2.5.34 - EXPORT_SYMBOL(reparent_to_init) for module build
With 2.5.34, in order to build a module that calls daemonize(), I had to export reparent_to_init:
kernel/ksyms.c +1 -0
@@ -521,6 +521,7 @@EXPORT_SYMBOL (securebits);EXPORT_SYMBOL (cap_bset);EXPORT_SYMBOL (daemonize); +EXPORT_SYMBOL (reparent_to_init);EXPORT_SYMBOL (csum_partial);/* for networking and md */ EXPORT_SYMBOL (seq_escape);EXPORT_SYMBOL (seq_printf);
-- Patrick Mansfield
Quoted 2 lines · Patrick Mansfield
With 2.5.34, in order to build a module that calls
daemonize(), I had to exportreparent_to_init:
I suggest you check the source of daemonize() in 2.5.34 ;)
regards
john
Signature · 3 lines
"This *is* Usenet, after all, where virtually every conversation that goes on is fairly ludicrous in the first place." - Godwin's Law FAQ
Quoted 9 lines · John Levon
Quoted 2 lines · Patrick Mansfield
With 2.5.34, in order to build a module that calls
daemonize(), I had to exportreparent_to_init:I suggest you check the source of
daemonize()in 2.5.34 ;)regards
john
OK, thanks, I thought I grepped the module source (qla v6b5 adapter, not included in the kernel tree), but missed the call to reparent_to_init(). I was wondering why I'd need an export for a function only called from within in its own file. I removed the reparent_to_init call and it worked fine.
Perhaps reparent_to_init should now be static?
Thanks.
-- Patrick Mansfield
Quoted 1 line · Patrick Mansfield
Perhaps
reparent_to_initshould now be static?
My grep certainly agrees with this. (Don't forget to remove the prototype in sched.h too).
regards
john
Signature · 3 lines
"This *is* Usenet, after all, where virtually every conversation that goes on is fairly ludicrous in the first place." - Godwin's Law FAQ
In message <[email protected]> you write:
Quoted 2 lines
With 2.5.34, in order to build a module that calls
daemonize(), I had to exportreparent_to_init:
Why?
Rusty.
--
Anyone who quotes me in their sig is an idiot. -- Rusty Russell.