Messages in this thread | | | Date | Mon, 26 Feb 2018 15:37:35 +0900 | From | Sergey Senozhatsky <> | Subject | Re: [PATCH v3] printk: Relocate wake_klogd check close to the end of console_unlock() |
| |
On (02/19/18 17:01), Petr Mladek wrote: [..] > - raw_spin_lock(&logbuf_lock); > retry = console_seq != log_next_seq; > + /* > + * Check whether userland needs notification. Do this only when really > + * leaving to avoid race with console_trylock_spinning(). > + */ > + if (seen_seq != log_next_seq && !retry) { > + wake_klogd = true; > + seen_seq = log_next_seq; > + }
Let's add the "why" part. This "!retry" might be hard to understand. We are looking at
- CPUa is about to leave console_unlock() - printk on CPUb appends a new message - CPUa detects that `console_seq != log_next_seq', updates `seen_seq' - printk on CPUb is getting preempted - CPUa re-takes the console_sem via retry path - printk CPUb is becoming TASK_RUNNING again - it now spins for console_sem, since we have an active console_sem owner - CPUa detects that there is a console_sem waiter, so it offloads the printing task, without ever waking up klogd
Either we can have that complex "seen_seq != log_next_seq && !retry" check - or we simply can add
if (console_lock_spinning_disable_and_check()) { printk_safe_exit_irqrestore(flags); if (wake_klogd) wake_up_klogd(); }
to the offloading return path.
The later is *may be* simpler to follow. The rule is: every !console_suspend and !cant-use-consoles return path from console_unlock() must wake_up_klogd() [if needed].
-ss
| |