Skip to contentSkip to conversation

Child exit when booting uml 2.4.20 on 2.6.0-test4 host

uml-devel Etay Meiri 4 participants 7 messages Sep 6, 2003 – Nov 10, 2003 1 message not in the archive User-mode Linux (UML)

lore
Share & tools
lore

Etay Meiri

author
#0 lore raw

Hi,

I have been running a UML kernel version 2.4.20 on a RH9 host with kernel 2.4.20-8. Everything is working properly in that environment. Now that I've upgraded the host system to 2.6.0-test4 I can't boot that same UML kernel as before. I get the following error:

Checking root filesystem
/dev/ubd/0: clean, 43713/75520 files, 136981/151040 blocks [/sbin/fsck.ext3 (1) -- /] fsck.ext3 -a /dev/ubd/0 [ OK ]
Remounting root filesystem in read-write mode: [ OK ]

Activating swap partitions:  [  OK  ]
Finding module dependencies:  [  OK  ]

Checking filesystems
Checking all file systems.
[ OK ]
Mounting local filesystems: Child 15433 exited with signal 11

The uml fs seems to be ok because I can mount it as a loop device in the host system.

This is the command line I use for running UML:

./linux-2.4.20 ubd0=./root_fs_rh_9 mem=48M eth0=ethertap,tap1,,192.168.1.13 con=pty con0=fd:0,fd:1

TIA,

Signature · 15 lines
--------------------------------
Etay Meiri
[email protected]
--------------------------------



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
User-mode-linux-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel
author
#2 lore raw
In reply to a message not in the archive
Quoted 11 lines · Jeff Dike
Quoted 1 line · [email protected] said:

Mounting local filesystems: Child 15433 exited with signal 11

Try it with the latest patch. It should be fixed - I applied a different fix from the one suggested by Michael Young.

Jeff

I tried it with patch 2.6.0-test3-1 (from Aug 15). Isn't that the latest one?

Signature · 15 lines
--------------------------------
Etay Meiri
[email protected]
--------------------------------



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
User-mode-linux-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel
#3 lore raw
Quoted 1 line · [email protected] said:

I tried it with patch 2.6.0-test3-1 (from Aug 15).

No, I've released a whole lot of patches since then.

Jeff

-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
User-mode-linux-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel

#4 lore raw
In reply to a message not in the archive
Quoted 2 lines · Jeff Dike

Try it with the latest patch. It should be fixed - I applied a different fix from the one suggested by Michael Young.

My limited testing suggest the problem is indeed fixed in 2.4.22-3um (and presumably in -2um as well though I haven't tried it).

Michael Young

-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
User-mode-linux-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel

#5 lore raw

Alle 22:03, venerdì 12 settembre 2003, M A Young ha scritto:

Quoted 6 lines · Jeff Dike
Quoted 2 lines

Try it with the latest patch. It should be fixed - I applied a different fix from the one suggested by Michael Young.

My limited testing suggest the problem is indeed fixed in 2.4.22-3um (and presumably in -2um as well though I haven't tried it).

I could see the bug with 2.6.0-test4/test5, and for me, too it was solved using 2.4.22-3um. So that's ok.
About 2.6, there is a little patch to implement in the uml_utilities, since insmod tun doesn't work any more with the new module-init-tools. That must change to modprobe tun. That is, replace "insmod" by "modprobe"(including quotes!) all over uml_net/ethertap.c and uml_net/tuntap.c. If you have little time, I can post this together with a patch to the spec file which is again out of date(some files are not listed).

Signature · 15 lines
cat <<EOSIGN
Paolo Giarrusso, aka Blaisorblade
Linux Kernel 2.4.21/2.6.0-test on an i686; Linux registered user n. 292729
EOSIGN



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
User-mode-linux-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel
#6 lore raw
Quoted 2 lines · BlaisorBlade

I could see the bug with 2.6.0-test4/test5, and for me, too it was solved using 2.4.22-3um. So that's ok.

If you want to try it on 2.6, I got impatient waiting for Jeff's patch to appear in his 2.6 tree and forward ported his patch (attached), with some minor cleanups.

Michael Young

arch/um/kernel/process.c +2 -2
diff -Naur linux-/arch/um/kernel/process.c linux+/arch/um/kernel/process.c
@@ -229,11 +229,11 @@
 
 int run_kernel_thread(int (*fn)(void *), void *arg, void **jmp_ptr)
 {
-	jmp_buf buf;
+	sigjmp_buf buf;
 	int n;
 
 	*jmp_ptr = &buf;
-	n = setjmp(buf);
+	n = sigsetjmp(buf,1);
 	if(n != 0)
 		return(n);
 	(*fn)(arg);
arch/um/kernel/skas/process.c +19 -19
diff -Naur linux-/arch/um/kernel/skas/process.c linux+/arch/um/kernel/skas/process.c
@@ -184,12 +184,12 @@
 void new_thread(void *stack, void **switch_buf_ptr, void **fork_buf_ptr,
 		void (*handler)(int))
 {
-	jmp_buf switch_buf, fork_buf;
+	sigjmp_buf switch_buf, fork_buf;
 
 	*switch_buf_ptr = &switch_buf;
 	*fork_buf_ptr = &fork_buf;
 
-	if(setjmp(fork_buf) == 0)
+	if(sigsetjmp(fork_buf,1) == 0)
 		new_thread_proc(stack, handler);
 
 	remove_sigstack();
@@ -197,12 +197,12 @@
 
 void thread_wait(void *sw, void *fb)
 {
-	jmp_buf buf, **switch_buf = sw, *fork_buf;
+	sigjmp_buf buf, **switch_buf = sw, *fork_buf;
 
 	*switch_buf = &buf;
 	fork_buf = fb;
-	if(setjmp(buf) == 0)
-		longjmp(*fork_buf, 1);
+	if(sigsetjmp(buf,1) == 0)
+		siglongjmp(*fork_buf, 1);
 }
 
 static int move_registers(int int_op, int fp_op, union uml_pt_regs *regs,
@@ -257,34 +257,34 @@
 
 void switch_threads(void *me, void *next)
 {
-	jmp_buf my_buf, **me_ptr = me, *next_buf = next;
+	sigjmp_buf my_buf, **me_ptr = me, *next_buf = next;
 	
 	*me_ptr = &my_buf;
-	if(setjmp(my_buf) == 0)
-		longjmp(*next_buf, 1);
+	if(sigsetjmp(my_buf,1) == 0)
+		siglongjmp(*next_buf, 1);
 }
 
-static jmp_buf initial_jmpbuf;
+static sigjmp_buf initial_jmpbuf;
 
 /* XXX Make these percpu */
 static void (*cb_proc)(void *arg);
 static void *cb_arg;
-static jmp_buf *cb_back;
+static sigjmp_buf *cb_back;
 
 int start_idle_thread(void *stack, void *switch_buf_ptr, void **fork_buf_ptr)
 {
-	jmp_buf **switch_buf = switch_buf_ptr;
+	sigjmp_buf **switch_buf = switch_buf_ptr;
 	int n;
 
 	*fork_buf_ptr = &initial_jmpbuf;
-	n = setjmp(initial_jmpbuf);
+	n = sigsetjmp(initial_jmpbuf,1);
 	if(n == 0)
 		new_thread_proc((void *) stack, new_thread_handler);
 	else if(n == 1)
 		remove_sigstack();
 	else if(n == 2){
 		(*cb_proc)(cb_arg);
-		longjmp(*cb_back, 1);
+		siglongjmp(*cb_back, 1);
 	}
 	else if(n == 3){
 		kmalloc_ok = 0;
@@ -294,7 +294,7 @@
 		kmalloc_ok = 0;
 		return(1);
 	}
-	longjmp(**switch_buf, 1);
+	siglongjmp(**switch_buf, 1);
 }
 
 void remove_sigstack(void)
@@ -309,15 +309,15 @@
 
 void initial_thread_cb_skas(void (*proc)(void *), void *arg)
 {
-	jmp_buf here;
+	sigjmp_buf here;
 
 	cb_proc = proc;
 	cb_arg = arg;
 	cb_back = &here;
 
 	block_signals();
-	if(setjmp(here) == 0)
-		longjmp(initial_jmpbuf, 2);
+	if(sigsetjmp(here,1) == 0)
+		siglongjmp(initial_jmpbuf, 2);
 	unblock_signals();
 
 	cb_proc = NULL;
@@ -328,13 +328,13 @@
 void halt_skas(void)
 {
 	block_signals();
-	longjmp(initial_jmpbuf, 3);
+	siglongjmp(initial_jmpbuf, 3);
 }
 
 void reboot_skas(void)
 {
 	block_signals();
-	longjmp(initial_jmpbuf, 4);
+	siglongjmp(initial_jmpbuf, 4);
 }
 
 int new_mm(int from)
arch/um/kernel/trap_user.c +2 -2
diff -Naur linux-/arch/um/kernel/trap_user.c linux+/arch/um/kernel/trap_user.c
@@ -123,9 +123,9 @@
 
 void do_longjmp(void *b, int val)
 {
-	jmp_buf *buf = b;
+	sigjmp_buf *buf = b;
 
-	longjmp(*buf, val);
+	siglongjmp(*buf, val);
 }
 
 /*
arch/um/kernel/tt/uaccess_user.c +2 -2
diff -Naur linux-/arch/um/kernel/tt/uaccess_user.c linux+/arch/um/kernel/tt/uaccess_user.c
@@ -72,10 +72,10 @@
 	struct tt_regs save = TASK_REGS(get_current())->tt;
 	int ret;
 	unsigned long *faddrp = (unsigned long *)fault_addr;
-	jmp_buf jbuf;
+	sigjmp_buf jbuf;
 
 	*fault_catcher = &jbuf;
-	if(setjmp(jbuf) == 0){
+	if(sigsetjmp(jbuf,1) == 0){
 		ret = strlen(str) + 1;
 	} 
 	else {
arch/um/kernel/uaccess_user.c +2 -2
diff -Naur linux-/arch/um/kernel/uaccess_user.c linux+/arch/um/kernel/uaccess_user.c
@@ -18,9 +18,9 @@
 {
 	unsigned long *faddrp = (unsigned long *) fault_addr, ret;
 
-	jmp_buf jbuf;
+	sigjmp_buf jbuf;
 	*fault_catcher = &jbuf;
-	if(setjmp(jbuf) == 0){
+	if(sigsetjmp(jbuf,1) == 0){
 		(*op)(to, from, n);
 		ret = 0;
 		*faulted_out = 0;
#7 lore raw
In reply to BlaisorBlade
Quoted 5 lines · [email protected] said:

About 2.6, there is a little patch to implement in the uml_utilities, since insmod tun doesn't work any more with the new module-init-tools. That must change to modprobe tun. That is, replace "insmod" by "modprobe"(including quotes!) all over uml_net/ethertap.c and uml_net/tuntap.c.

Yup, fixed. There were only three of them, though.

Quoted 3 lines

If you have little time, I can post this
together with a patch to the spec file which is again out of date(some files are not listed).

Send me a patch to the spec file, please.

Jeff

-------------------------------------------------------
This SF.Net email sponsored by: ApacheCon 2003, 16-19 November in Las Vegas. Learn firsthand the latest developments in Apache, PHP, Perl, XML, Java, MySQL, WebDAV, and more! http://www.apachecon.com/
_______________________________________________
User-mode-linux-devel mailing list [email protected]
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel