- 28 Oct, 2010 40 commits
-
-
KAMEZAWA Hiroyuki authored
/proc/stat shows the total number of all interrupts to each cpu. But when the number of IRQs are very large, it take very long time and 'cat /proc/stat' takes more than 10 secs. This is because sum of all irq events are counted when /proc/stat is read. This patch adds "sum of all irq" counter percpu and reduce read costs. The cost of reading /proc/stat is important because it's used by major applications as 'top', 'ps', 'w', etc.... A test on a mechin (4096cpu, 256 nodes, 4592 irqs) shows %time cat /proc/stat > /dev/null Before Patch: 12.627 sec After Patch: 2.459 sec Signed-off-by:
KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Tested-by:
Jack Steiner <steiner@sgi.com> Acked-by:
Andrew Morton <akpm@linux-foundation.org> Signed-off-by:
Linus Torvalds <torvalds@linux-foundation.org>
-
Davidlohr Bueso authored
The length of the BLOCK_IPOLL string is making i's value be printed too far to the right. This patch fixes this and makes the output a bit neater. Currently: CPU0 HI: 0 TIMER: 599792 NET_TX: 2 NET_RX: 6 BLOCK: 80807 BLOCK_IOPOLL: 0 TASKLET: 20012 SCHED: 0 HRTIMER: 63 RCU: 619279 With patch: CPU0 HI: 0 TIMER: 585582 NET_TX: 2 NET_RX: 6 BLOCK: 80320 BLOCK_IOPOLL: 0 TASKLET: 19287 SCHED: 0 HRTIMER: 62 RCU: 604441 Signed-off-by:Davidlohr Bueso <dave@gnu.org> Acked-by:
Keika Kobayashi <kobayashi.kk@ncos.nec.co.jp> Signed-off-by:
-
Nikanth Karthikesan authored
Document /proc/pid/pagemap in Documentation/filesystems/proc.txt Signed-off-by:
Nikanth Karthikesan <knikanth@suse.de> Cc: Richard Guenther <rguenther@suse.de> Cc: Balbir Singh <balbir@linux.vnet.ibm.com> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Acked-by:
Matt Mackall <mpm@selenic.com> Signed-off-by:
-
Nikanth Karthikesan authored
Export the number of anonymous pages in a mapping via smaps. Even the private pages in a mapping backed by a file, would be marked as anonymous, when they are modified. Export this information to user-space via smaps. Exporting this count will help gdb to make a better decision on which areas need to be dumped in its coredump; and should be useful to others studying the memory usage of a process. Signed-off-by:
Hugh Dickins <hughd@google.com> Reviewed-by:
KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Cc: Matt Mackall <mpm@selenic.com> Signed-off-by:
-
Namhyung Kim authored
find_new_reaper() releases and regrabs tasklist_lock but was missing proper annotations. Add it. This remove following sparse warning: warning: context imbalance in 'find_new_reaper' - unexpected unlock Signed-off-by:
Namhyung Kim <namhyung@gmail.com> Acked-by:
Roland McGrath <roland@redhat.com> Signed-off-by:
-
Roland McGrath authored
The userland ELF tools have been coping with partial-segments core files for a few years now. Multiple distro builds are now setting this option. It behooves everyone who ever deals with core files to have more info dumped in there, especially as more and more people's compilers are producing build IDs. Make it the default. Anyone using older tools confused by these core files can configure this option off, or just change /proc/PID/coredump_filter after boot. Signed-off-by:
-
Xiaotian Feng authored
We met a parameter truncated issue, consider following: > echo "|/root/core_pattern_pipe_test %p /usr/libexec/blah-blah-blah \ %s %c %p %u %g 11 12345678901234567890123456789012345678 %t" > \ /proc/sys/kernel/core_pattern This is okay because the strings is less than CORENAME_MAX_SIZE. "cat /proc/sys/kernel/core_pattern" shows the whole string. but after we run core_pattern_pipe_test in man page, we found last parameter was truncated like below: argc[10]=<12807486> The root cause is core_pattern allows % specifiers, which need to be replaced during parse time, but the replace may expand the strings to larger than CORENAME_MAX_SIZE. So if the last parameter is % specifiers, the replace code is using snprintf(out_ptr, out_end - out_ptr, ...), this will write out of corename array. [akpm@linux-foundation.org: coding-style fixes] Signed-off-by:Xiaotian Feng <dfeng@redhat.com> Cc: Alexander Viro <viro@zeniv.linux.org.uk> Cc: Oleg Nesterov <oleg@redhat.com> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Reviewed-by:
Neil Horman <nhorman@tuxdriver.com> Cc: Roland McGrath <roland@redhat.com> Signed-off-by:
-
KOSAKI Motohiro authored
Oleg Nesterov pointed out we have to prevent multiple-threads-inside-exec itself and we can reuse ->cred_guard_mutex for it. Yes, concurrent execve() has no worth. Let's move ->cred_guard_mutex from task_struct to signal_struct. It naturally prevent multiple-threads-inside-exec. Signed-off-by:
Oleg Nesterov <oleg@redhat.com> Acked-by:
David Howells <dhowells@redhat.com> Signed-off-by:
-
Namhyung Kim authored
ptrace_stop() releases and regrabs current->sighand->siglock but was missing proper annotation. Add it. Signed-off-by:
-
Namhyung Kim authored
lock_task_sighand() grabs sighand->siglock in case of returning non-NULL but unlock_task_sighand() releases it unconditionally. This leads sparse to complain about the lock context imbalance. Rename and wrap lock_task_sighand() using __cond_lock() macro to make sparse happy. Suggested-by:
Eric Dumazet <eric.dumazet@gmail.com> Signed-off-by:
-
Namhyung Kim authored
Use new 'datap' variable in order to remove unnecessary castings. Signed-off-by:
-
Namhyung Kim authored
Remove unnecessary castings using void pointer and fix copy_to_user() return value. Also add missing __user markup on the argument of arch_ptrctl(). Signed-off-by:
-
Namhyung Kim authored
Remove checking @addr less than 0 because @addr is now unsigned. Signed-off-by:
Chris Metcalf <cmetcalf@tilera.com> Signed-off-by:
-
Namhyung Kim authored
Factor out struct fps and remove redundant castings. Signed-off-by:
David S. Miller <davem@davemloft.net> Signed-off-by:
-
Namhyung Kim authored
Remove unnecessary castings and get rid of dummy pointer in favor of offsetof() macro in ptrace_32.c. Also use temporary variables and break long lines in order to improve readability. Signed-off-by:
-
Namhyung Kim authored
Remove unnecessary castings. Signed-off-by:
-
Namhyung Kim authored
Use new 'datavp' and 'datalp' variables in order to remove unnecessary castings. Signed-off-by:
-
Namhyung Kim authored
Add missing __user markup on the argument of put_user(). Signed-off-by:
-
Namhyung Kim authored
Use new 'datap' variable in order to remove unnecessary castings. Also remove checking @addr less than 0 because @addr is now unsigned. Signed-off-by:
-
Namhyung Kim authored
Use new 'addrp', 'datavp' and 'datalp' variables in order to remove unnecessary castings. Signed-off-by:
-
Namhyung Kim authored
Remove checking @addr greater than 0 because @addr is now unsigned. Signed-off-by:
-
Namhyung Kim authored
Use new 'regno', 'datap' variables in order to remove duplicated expressions and unnecessary castings. Alse remove checking @addr less than 0 because addr is now unsigned. Signed-off-by:
Greg Ungerer <gerg@uclinux.org> Signed-off-by:
-
Namhyung Kim authored
Use new 'regno', 'datap' variables in order to remove duplicated expressions and unnecessary castings. Signed-off-by:
Geert Uytterhoeven <geert@linux-m68k.org> Cc: Roman Zippel <zippel@linux-m68k.org> Signed-off-by:
-
Namhyung Kim authored
Use new 'datap' variable in order to remove duplicated castings. Signed-off-by:
-
Namhyung Kim authored
Use new 'regno', 'datap' variables in order to remove duplicated expressions and unnecessary castings. Alse remove checking @addr less than 0 because addr is now unsigned. Signed-off-by:
-
Namhyung Kim authored
Use new 'regno', 'datap' variables in order to remove duplicated expressions and unnecessary castings. Alse remove checking @addr less than 0 because addr is now unsigned. Signed-off-by:
-
Namhyung Kim authored
Use new 'regno' variable in order to remove redandunt expression and remove checking @addr less than 0 because @addr is now unsigned. Also update 'datap' on PTRACE_GET/SETREGS to fix a bug on arch-v10. Signed-off-by:
Mikael Starvik <starvik@axis.com> Cc: Jesper Nilsson <jesper.nilsson@axis.com> Signed-off-by:
-
Namhyung Kim authored
Change signature of get/put_reg() according to the change of arch_ptrace() and remove unnecessary castings. Signed-off-by:
Mike Frysinger <vapier@gentoo.org> Signed-off-by:
-
Namhyung Kim authored
use new 'datap' variable type of void pointer in order to remove unnecessary castings. Signed-off-by:
Haavard Skinnemoen <hskinnemoen@atmel.com> Cc: Hans-Christian Egtvedt <hans-christian.egtvedt@atmel.com> Signed-off-by:
-
Namhyung Kim authored
use new 'datap' variable in order to remove unnecessary castings. Signed-off-by:
-
Namhyung Kim authored
Remove checking @addr less than 0 because @addr is now unsigned and use new udescp variable in order to remove unnecessary castings. [akpm@linux-foundation.org: fix unused variable 'udescp'] Signed-off-by:
-
Namhyung Kim authored
Fix up the arguments to arch_ptrace() to take account of the fact that @addr and @data are now unsigned long rather than long as of a preceding patch in this series. Signed-off-by:
-
Namhyung Kim authored
Use new 'datavp' and 'datalp' variables to remove unnecesary castings. Signed-off-by:
-
Namhyung Kim authored
Since userspace API of ptrace syscall defines @addr and @data as void pointers, it would be more appropriate to define them as unsigned long in kernel. Therefore related functions are changed also. 'unsigned long' is typically used in other places in kernel as an opaque data type and that using this helps cleaning up a lot of warnings from sparse. Suggested-by:
Arnd Bergmann <arnd@arndb.de> Signed-off-by:
-
Namhyung Kim authored
exit_ptrace() releases and regrabs tasklist_lock but was missing proper annotation. Add it. Signed-off-by:
-
KAMEZAWA Hiroyuki authored
This patch extracts the core logic from mem_cgroup_update_file_mapped() as mem_cgroup_update_file_stat() and adds a wrapper. As a planned future update, memory cgroup has to count dirty pages to implement dirty_ratio/limit. And more, the number of dirty pages is required to kick flusher thread to start writeback. (Now, no kick.) This patch is preparation for it and makes other statistics implementation clearer. Just a clean up. Signed-off-by:
Balbir Singh <balbir@linux.vnet.ibm.com> Reviewed-by:
Greg Thelen <gthelen@google.com> Cc: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by:
-
KAMEZAWA Hiroyuki authored
An event counter MEM_CGROUP_ON_MOVE is used for quick check whether file stat update can be done in async manner or not. Now, it use percpu counter and for_each_possible_cpu to update. This patch replaces for_each_possible_cpu to for_each_online_cpu and adds necessary synchronization logic at CPU HOTPLUG. [akpm@linux-foundation.org: coding-style fixes] Signed-off-by:
-
KAMEZAWA Hiroyuki authored
Now, memcgroup's per cpu coutner uses for_each_possible_cpu() to get the value. It's better to use for_each_online_cpu() and a cpu hotplug handler. This patch only handles statistics counter. MEM_CGROUP_ON_MOVE will be handled in another patch. Signed-off-by:
-
KAMEZAWA Hiroyuki authored
In memory cgroup management, we sometimes have to walk through subhierarchy of cgroup to gather informaiton, or lock something, etc. Now, to do that, mem_cgroup_walk_tree() function is provided. It calls given callback function per cgroup found. But the bad thing is that it has to pass a fixed style function and argument, "void*" and it adds much type casting to memcontrol.c. To make the code clean, this patch replaces walk_tree() with for_each_mem_cgroup_tree(iter, root) An iterator style call. The good point is that iterator call doesn't have to assume what kind of function is called under it. A bad point is that it may cause reference-count leak if a caller use "break" from the loop by mistake. I think the benefit is larger. The modified code seems straigtforward and easy to read because we don't have misterious callbacks and pointer cast. Signed-off-by:
-
KAMEZAWA Hiroyuki authored
At accounting file events per memory cgroup, we need to find memory cgroup via page_cgroup->mem_cgroup. Now, we use lock_page_cgroup() for guarantee pc->mem_cgroup is not overwritten while we make use of it. But, considering the context which page-cgroup for files are accessed, we can use alternative light-weight mutual execusion in the most case. At handling file-caches, the only race we have to take care of is "moving" account, IOW, overwriting page_cgroup->mem_cgroup. (See comment in the patch) Unlike charge/uncharge, "move" happens not so frequently. It happens only when rmdir() and task-moving (with a special settings.) This patch adds a race-checker for file-cache-status accounting v.s. account moving. The new per-cpu-per-memcg counter MEM_CGROUP_ON_MOVE is added. The routine for account move 1. Increment it before start moving 2. Call synchronize_rcu() 3. Decrement it after the end of moving. By this, file-status-counting routine can check it needs to call lock_page_cgroup(). In most case, I doesn't need to call it. Following is a perf data of a process which mmap()/munmap 32MB of file cache in a minute. Before patch: 28.25% mmap mmap [.] main 22.64% mmap [kernel.kallsyms] [k] page_fault 9.96% mmap [kernel.kallsyms] [k] mem_cgroup_update_file_mapped 3.67% mmap [kernel.kallsyms] [k] filemap_fault 3.50% mmap [kernel.kallsyms] [k] unmap_vmas 2.99% mmap [kernel.kallsyms] [k] __do_fault 2.76% mmap [kernel.kallsyms] [k] find_get_page After patch: 30.00% mmap mmap [.] main 23.78% mmap [kernel.kallsyms] [k] page_fault 5.52% mmap [kernel.kallsyms] [k] mem_cgroup_update_file_mapped 3.81% mmap [kernel.kallsyms] [k] unmap_vmas 3.26% mmap [kernel.kallsyms] [k] find_get_page 3.18% mmap [kernel.kallsyms] [k] __do_fault 3.03% mmap [kernel.kallsyms] [k] filemap_fault 2.40% mmap [kernel.kallsyms] [k] handle_mm_fault 2.40% mmap [kernel.kallsyms] [k] do_page_fault This patch reduces memcg's cost to some extent. (mem_cgroup_update_file_mapped is called by both of map/unmap) Note: It seems some more improvements are required..but no idea. maybe removing set/unset flag is required. Signed-off-by:Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Cc: Balbir Singh <balbir@in.ibm.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by:
-
