Re: Fixed - pthread altsigstack problem

[KUROSAWA Takahiro]

On Mon, 12 Mar 2001 18:46:16 -0500
James FitzGibbon [EMAIL PROTECTED] wrote:

 Both of the patches below fix the problem mentioned in PR bin/25110.  The
 first one fixes it inside of kern_fork.c and would appear to apply the
 corrective behaviour regardless of whether the process uses libc_r or not. 

The first patch that I sent to you makes the SS_DISABLE flag setting 
be inherited to the child process (especially when the parent have
cleared the flag),  as I had mentioned in the mail to -current:

http://docs.freebsd.org/cgi/getmsg.cgi?fetch=332928+0+archive/2001/freebsd-current/20010304.freebsd-current

I'm not sure that the SS_DISABLE flag setting should be inherited to
the child process,  but at least I can say the behavior of 5-CURRENT
(and 4-STABLE) is different from that of Solaris 7, NetBSD 1.5 and
FreeBSD 3.5-STABLE.

To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message

Fixed - pthread altsigstack problem

[James FitzGibbon]

Both of the patches below fix the problem mentioned in PR bin/25110.  The
first one fixes it inside of kern_fork.c and would appear to apply the
corrective behaviour regardless of whether the process uses libc_r or not. 
The second patch fixes the problem inside of uthread_fork.c.  Whether the
first approach imposes an extra cost on non-threaded applications I'm not
kernel-experienced enough to say, but hopefully between two of you gents you
can decide which is the better fix to apply.

As I mentioned in my original post, this is a bug that we are experiencing
in 4.3-BETA, so if this could make it into 4.3-RELEASE it would be of great
help.  One note regarding the second (libc_r) patch: the reference to
__sys_sigaltstack needs to be changed to _thread_sys_sigaltstack in order to
prevent undefined symbols on 4.x systems.

Patch #1 (kernel fix):

--- kern_fork.c.origSat Mar 10 12:17:40 2001
+++ kern_fork.c Sat Mar 10 12:20:39 2001
@@ -434,7 +434,7 @@ 
 * Preserve some more flags in subprocess.  P_PROFIL has already
 * been preserved.  
 */ 
-   p2-p_flag |= p1-p_flag  P_SUGID; 
+   p2-p_flag |= p1-p_flag  (P_SUGID | P_ALTSTACK);  
if (p1-p_session-s_ttyvp != NULL  p1-p_flag  P_CONTROLT)  
p2-p_flag |= P_CONTROLT;   
if (flags  RFPPWAIT)

Patch #2 (libc_r fix):

--- uthread_fork.c  2001/01/24 13:03:33 1.21   
   
+++ uthread_fork.c  2001/03/09 17:53:37
+   
@@ -32,6 +32,7 @@  
   
  * $FreeBSD: src/lib/libc_r/uthread/uthread_fork.c,v 1.21 2001/01/24 13:03:33 
deischen Exp $ 
  
  */   
   
 #include errno.h
   
+#include signal.h   
+   
 #include string.h   
   
 #include stdlib.h   
   
 #include unistd.h   
   
@@ -110,7 +111,16 @@   
   
else if (_pq_init(_readyq) != 0) {
   
/* Abort this application: */  
   
PANIC("Cannot initialize priority ready queue.");  
   
-   } else {   
   
+   } else if ((_thread_sigstack.ss_sp == NULL)  
+   
+   ((_thread_sigstack.ss_sp = malloc(SIGSTKSZ)) == NULL)) 
+   
+   PANIC("Unable to allocate alternate signal stack");
+   
+   else { 
+   
+   /* Install the alternate signal stack: */  
+   
+   _thread_sigstack.ss_size = SIGSTKSZ;   
+   
+   _thread_sigstack.ss_flags = 0; 
+   
+   if (__sys_sigaltstack(_thread_sigstack, NULL) != 0)   
+   
+   PANIC("Unable to install alternate signal stack"); 
+   
+  
+   
/* 
   
 * Enter a loop to remove all threads other than   
   
 * the running thread from the thread list:
  

Thanks for the help guys.

-- 
j.

To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe