Skip to content

Commit

Permalink
[PATCH] reorder struct files_struct
Browse files Browse the repository at this point in the history
The file_lock spinlock sits close to mostly read fields of 'struct
files_struct'

In SMP (and NUMA) environments, each time a thread wants to open or close
a file, it has to acquire the spinlock, thus invalidating the cache line
containing this spinlock on other CPUS.  So other threads doing
read()/write()/...  calls that use RCU to access the file table are going
to ask further memory (possibly NUMA) transactions to read again this
memory line.

Move the spinlock to another cache line, so that concurrent threads can
share the cache line containing 'count' and 'fdt' fields.

It's worth up to 9% on a microbenchmark using a 4-thread 2-package x86
machine.  See
http://marc.theaimsgroup.com/?l=linux-kernel&m=112680448713342&w=2

Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
  • Loading branch information
Eric Dumazet authored and Linus Torvalds committed Nov 14, 2005
1 parent d6c7ac0 commit 95e861d
Showing 1 changed file with 5 additions and 5 deletions.
10 changes: 5 additions & 5 deletions include/linux/file.h
Original file line number Diff line number Diff line change
Expand Up @@ -33,13 +33,13 @@ struct fdtable {
* Open file table structure
*/
struct files_struct {
atomic_t count;
spinlock_t file_lock; /* Protects all the below members. Nests inside tsk->alloc_lock */
atomic_t count;
struct fdtable *fdt;
struct fdtable fdtab;
fd_set close_on_exec_init;
fd_set open_fds_init;
struct file * fd_array[NR_OPEN_DEFAULT];
fd_set close_on_exec_init;
fd_set open_fds_init;
struct file * fd_array[NR_OPEN_DEFAULT];
spinlock_t file_lock; /* Protects concurrent writers. Nests inside tsk->alloc_lock */
};

#define files_fdtable(files) (rcu_dereference((files)->fdt))
Expand Down

0 comments on commit 95e861d

Please sign in to comment.