[uml-devel] [patch] uml fixes, 2.6.1-mm5-A2

[Ingo Molnar <mingo@elte.hu>]

the patch below fixes a bug in UML that i've reproduced under 2.6.1-mm5. 

The bug hits in the following race scenario: if userspace() is just
about to handle a SIGTRAP signal that it wants to relay_signal(), and
SIGVTALRM hits the context before this happens but after is_user is set
to 1 in userspace(), then we get a kernelspace-signal panic. Ought to be
rare but definitely triggerable.

	Ingo

--- linux/arch/um/kernel/skas/trap_user.c.orig	
+++ linux/arch/um/kernel/skas/trap_user.c	
@@ -19,8 +19,10 @@ void sig_handler_common_skas(int sig, vo
 	struct skas_regs *r;
 	struct signal_info *info;
 	int save_errno = errno;
+	int save_user;
 
 	r = &TASK_REGS(get_current())->skas;
+	save_user = r->is_user;
 	r->is_user = 0;
 	r->fault_addr = SC_FAULT_ADDR(sc);
 	r->fault_type = SC_FAULT_TYPE(sc);
@@ -32,6 +34,7 @@ void sig_handler_common_skas(int sig, vo
 
 	(*info->handler)(sig, (union uml_pt_regs *) r);
 
+	r->is_user = save_user;
 	errno = save_errno;
 }
 


-------------------------------------------------------
The SF.Net email is sponsored by EclipseCon 2004
Premiere Conference on Open Tools Development and Integration
See the breadth of Eclipse activity. February 3-5 in Anaheim, CA.
http://www.eclipsecon.org/osdn
_______________________________________________
User-mode-linux-devel mailing list
User-mode-linux-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel