* [PATCH] proc: rearrange struct proc_dir_entry
@ 2017-12-20 21:59 Alexey Dobriyan
2017-12-20 23:10 ` Randy Dunlap
0 siblings, 1 reply; 3+ messages in thread
From: Alexey Dobriyan @ 2017-12-20 21:59 UTC (permalink / raw)
To: akpm; +Cc: linux-kernel
struct proc_dir_entry became bit messy over years:
* move 16-bit ->mode_t before namelen to get rid of padding
* make ->in_use first field: it seems to be most used resulting in
smaller code on x86_64 (defconfig):
add/remove: 0/0 grow/shrink: 7/13 up/down: 24/-67 (-43)
Function old new delta
proc_readdir_de 451 455 +4
proc_get_inode 282 286 +4
pde_put 65 69 +4
remove_proc_subtree 294 297 +3
remove_proc_entry 297 300 +3
proc_register 295 298 +3
proc_notify_change 94 97 +3
unuse_pde 27 26 -1
proc_reg_write 89 85 -4
proc_reg_unlocked_ioctl 85 81 -4
proc_reg_read 89 85 -4
proc_reg_llseek 87 83 -4
proc_reg_get_unmapped_area 123 119 -4
proc_entry_rundown 139 135 -4
proc_reg_poll 91 85 -6
proc_reg_mmap 79 73 -6
proc_get_link 55 49 -6
proc_reg_release 108 101 -7
proc_reg_open 298 291 -7
close_pdeo 228 218 -10
* move writeable fields together to a first cacheline (on x86_64),
those include
* ->in_use: reference count, taken every open/read/write/close etc
* ->count: reference count, taken at readdir on every entry
* ->pde_openers: tracks (nearly) every open, dirtied
* ->pde_unload_lock: spinlock protecting ->pde_openers
* ->proc_iops, ->proc_fops, ->data: writeonce fields,
used right together with previous group.
* other rarely written fields go into 1st/2nd and 2nd/3rd cacheline on
32-bit and 64-bit respectively.
Additionally on 32-bit, ->subdir, ->subdir_node, ->namelen, ->name
go fully into 2nd cacheline, separated from writeable fields.
They are all used during lookup.
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
---
fs/proc/internal.h | 23 +++++++++++++----------
1 file changed, 13 insertions(+), 10 deletions(-)
--- a/fs/proc/internal.h
+++ b/fs/proc/internal.h
@@ -31,24 +31,27 @@ struct mempolicy;
* subdir_node is used to build the rb tree "subdir" of the parent.
*/
struct proc_dir_entry {
+ /*
+ * number of callers into module in progress;
+ * negative -> it's going away RSN
+ */
+ atomic_t in_use;
+ atomic_t count; /* use count */
+ struct list_head pde_openers; /* who did ->open, but not ->release */
+ spinlock_t pde_unload_lock; /* proc_fops checks and pde_users bumps */
+ struct completion *pde_unload_completion;
+ const struct inode_operations *proc_iops;
+ const struct file_operations *proc_fops;
+ void *data;
unsigned int low_ino;
- umode_t mode;
nlink_t nlink;
kuid_t uid;
kgid_t gid;
loff_t size;
- const struct inode_operations *proc_iops;
- const struct file_operations *proc_fops;
struct proc_dir_entry *parent;
struct rb_root_cached subdir;
struct rb_node subdir_node;
- void *data;
- atomic_t count; /* use count */
- atomic_t in_use; /* number of callers into module in progress; */
- /* negative -> it's going away RSN */
- struct completion *pde_unload_completion;
- struct list_head pde_openers; /* who did ->open, but not ->release */
- spinlock_t pde_unload_lock; /* proc_fops checks and pde_users bumps */
+ umode_t mode;
u8 namelen;
char name[];
} __randomize_layout;
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [PATCH] proc: rearrange struct proc_dir_entry
2017-12-20 21:59 [PATCH] proc: rearrange struct proc_dir_entry Alexey Dobriyan
@ 2017-12-20 23:10 ` Randy Dunlap
2017-12-21 6:25 ` Alexey Dobriyan
0 siblings, 1 reply; 3+ messages in thread
From: Randy Dunlap @ 2017-12-20 23:10 UTC (permalink / raw)
To: Alexey Dobriyan, akpm; +Cc: linux-kernel
On 12/20/2017 01:59 PM, Alexey Dobriyan wrote:
> struct proc_dir_entry became bit messy over years:
>
> * move 16-bit ->mode_t before namelen to get rid of padding
> * make ->in_use first field: it seems to be most used resulting in
> smaller code on x86_64 (defconfig):
>
> add/remove: 0/0 grow/shrink: 7/13 up/down: 24/-67 (-43)
> Function old new delta
> proc_readdir_de 451 455 +4
> proc_get_inode 282 286 +4
> pde_put 65 69 +4
> remove_proc_subtree 294 297 +3
> remove_proc_entry 297 300 +3
> proc_register 295 298 +3
> proc_notify_change 94 97 +3
> unuse_pde 27 26 -1
> proc_reg_write 89 85 -4
> proc_reg_unlocked_ioctl 85 81 -4
> proc_reg_read 89 85 -4
> proc_reg_llseek 87 83 -4
> proc_reg_get_unmapped_area 123 119 -4
> proc_entry_rundown 139 135 -4
> proc_reg_poll 91 85 -6
> proc_reg_mmap 79 73 -6
> proc_get_link 55 49 -6
> proc_reg_release 108 101 -7
> proc_reg_open 298 291 -7
> close_pdeo 228 218 -10
>
> * move writeable fields together to a first cacheline (on x86_64),
> those include
> * ->in_use: reference count, taken every open/read/write/close etc
> * ->count: reference count, taken at readdir on every entry
> * ->pde_openers: tracks (nearly) every open, dirtied
> * ->pde_unload_lock: spinlock protecting ->pde_openers
> * ->proc_iops, ->proc_fops, ->data: writeonce fields,
> used right together with previous group.
>
> * other rarely written fields go into 1st/2nd and 2nd/3rd cacheline on
> 32-bit and 64-bit respectively.
>
> Additionally on 32-bit, ->subdir, ->subdir_node, ->namelen, ->name
> go fully into 2nd cacheline, separated from writeable fields.
> They are all used during lookup.
Does
> } __randomize_layout;
pay attention to any of that?
> Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
> ---
>
> fs/proc/internal.h | 23 +++++++++++++----------
> 1 file changed, 13 insertions(+), 10 deletions(-)
>
> --- a/fs/proc/internal.h
> +++ b/fs/proc/internal.h
> @@ -31,24 +31,27 @@ struct mempolicy;
> * subdir_node is used to build the rb tree "subdir" of the parent.
> */
> struct proc_dir_entry {
> + /*
> + * number of callers into module in progress;
> + * negative -> it's going away RSN
> + */
> + atomic_t in_use;
> + atomic_t count; /* use count */
> + struct list_head pde_openers; /* who did ->open, but not ->release */
> + spinlock_t pde_unload_lock; /* proc_fops checks and pde_users bumps */
> + struct completion *pde_unload_completion;
> + const struct inode_operations *proc_iops;
> + const struct file_operations *proc_fops;
> + void *data;
> unsigned int low_ino;
> - umode_t mode;
> nlink_t nlink;
> kuid_t uid;
> kgid_t gid;
> loff_t size;
> - const struct inode_operations *proc_iops;
> - const struct file_operations *proc_fops;
> struct proc_dir_entry *parent;
> struct rb_root_cached subdir;
> struct rb_node subdir_node;
> - void *data;
> - atomic_t count; /* use count */
> - atomic_t in_use; /* number of callers into module in progress; */
> - /* negative -> it's going away RSN */
> - struct completion *pde_unload_completion;
> - struct list_head pde_openers; /* who did ->open, but not ->release */
> - spinlock_t pde_unload_lock; /* proc_fops checks and pde_users bumps */
> + umode_t mode;
> u8 namelen;
> char name[];
> } __randomize_layout;
>
--
~Randy
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [PATCH] proc: rearrange struct proc_dir_entry
2017-12-20 23:10 ` Randy Dunlap
@ 2017-12-21 6:25 ` Alexey Dobriyan
0 siblings, 0 replies; 3+ messages in thread
From: Alexey Dobriyan @ 2017-12-21 6:25 UTC (permalink / raw)
To: Randy Dunlap; +Cc: akpm, linux-kernel
On Wed, Dec 20, 2017 at 03:10:48PM -0800, Randy Dunlap wrote:
> On 12/20/2017 01:59 PM, Alexey Dobriyan wrote:
> > struct proc_dir_entry became bit messy over years:
> >
> > * move 16-bit ->mode_t before namelen to get rid of padding
> > * make ->in_use first field: it seems to be most used resulting in
> > smaller code on x86_64 (defconfig):
> >
> > add/remove: 0/0 grow/shrink: 7/13 up/down: 24/-67 (-43)
> > Function old new delta
> > proc_readdir_de 451 455 +4
> > proc_get_inode 282 286 +4
> > pde_put 65 69 +4
> > remove_proc_subtree 294 297 +3
> > remove_proc_entry 297 300 +3
> > proc_register 295 298 +3
> > proc_notify_change 94 97 +3
> > unuse_pde 27 26 -1
> > proc_reg_write 89 85 -4
> > proc_reg_unlocked_ioctl 85 81 -4
> > proc_reg_read 89 85 -4
> > proc_reg_llseek 87 83 -4
> > proc_reg_get_unmapped_area 123 119 -4
> > proc_entry_rundown 139 135 -4
> > proc_reg_poll 91 85 -6
> > proc_reg_mmap 79 73 -6
> > proc_get_link 55 49 -6
> > proc_reg_release 108 101 -7
> > proc_reg_open 298 291 -7
> > close_pdeo 228 218 -10
> >
> > * move writeable fields together to a first cacheline (on x86_64),
> > those include
> > * ->in_use: reference count, taken every open/read/write/close etc
> > * ->count: reference count, taken at readdir on every entry
> > * ->pde_openers: tracks (nearly) every open, dirtied
> > * ->pde_unload_lock: spinlock protecting ->pde_openers
> > * ->proc_iops, ->proc_fops, ->data: writeonce fields,
> > used right together with previous group.
> >
> > * other rarely written fields go into 1st/2nd and 2nd/3rd cacheline on
> > 32-bit and 64-bit respectively.
> >
> > Additionally on 32-bit, ->subdir, ->subdir_node, ->namelen, ->name
> > go fully into 2nd cacheline, separated from writeable fields.
> > They are all used during lookup.
>
> Does
> > } __randomize_layout;
> pay attention to any of that?
No. You can randomize inside cachelines but it will look rather ugly.
__randomize_layout rearranges everything.
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2017-12-21 6:25 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2017-12-20 21:59 [PATCH] proc: rearrange struct proc_dir_entry Alexey Dobriyan
2017-12-20 23:10 ` Randy Dunlap
2017-12-21 6:25 ` Alexey Dobriyan
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).