module: Optimize search_module_extables()
PATCH 1/1 module: Optimize search_module_extables()
While looking through the __ex_table stuff I found that we do a linear lookup of the module. Also fix up a comment.
--- kernel/module.c | 27 ++++++++++++++------------- 1 file changed, 14 insertions(+), 13 deletions(-)
kernel/module.c +14 -13
@@ -4165,22 +4165,23 @@ const struct exception_table_entry *search_module_extables(unsigned long addr)struct module *mod;preempt_disable (); -list_for_each_entry_rcu (mod, &modules, list) { -if (mod->state ==MODULE_STATE_UNFORMED ) -continue ; -if (mod->num_exentries ==0 ) -continue ; + mod =__module_address (addr); +if (!mod) + goto out; - e =search_extable (mod->extable, - mod->extable + mod->num_exentries -1 , - addr); -if (e) -break ; - } +if (!mod->num_exentries) + goto out; + + e =search_extable (mod->extable, + mod->extable + mod->num_exentries -1 , + addr); +out:preempt_enable (); -/* Now, if we found one, we are running inside it now, hence -we cannot unload the module, hence no refcnt needed. */ +/* +* Now, if we found one, we are running inside it now, hence +* we cannot unload the module, hence no refcnt needed. +*/ return e; }
Quoted 5 lines · Peter Zijlstra
While looking through the
__ex_tablestuff I found that we do a linear lookup of the module. Also fix up a comment.Signed-off-by Peter Zijlstra (Intel) <[email protected]>
I'm not all that familiar with the module code, but from a quick scan of __module_address() this looks sane to me. FWIW:
Mark.
Quoted 45 lines
--- kernel/module.c | 27 ++++++++++++++------------- 1 file changed, 14 insertions(+), 13 deletions(-)kernel/module.c +14
-13@@ -4165,22 +4165,23 @@ const struct exception_table_entry *search_module_extables(unsigned long addr)struct module *mod;preempt_disable (); -list_for_each_entry_rcu (mod, &modules, list) { -if (mod->state ==MODULE_STATE_UNFORMED ) -continue ; -if (mod->num_exentries ==0 ) -continue ; + mod =__module_address (addr); +if (!mod) + goto out; - e =search_extable (mod->extable, - mod->extable + mod->num_exentries -1 , - addr); -if (e) -break ; - } +if (!mod->num_exentries) + goto out; + + e =search_extable (mod->extable, + mod->extable + mod->num_exentries -1 , + addr); +out:preempt_enable (); -/* Now, if we found one, we are running inside it now, hence -we cannot unload the module, hence no refcnt needed. */ +/* +* Now, if we found one, we are running inside it now, hence +* we cannot unload the module, hence no refcnt needed. +*/ return e; }
Jessica Yu
Applied+++ Peter Zijlstra [08/02/17 15:48 +0100]:
Quoted 5 lines
While looking through the
__ex_tablestuff I found that we do a linear lookup of the module. Also fix up a comment.Signed-off-by Peter Zijlstra (Intel) <[email protected]>
Applied, thanks.
Hm. A quick scan through module.c still shows a couple of places that use similar linear lookups, and may benefit from the same __module_address optimization. But I'll save that for a separate patch..
Jessica
Quoted 45 lines
---
kernel/module.c| 27 ++++++++++++++-------------1 file changed, 14 insertions(+), 13 deletions(-)kernel/module.c +14
-13@@ -4165,22 +4165,23 @@ const struct exception_table_entry *search_module_extables(unsigned long addr)struct module *mod;preempt_disable (); -list_for_each_entry_rcu (mod, &modules, list) { -if (mod->state ==MODULE_STATE_UNFORMED ) -continue ; -if (mod->num_exentries ==0 ) -continue ; + mod =__module_address (addr); +if (!mod) + goto out; - e =search_extable (mod->extable, - mod->extable + mod->num_exentries -1 , - addr); -if (e) -break ; - } +if (!mod->num_exentries) + goto out; + + e =search_extable (mod->extable, + mod->extable + mod->num_exentries -1 , + addr); +out:preempt_enable (); -/* Now, if we found one, we are running inside it now, hence -we cannot unload the module, hence no refcnt needed. */ +/* +* Now, if we found one, we are running inside it now, hence +* we cannot unload the module, hence no refcnt needed. +*/ return e; }