Linux kernel mirror (for testing)
git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
kernel
os
linux
tick/nohz: Update obsolete comments
Some comments are obsolete enough to assume that IRQ exit restarts the
tick in idle or RCU is turned on at the same time as the tick, among
other details.
Update them and add more.
Signed-off-by: Frederic Weisbecker <frederic@kernel.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20230912104406.312185-3-frederic@kernel.org
authored by
Frederic Weisbecker
and committed by
Thomas Gleixner
822deeed
dba428a6
+36
-10
+36
-10
kernel/time/tick-sched.c
+36
-10
kernel/time/tick-sched.c
···
1175
1175
}
1176
1176
1177
1177
/**
1178
-
* tick_nohz_irq_exit - update next tick event from interrupt exit
1178
+
* tick_nohz_irq_exit - Notify the tick about IRQ exit
1179
1179
*
1180
-
* When an interrupt fires while we are idle and it doesn't cause
1181
-
* a reschedule, it may still add, modify or delete a timer, enqueue
1182
-
* an RCU callback, etc...
1183
-
* So we need to re-calculate and reprogram the next tick event.
1180
+
* A timer may have been added/modified/deleted either by the current IRQ,
1181
+
* or by another place using this IRQ as a notification. This IRQ may have
1182
+
* also updated the RCU callback list. These events may require a
1183
+
* re-evaluation of the next tick. Depending on the context:
1184
+
*
1185
+
* 1) If the CPU is idle and no resched is pending, just proceed with idle
1186
+
* time accounting. The next tick will be re-evaluated on the next idle
1187
+
* loop iteration.
1188
+
*
1189
+
* 2) If the CPU is nohz_full:
1190
+
*
1191
+
* 2.1) If there is any tick dependency, restart the tick if stopped.
1192
+
*
1193
+
* 2.2) If there is no tick dependency, (re-)evaluate the next tick and
1194
+
* stop/update it accordingly.
1184
1195
*/
1185
1196
void tick_nohz_irq_exit(void)
1186
1197
{
···
1341
1330
}
1342
1331
1343
1332
/**
1344
-
* tick_nohz_idle_exit - restart the idle tick from the idle task
1333
+
* tick_nohz_idle_exit - Update the tick upon idle task exit
1345
1334
*
1346
-
* Restart the idle tick when the CPU is woken up from idle
1347
-
* This also exit the RCU extended quiescent state. The CPU
1348
-
* can use RCU again after this function is called.
1335
+
* When the idle task exits, update the tick depending on the
1336
+
* following situations:
1337
+
*
1338
+
* 1) If the CPU is not in nohz_full mode (most cases), then
1339
+
* restart the tick.
1340
+
*
1341
+
* 2) If the CPU is in nohz_full mode (corner case):
1342
+
* 2.1) If the tick can be kept stopped (no tick dependencies)
1343
+
* then re-eavaluate the next tick and try to keep it stopped
1344
+
* as long as possible.
1345
+
* 2.2) If the tick has dependencies, restart the tick.
1346
+
*
1349
1347
*/
1350
1348
void tick_nohz_idle_exit(void)
1351
1349
{
···
1384
1364
}
1385
1365
1386
1366
/*
1387
-
* The nohz low res interrupt handler
1367
+
* In low-resolution mode, the tick handler must be implemented directly
1368
+
* at the clockevent level. hrtimer can't be used instead because its
1369
+
* infrastructure actually relies on the tick itself as a backend in
1370
+
* low-resolution mode (see hrtimer_run_queues()).
1371
+
*
1372
+
* This low-resolution handler still makes use of some hrtimer APIs meanwhile
1373
+
* for commodity with expiration calculation and forwarding.
1388
1374
*/
1389
1375
static void tick_nohz_lowres_handler(struct clock_event_device *dev)
1390
1376
{