello!
This series provides SRCU updates.
1. Fix a typo in comment "amoritized"->"amortized", courtesy of
Ethon Paul.
2. Avoid local_irq_save() before acquiring spinlock_t, courtesy
of Sebastian Andrzej Siewior.
3. Remove KCSAN stubs.
Thanx, Paul
------------------------------------------------------------------------
srcutree.c | 29 ++++++++--------------------- 1 file changed, 8 insertions(+), 21 deletions(-)
[email protected]
maintainerPATCH 1/3 srcu: Fix a typo in comment "amoritized"->"amortized"
From: Ethon Paul <[email protected]>
This commit fixes a typo in a comment.
--- kernel/rcu/srcutree.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
kernel/rcu/srcutree.c +1 -1
@@ -766,7 +766,7 @@ static void srcu_flip(struct srcu_struct *ssp)* it, if this function was preempted for enough time for the counters * to wrap, it really doesn't matter whether or not we expedite the grace * period. The extra overhead of a needlessly expedited grace period is -* negligible when amoritized over that time period, and the extra latency +* negligible when amortized over that time period, and the extra latency * of a needlessly non-expedited grace period is similarly negligible. */ static bool srcu_might_be_idle (struct srcu_struct *ssp)
Signature · 1 line
2.9.5
[email protected]
maintainerPATCH 2/3 srcu: Avoid local_irq_save() before acquiring spinlock_t
In reply to Paul E. McKenneyFrom: Sebastian Andrzej Siewior <[email protected]>
SRCU disables interrupts to get a stable per-CPU pointer and then acquires the spinlock which is in the per-CPU data structure. The release uses spin_unlock_irqrestore(). While this is correct on a non-RT kernel, this conflicts with the RT semantics because the spinlock is converted to a 'sleeping' spinlock. Sleeping locks can obviously not be acquired with interrupts disabled.
Acquire the per-CPU pointer `ssp->sda' without disabling preemption and then acquire the spinlock_t of the per-CPU data structure. The lock will ensure that the data is consistent.
The added call to check_init_srcu_struct() is now needed because a statically defined srcu_struct may remain uninitialized until this point and the newly introduced locking operation requires an initialized spinlock_t.
This change was tested for four hours with 8*SRCU-N and 8*SRCU-P without causing any warnings.
--- kernel/rcu/srcutree.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-)
kernel/rcu/srcutree.c +7 -7
@@ -777,14 +777,15 @@ static bool srcu_might_be_idle(struct srcu_struct *ssp)unsigned long t;unsigned long tlast; +check_init_srcu_struct (ssp);/* If the local srcu_data structure has callbacks, not idle. */ -local_irq_save (flags); - sdp =this_cpu_ptr (ssp->sda); + sdp =raw_cpu_ptr (ssp->sda); +spin_lock_irqsave_rcu_node (sdp, flags);if (rcu_segcblist_pend_cbs (&sdp->srcu_cblist)) { -local_irq_restore (flags); +spin_unlock_irqrestore_rcu_node (sdp, flags);return false;/* Callbacks already present, so not idle. */ } -local_irq_restore (flags); +spin_unlock_irqrestore_rcu_node (sdp, flags);/* * No local callbacks, so probabalistically probe global state.
@@ -864,9 +865,8 @@ static void __call_srcu(struct srcu_struct *ssp, struct rcu_head *rhp, } rhp->func = func; idx =srcu_read_lock (ssp); -local_irq_save (flags); - sdp =this_cpu_ptr (ssp->sda); -spin_lock_rcu_node (sdp); + sdp =raw_cpu_ptr (ssp->sda); +spin_lock_irqsave_rcu_node (sdp, flags);rcu_segcblist_enqueue (&sdp->srcu_cblist, rhp);rcu_segcblist_advance (&sdp->srcu_cblist,rcu_seq_current (&ssp->srcu_gp_seq));
Signature · 1 line
2.9.5
[email protected]
maintainerPATCH 3/3 srcu: Remove KCSAN stubs
In reply to Paul E. McKenneyFrom: "Paul E. McKenney" <[email protected]>
KCSAN is now in mainline, so this commit removes the stubs for the data_race(), ASSERT_EXCLUSIVE_WRITER(), and ASSERT_EXCLUSIVE_ACCESS() macros.
--- kernel/rcu/srcutree.c | 13 ------------- 1 file changed, 13 deletions(-)
kernel/rcu/srcutree.c +0 -13
@@ -29,19 +29,6 @@#include "rcu.h" #include "rcu_segcblist.h" -#ifndef data_race -#define data_race (expr) \ - ({ \ - expr; \ - }) -#endif -#ifndef ASSERT_EXCLUSIVE_WRITER -#define ASSERT_EXCLUSIVE_WRITER (var) do { } while (0) -#endif -#ifndef ASSERT_EXCLUSIVE_ACCESS -#define ASSERT_EXCLUSIVE_ACCESS (var) do { } while (0) -#endif -/* Holdoff in nanoseconds for auto-expediting. */ #define DEFAULT_SRCU_EXP_HOLDOFF (25 * 1000)static ulong exp_holdoff =DEFAULT_SRCU_EXP_HOLDOFF ;
Signature · 1 line
2.9.5
kernel test robot
build botRe: [PATCH tip/core/rcu 3/3] srcu: Remove KCSAN stubs
I love your patch! Yet something to improve:
Hi,
I love your patch! Yet something to improve:
[auto build test ERROR on tip/core/rcu]
url: https://github.com/0day-ci/linux/commits/paulmck-kernel-org/srcu-Fix-a-typo-in-comment-amoritized-amortized/20200623-085757 base: https://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git cb3cb6733fbd8fd8d2c716095fdca42dadba2063 config: x86_64-rhel-7.6 (attached as .config) compiler: gcc-9 (Debian 9.3.0-13) 9.3.0 reproduce (this is a W=1 build): # save the attached .config to linux build tree make W=1 ARCH=x86_64
If you fix the issue, kindly add following tag as appropriate
All errors (new ones prefixed by >>):
kernel/rcu/srcutree.c: In function 'srcu_torture_stats_print':
>> kernel/rcu/srcutree.c:1271:8: error: implicit declaration of function 'data_race' [-Werror=implicit-function-declaration]
1271 | u0 = data_race(sdp->srcu_unlock_count[!idx]);
| ^~~~~~~~~cc1: some warnings being treated as errors
vim +/data_race +1271 kernel/rcu/srcutree.c
1f4f6da1c80905 Paul E. McKenney 2017-04-21 1254
aacb5d91ab1bfb Paul E. McKenney 2018-10-28 1255 void srcu_torture_stats_print(struct srcu_struct *ssp, char *tt, char *tf)
115a1a5285664f Paul E. McKenney 2017-05-22 1256 {
115a1a5285664f Paul E. McKenney 2017-05-22 1257 int cpu;
115a1a5285664f Paul E. McKenney 2017-05-22 1258 int idx;
ac3748c6042660 Paul E. McKenney 2017-05-22 1259 unsigned long s0 = 0, s1 = 0;
115a1a5285664f Paul E. McKenney 2017-05-22 1260
aacb5d91ab1bfb Paul E. McKenney 2018-10-28 1261 idx = ssp->srcu_idx & 0x1;
52e17ba1d063ab Paul E. McKenney 2018-06-19 1262 pr_alert("%s%s Tree SRCU g%ld per-CPU(idx=%d):",
aacb5d91ab1bfb Paul E. McKenney 2018-10-28 1263 tt, tf, rcu_seq_current(&ssp->srcu_gp_seq), idx);
115a1a5285664f Paul E. McKenney 2017-05-22 1264 for_each_possible_cpu(cpu) {
115a1a5285664f Paul E. McKenney 2017-05-22 1265 unsigned long l0, l1;
115a1a5285664f Paul E. McKenney 2017-05-22 1266 unsigned long u0, u1;
115a1a5285664f Paul E. McKenney 2017-05-22 1267 long c0, c1;
5ab07a8df4d6c9 Paul E. McKenney 2018-05-22 1268 struct srcu_data *sdp;
115a1a5285664f Paul E. McKenney 2017-05-22 1269
aacb5d91ab1bfb Paul E. McKenney 2018-10-28 1270 sdp = per_cpu_ptr(ssp->sda, cpu);
b68c6146512d92 Paul E. McKenney 2020-01-03 @1271 u0 = data_race(sdp->srcu_unlock_count[!idx]);
b68c6146512d92 Paul E. McKenney 2020-01-03 1272 u1 = data_race(sdp->srcu_unlock_count[idx]);
115a1a5285664f Paul E. McKenney 2017-05-22 1273
115a1a5285664f Paul E. McKenney 2017-05-22 1274 /*
115a1a5285664f Paul E. McKenney 2017-05-22 1275 * Make sure that a lock is always counted if the corresponding
115a1a5285664f Paul E. McKenney 2017-05-22 1276 * unlock is counted.
115a1a5285664f Paul E. McKenney 2017-05-22 1277 */
115a1a5285664f Paul E. McKenney 2017-05-22 1278 smp_rmb();
115a1a5285664f Paul E. McKenney 2017-05-22 1279
b68c6146512d92 Paul E. McKenney 2020-01-03 1280 l0 = data_race(sdp->srcu_lock_count[!idx]);
b68c6146512d92 Paul E. McKenney 2020-01-03 1281 l1 = data_race(sdp->srcu_lock_count[idx]);
115a1a5285664f Paul E. McKenney 2017-05-22 1282
115a1a5285664f Paul E. McKenney 2017-05-22 1283 c0 = l0 - u0;
115a1a5285664f Paul E. McKenney 2017-05-22 1284 c1 = l1 - u1;
7e210a653ec944 Paul E. McKenney 2019-06-28 1285 pr_cont(" %d(%ld,%ld %c)",
7e210a653ec944 Paul E. McKenney 2019-06-28 1286 cpu, c0, c1,
7e210a653ec944 Paul E. McKenney 2019-06-28 1287 "C."[rcu_segcblist_empty(&sdp->srcu_cblist)]);
ac3748c6042660 Paul E. McKenney 2017-05-22 1288 s0 += c0;
ac3748c6042660 Paul E. McKenney 2017-05-22 1289 s1 += c1;
115a1a5285664f Paul E. McKenney 2017-05-22 1290 }
ac3748c6042660 Paul E. McKenney 2017-05-22 1291 pr_cont(" T(%ld,%ld)\n", s0, s1);
115a1a5285664f Paul E. McKenney 2017-05-22 1292 }
115a1a5285664f Paul E. McKenney 2017-05-22 1293 EXPORT_SYMBOL_GPL(srcu_torture_stats_print);
115a1a5285664f Paul E. McKenney 2017-05-22 1294 ---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all(a)lists.01.org
Paul E. McKenney
maintainerRe: [PATCH tip/core/rcu 3/3] srcu: Remove KCSAN stubs
Quoted 17 lines · kernel test robot
Hi,
I love your patch! Yet something to improve:
[auto build test ERROR on tip/core/rcu]
url: https://github.com/0day-ci/linux/commits/paulmck-kernel-org/srcu-Fix-a-typo-in-comment-amoritized-amortized/20200623-085757 base: https://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git cb3cb6733fbd8fd8d2c716095fdca42dadba2063 config: x86_64-rhel-7.6 (attached as .config) compiler: gcc-9 (Debian 9.3.0-13) 9.3.0 reproduce (this is a W=1 build): # save the attached .config to linux build tree make W=1 ARCH=x86_64If you fix the issue, kindly add following tag as appropriate
Reported-by kernel test robot <[email protected]>
Quoted 7 lines
All errors (new ones prefixed by >>):
kernel/rcu/srcutree.c: In function 'srcu_torture_stats_print':Quoted 1 line
Quoted 1 line
kernel/rcu/srcutree.c:1271:8: error: implicit declaration of function 'data_race' [-Werror=implicit-function-declaration]1271 | u0 = data_race(sdp->srcu_unlock_count[!idx]); | ^~~~~~~~~cc1: some warnings being treated as errors
This definition is in mainline, just not in v5.7. So if you apply this patch to v5.8-rc1, it should work fine. I have nevertheless moved this patch to my v5.10 stack to avoid confusion.
Thanx, Paul
Quoted 47 lines
vim +/data_race +1271
kernel/rcu/srcutree.c1f4f6da1c80905 Paul E. McKenney 2017-04-21 1254 aacb5d91ab1bfb Paul E. McKenney 2018-10-28 1255 void srcu_torture_stats_print(struct srcu_struct *ssp, char *tt, char *tf) 115a1a5285664f Paul E. McKenney 2017-05-22 1256 { 115a1a5285664f Paul E. McKenney 2017-05-22 1257 int cpu; 115a1a5285664f Paul E. McKenney 2017-05-22 1258 int idx; ac3748c6042660 Paul E. McKenney 2017-05-22 1259 unsigned long s0 = 0, s1 = 0; 115a1a5285664f Paul E. McKenney 2017-05-22 1260 aacb5d91ab1bfb Paul E. McKenney 2018-10-28 1261 idx = ssp->srcu_idx & 0x1; 52e17ba1d063ab Paul E. McKenney 2018-06-19 1262 pr_alert("%s%s Tree SRCU g%ld per-CPU(idx=%d):", aacb5d91ab1bfb Paul E. McKenney 2018-10-28 1263 tt, tf, rcu_seq_current(&ssp->srcu_gp_seq), idx); 115a1a5285664f Paul E. McKenney 2017-05-22 1264 for_each_possible_cpu(cpu) { 115a1a5285664f Paul E. McKenney 2017-05-22 1265 unsigned long l0, l1; 115a1a5285664f Paul E. McKenney 2017-05-22 1266 unsigned long u0, u1; 115a1a5285664f Paul E. McKenney 2017-05-22 1267 long c0, c1; 5ab07a8df4d6c9 Paul E. McKenney 2018-05-22 1268 struct srcu_data *sdp; 115a1a5285664f Paul E. McKenney 2017-05-22 1269 aacb5d91ab1bfb Paul E. McKenney 2018-10-28 1270 sdp = per_cpu_ptr(ssp->sda, cpu); b68c6146512d92 Paul E. McKenney 2020-01-03 @1271 u0 = data_race(sdp->srcu_unlock_count[!idx]); b68c6146512d92 Paul E. McKenney 2020-01-03 1272 u1 = data_race(sdp->srcu_unlock_count[idx]); 115a1a5285664f Paul E. McKenney 2017-05-22 1273 115a1a5285664f Paul E. McKenney 2017-05-22 1274 /* 115a1a5285664f Paul E. McKenney 2017-05-22 1275 * Make sure that a lock is always counted if the corresponding 115a1a5285664f Paul E. McKenney 2017-05-22 1276 * unlock is counted. 115a1a5285664f Paul E. McKenney 2017-05-22 1277 */ 115a1a5285664f Paul E. McKenney 2017-05-22 1278 smp_rmb(); 115a1a5285664f Paul E. McKenney 2017-05-22 1279 b68c6146512d92 Paul E. McKenney 2020-01-03 1280 l0 = data_race(sdp->srcu_lock_count[!idx]); b68c6146512d92 Paul E. McKenney 2020-01-03 1281 l1 = data_race(sdp->srcu_lock_count[idx]); 115a1a5285664f Paul E. McKenney 2017-05-22 1282 115a1a5285664f Paul E. McKenney 2017-05-22 1283 c0 = l0 - u0; 115a1a5285664f Paul E. McKenney 2017-05-22 1284 c1 = l1 - u1; 7e210a653ec944 Paul E. McKenney 2019-06-28 1285 pr_cont(" %d(%ld,%ld %c)", 7e210a653ec944 Paul E. McKenney 2019-06-28 1286 cpu, c0, c1, 7e210a653ec944 Paul E. McKenney 2019-06-28 1287 "C."[rcu_segcblist_empty(&sdp->srcu_cblist)]); ac3748c6042660 Paul E. McKenney 2017-05-22 1288 s0 += c0; ac3748c6042660 Paul E. McKenney 2017-05-22 1289 s1 += c1; 115a1a5285664f Paul E. McKenney 2017-05-22 1290 } ac3748c6042660 Paul E. McKenney 2017-05-22 1291 pr_cont(" T(%ld,%ld)\n", s0, s1); 115a1a5285664f Paul E. McKenney 2017-05-22 1292 } 115a1a5285664f Paul E. McKenney 2017-05-22 1293 EXPORT_SYMBOL_GPL(srcu_torture_stats_print); 115a1a5285664f Paul E. McKenney 2017-05-22 1294---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all(a)lists.01.org