Christoph Lameter demonstrated very poor scalability on the SGI 512-way, with
a many-threaded application which concurrently initializes different parts of
a large anonymous area.
This patch corrects that, by using a separate spinlock per page table page, to
guard the page table entries in that page, instead of using the mm's single
page_table_lock. (But even then, page_table_lock is still used to guard page
table allocation, and anon_vma allocation.)
In this implementation, the spinlock is tucked inside the struct page of the
page table page: with a BUILD_BUG_ON in case it overflows - which it would in
the case of 32-bit PA-RISC with spinlock debugging enabled.
Splitting the lock is not quite for free: another cacheline access. Ideally,
I suppose we would use split ptlock only for multi-threaded processes on
multi-cpu machines; but deciding that dynamically would have its own costs.
So for now enable it by config, at some number of cpus - since the Kconfig
language doesn't support inequalities, let preprocessor compare that with
NR_CPUS. But I don't think it's worth being user-configurable: for good
testing of both split and unsplit configs, split now at 4 cpus, and perhaps
change that to 8 later.
There is a benefit even for singly threaded processes: kswapd can be attacking
one part of the mm while another part is busy faulting.
Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Final step in pushing down common core's page_table_lock. follow_page no
longer wants caller to hold page_table_lock, uses pte_offset_map_lock itself;
and so no page_table_lock is taken in get_user_pages itself.
But get_user_pages (and get_futex_key) do then need follow_page to pin the
page for them: take Daniel's suggestion of bitflags to follow_page.
Need one for WRITE, another for TOUCH (it was the accessed flag before:
vanished along with check_user_page_readable, but surely get_numa_maps is
wrong to mark every page it finds as accessed), another for GET.
And another, ANON to dispose of untouched_anonymous_page: it seems silly for
that to descend a second time, let follow_page observe if there was no page
table and return ZERO_PAGE if so. Fix minor bug in that: check VM_LOCKED -
make_pages_present ought to make readonly anonymous present.
Give get_numa_maps a cond_resched while we're there.
Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Second step in pushing down the page_table_lock. Remove the temporary
bridging hack from __pud_alloc, __pmd_alloc, __pte_alloc: expect callers not
to hold page_table_lock, whether it's on init_mm or a user mm; take
page_table_lock internally to check if a racing task already allocated.
Convert their callers from common code. But avoid coming back to change them
again later: instead of moving the spin_lock(&mm->page_table_lock) down,
switch over to new macros pte_alloc_map_lock and pte_unmap_unlock, which
encapsulate the mapping+locking and unlocking+unmapping together, and in the
end may use alternatives to the mm page_table_lock itself.
These callers all hold mmap_sem (some exclusively, some not), so at no level
can a page table be whipped away from beneath them; and pte_alloc uses the
"atomic" pmd_present to test whether it needs to allocate. It appears that on
all arches we can safely descend without page_table_lock.
Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
update_mem_hiwater has attracted various criticisms, in particular from those
concerned with mm scalability. Originally it was called whenever rss or
total_vm got raised. Then many of those callsites were replaced by a timer
tick call from account_system_time. Now Frank van Maarseveen reports that to
be found inadequate. How about this? Works for Frank.
Replace update_mem_hiwater, a poor combination of two unrelated ops, by macros
update_hiwater_rss and update_hiwater_vm. Don't attempt to keep
mm->hiwater_rss up to date at timer tick, nor every time we raise rss (usually
by 1): those are hot paths. Do the opposite, update only when about to lower
rss (usually by many), or just before final accounting in do_exit. Handle
mm->hiwater_vm in the same way, though it's much less of an issue. Demand
that whoever collects these hiwater statistics do the work of taking the
maximum with rss or total_vm.
And there has been no collector of these hiwater statistics in the tree. The
new convention needs an example, so match Frank's usage by adding a VmPeak
line above VmSize to /proc/<pid>/status, and also a VmHWM line above VmRSS
(High-Water-Mark or High-Water-Memory).
There was a particular anomaly during mremap move, that hiwater_vm might be
captured too high. A fleeting such anomaly remains, but it's quickly
corrected now, whereas before it would stick.
What locking? None: if the app is racy then these statistics will be racy,
it's not worth any overhead to make them exact. But whenever it suits,
hiwater_vm is updated under exclusive mmap_sem, and hiwater_rss under
page_table_lock (for now) or with preemption disabled (later on): without
going to any trouble, minimize the time between reading current values and
updating, to minimize those occasions when a racing thread bumps a count up
and back down in between.
Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Remove PageReserved() calls from core code by tightening VM_RESERVED
handling in mm/ to cover PageReserved functionality.
PageReserved special casing is removed from get_page and put_page.
All setting and clearing of PageReserved is retained, and it is now flagged
in the page_alloc checks to help ensure we don't introduce any refcount
based freeing of Reserved pages.
MAP_PRIVATE, PROT_WRITE of VM_RESERVED regions is tentatively being
deprecated. We never completely handled it correctly anyway, and is be
reintroduced in future if required (Hugh has a proof of concept).
Once PageReserved() calls are removed from kernel/power/swsusp.c, and all
arch/ and driver code, the Set and Clear calls, and the PG_reserved bit can
be trivially removed.
Last real user of PageReserved is swsusp, which uses PageReserved to
determine whether a struct page points to valid memory or not. This still
needs to be addressed (a generic page_is_ram() should work).
A last caveat: the ZERO_PAGE is now refcounted and managed with rmap (and
thus mapcounted and count towards shared rss). These writes to the struct
page could cause excessive cacheline bouncing on big systems. There are a
number of ways this could be addressed if it is an issue.
Signed-off-by: Nick Piggin <npiggin@suse.de>
Refcount bug fix for filemap_xip.c
Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
One anomaly remains from when Andrea rationalized the responsibilities of
mmap_sem and page_table_lock: in dup_mmap we add vmas to the child holding its
page_table_lock, but not the mmap_sem which normally guards the vma list and
rbtree. Which could be an issue for unuse_mm: though since it just walks down
the list (today with page_table_lock, tomorrow not), it's probably okay. Will
need a memory barrier? Oh, keep it simple, Nick and I agreed, no harm in
taking child's mmap_sem here.
Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Use the parent's oldmm throughout dup_mmap, instead of perversely going back
to current->mm. (Can you hear the sigh of relief from those mpnts? Usually I
squash them, but not today.)
Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
I was lazy when we added anon_rss, and chose to change as few places as
possible. So currently each anonymous page has to be counted twice, in rss
and in anon_rss. Which won't be so good if those are atomic counts in some
configurations.
Change that around: keep file_rss and anon_rss separately, and add them
together (with get_mm_rss macro) when the total is needed - reading two
atomics is much cheaper than updating two atomics. And update anon_rss
upfront, typically in memory.c, not tucked away in page_add_anon_rmap.
Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
How is anon_rss initialized? In dup_mmap, and by mm_alloc's memset; but
that's not so good if an mm_counter_t is a special type. And how is rss
initialized? By set_mm_counter, all over the place. Come on, we just need to
initialize them both at once by set_mm_counter in mm_init (which follows the
memcpy when forking).
Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
The original vm_stat_account has fallen into disuse, with only one user, and
only one user of vm_stat_unaccount. It's easier to keep track if we convert
them all to __vm_stat_account, then free it from its __shackles.
Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
frv, sh64, ia64 and sparc64 do not have do_settimeofday() exported (the
last two are using variant in kernel/time.c). Exports added to match
the rest of architectures.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
exit_signal() (called from copy_process's error path) should decrement
->signal->live, otherwise forking process will miss 'group_dead' in
do_exit().
Signed-off-by: Oleg Nesterov <oleg@tv-sign.ru>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This just makes sure that a thread's expiry times can't get reset after
it clears them in do_exit.
This is what allowed us to re-introduce the stricter BUG_ON() check in
a362f463a6.
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This reverts commit 3de463c7d9.
Roland has another patch that allows us to leave the BUG_ON() in place
by just making sure that the condition it tests for really is always
true.
That goes in next.
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
There's a silly off-by-one error in the code that updates the expiration
of posix CPU timers, causing them to not be properly updated when they
hit exactly on their expiration time (which should be the normal case).
This causes them to then fire immediately again, and only _then_ get
properly updated.
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
With CONFIG_SMP=n:
*** Warning: "cpu_online_map" [drivers/firmware/dcdbas.ko] undefined!
due to set_cpus_allowed().
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This might be harmless, but looks like a race from code inspection (I
was unable to trigger it). I must admit, I don't understand why we
can't return TIMER_RETRY after 'spin_unlock(&p->sighand->siglock)'
without doing bump_cpu_timer(), but this is what original code does.
posix_cpu_timer_set:
read_lock(&tasklist_lock);
spin_lock(&p->sighand->siglock);
list_del_init(&timer->it.cpu.entry);
spin_unlock(&p->sighand->siglock);
We are probaly deleting the timer from run_posix_cpu_timers's 'firing'
local list_head while run_posix_cpu_timers() does list_for_each_safe.
Various bad things can happen, for example we can just delete this timer
so that list_for_each() will not notice it and run_posix_cpu_timers()
will not reset '->firing' flag. In that case,
....
if (timer->it.cpu.firing) {
read_unlock(&tasklist_lock);
timer->it.cpu.firing = -1;
return TIMER_RETRY;
}
sys_timer_settime() goes to 'retry:', calls posix_cpu_timer_set() again,
it returns TIMER_RETRY ...
Signed-off-by: Oleg Nesterov <oleg@tv-sign.ru>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
do_exit() clears ->it_##clock##_expires, but nothing prevents
another cpu to attach the timer to exiting process after that.
After exit_notify() does 'write_unlock_irq(&tasklist_lock)' and
before do_exit() calls 'schedule() local timer interrupt can find
tsk->exit_state != 0. If that state was EXIT_DEAD (or another cpu
does sys_wait4) interrupted task has ->signal == NULL.
At this moment exiting task has no pending cpu timers, they were cleaned
up in __exit_signal()->posix_cpu_timers_exit{,_group}(), so we can just
return from irq.
Signed-off-by: Oleg Nesterov <oleg@tv-sign.ru>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
1. cleanup_timers() sets timer->task = NULL under tasklist + ->sighand locks.
That means that this code in posix_cpu_timer_del() and posix_cpu_timer_set()
lock_timer(timer);
if (timer->task == NULL)
return;
read_lock(tasklist);
put_task_struct(timer->task)
is racy. With this patch timer->task modified and accounted only under
timer->it_lock. Sadly, this means that dead task_struct won't be freed
until timer deleted or armed.
2. run_posix_cpu_timers() collects expired timers into local list under
tasklist + ->sighand again. That means that posix_cpu_timer_del()
should check timer->it.cpu.firing under these locks too.
Signed-off-by: Oleg Nesterov <oleg@tv-sign.ru>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Bursty timers aren't good for anybody, very much including latency for
other programs when we trigger lots of timers in interrupt context. So
set a random limit, after which we'll handle the rest on the next timer
tick.
Noted by Oleg Nesterov <oleg@tv-sign.ru>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
