rcutorture: don't mark torture_init_{begin,end} with __init
PATCH 1/1 rcutorture: don't mark torture_init_{begin,end} with __init
By the time rcutorture is initialised via its module_init callback, the .init.text section of torture.ko has already been freed, so the calls to torture_init{begin,end} end up branching to junk.
This patch removes the __init annotations from torture_init_{begin,end} to make sure the code is there when we want to call it.
---
Just lost a day to this -- I was sure my static relocation handling must be to blame!
kernel/torture.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
kernel/torture.c +2 -2
@@ -596,7 +596,7 @@ static void torture_stutter_cleanup(void)* The runnable parameter points to a flag that controls whether or not * the test is currently runnable. If there is no such flag, pass in NULL. */ -void __init torture_init_begin (char *ttype,bool v,int *runnable) +void torture_init_begin (char *ttype,bool v,int *runnable) {mutex_lock (&fullstop_mutex); torture_type = ttype;
@@ -610,7 +610,7 @@ EXPORT_SYMBOL_GPL(torture_init_begin);/* * Tell the torture module that initialization is complete. */ -void __init torture_init_end (void ) +void torture_init_end (void ) {mutex_unlock (&fullstop_mutex);register_reboot_notifier (&torture_shutdown_nb);
Signature · 1 line
1.9.2
Quoted 10 lines · Will Deacon
By the time rcutorture is initialised via its
module_initcallback, the .init.text section of torture.ko has already been freed, so the calls to torture_init{begin,end} end up branching to junk.This patch removes the
__initannotations from torture_init_{begin,end} to make sure the code is there when we want to call it.Cc Paul E. McKenney <[email protected]>Cc Josh Triplett <[email protected]>Signed-off-by Will Deacon <[email protected]>
Good catch, and apologies for the wasted day, but someone beat you to this. By only 12 days, which is a bit surprising given how long this has been lurking.
Thanx, Paul
Quoted 33 lines
---
Just lost a day to this -- I was sure my static relocation handling must be to blame!
kernel/torture.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)kernel/torture.c +2
-2@@ -596,7 +596,7 @@ static void torture_stutter_cleanup(void)* The runnable parameter points to a flag that controls whether or not * the test is currently runnable. If there is no such flag, pass in NULL. */ -void __init torture_init_begin (char *ttype,bool v,int *runnable) +void torture_init_begin (char *ttype,bool v,int *runnable) {mutex_lock (&fullstop_mutex); torture_type = ttype;@@ -610,7 +610,7 @@ EXPORT_SYMBOL_GPL(torture_init_begin);/* * Tell the torture module that initialization is complete. */ -void __init torture_init_end (void ) +void torture_init_end (void ) {mutex_unlock (&fullstop_mutex);register_reboot_notifier (&torture_shutdown_nb);Signature · 1 line
1.9.2
Quoted 8 lines · Paul E. McKenney
Quoted 7 lines · Will Deacon
By the time rcutorture is initialised via its
module_initcallback, the .init.text section of torture.ko has already been freed, so the calls to torture_init{begin,end} end up branching to junk.This patch removes the
__initannotations from torture_init_{begin,end} to make sure the code is there when we want to call it.
Quoted 7 lines
Quoted 3 lines
Cc Paul E. McKenney <[email protected]>Cc Josh Triplett <[email protected]>Signed-off-by Will Deacon <[email protected]>Good catch, and apologies for the wasted day, but someone beat you to this. By only 12 days, which is a bit surprising given how long this has been lurking.
Oh well, at least it's getting fixed!
Will