Skip to contentSkip to conversation

Module Alias back compat code

PATCH In review Rusty Russell 2 participants 3 messages Dec 19, 2003 – Dec 22, 2003

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 Module Alias back compat code

The provides backwards compat for old char and block aliases.

MODULE_ALIAS_BLOCK() and MODULE_ALIAS_CHAR() define aliases of form "XXX-<major>-<minor>", so we should probe for modules using this form. Unfortunately in 2.4, block aliases were "XXX-<major>" and char aliases were of both forms.

Ideally, all modules would now be using MODULE_ALIAS() macros to define their aliases, and the old configuration files wouldn't matter as much. Unfortunately, this hasn't happened, so we make request_module() return the exit status of modprobe, and then do fallback when probing for char and block devices.

Please apply.
Rusty.

Name: Backwards Compatibility For Old block and char Aliases
Author: Rusty Russell
Status: Tested on 2.6.0

D: MODULE_ALIAS_BLOCK() and MODULE_ALIAS_CHAR() define aliases of form
D: "XXX-<major>-<minor>", so we should probe for modules using this
D: form. Unfortunately in 2.4, block aliases were "XXX-<major>" and
D: char aliases were of both forms.
D:
D: Ideally, all modules would now be using MODULE_ALIAS() macros to
D: define their aliases, and the old configuration files wouldn't
D: matter as much. Unfortunately, this hasn't happened, so we make
D: request_module() return the exit status of modprobe, and then
D: do fallback when probing for char and block devices.

drivers/block/genhd.c +3 -1
diff -urpN --exclude TAGS -X /home/rusty/devel/kernel/kernel-patches/current-dontdiff --minimal .1494-linux-2.6.0/drivers/block/genhd.c .1494-linux-2.6.0.updated/drivers/block/genhd.c
@@ -296,7 +296,9 @@ extern int blk_dev_init(void);
 
 static struct kobject *base_probe(dev_t dev, int *part, void *data)
 {
-	request_module("block-major-%d", MAJOR(dev));
+	if (request_module("block-major-%d-%d", MAJOR(dev), MINOR(dev)) > 0)
+		/* Make old-style 2.4 aliases work */
+		request_module("block-major-%d", MAJOR(dev));
 	return NULL;
 }
 
fs/char_dev.c +3 -1
diff -urpN --exclude TAGS -X /home/rusty/devel/kernel/kernel-patches/current-dontdiff --minimal .1494-linux-2.6.0/fs/char_dev.c .1494-linux-2.6.0.updated/fs/char_dev.c
@@ -434,7 +434,9 @@ void cdev_init(struct cdev *cdev, struct
 
 static struct kobject *base_probe(dev_t dev, int *part, void *data)
 {
-	request_module("char-major-%d-%d", MAJOR(dev), MINOR(dev));
+	if (request_module("char-major-%d-%d", MAJOR(dev), MINOR(dev)) > 0)
+		/* Make old-style 2.4 aliases work */
+		request_module("char-major-%d", MAJOR(dev));
 	return NULL;
 }
 
include/linux/kmod.h +2 -0
diff -urpN --exclude TAGS -X /home/rusty/devel/kernel/kernel-patches/current-dontdiff --minimal .1494-linux-2.6.0/include/linux/kmod.h .1494-linux-2.6.0.updated/include/linux/kmod.h
@@ -24,6 +24,8 @@
 #include <linux/compiler.h>
 
 #ifdef CONFIG_KMOD
+/* modprobe exit status on success, -ve on error.  Return value
+ * usually useless though. */
 extern int request_module(const char * name, ...) __attribute__ ((format (printf, 1, 2)));
 #else
 static inline int request_module(const char * name, ...) { return -ENOSYS; }
kernel/kmod.c +10 -5
diff -urpN --exclude TAGS -X /home/rusty/devel/kernel/kernel-patches/current-dontdiff --minimal .1494-linux-2.6.0/kernel/kmod.c .1494-linux-2.6.0.updated/kernel/kmod.c
@@ -182,16 +182,21 @@ static int wait_for_helper(void *data)
 {
 	struct subprocess_info *sub_info = data;
 	pid_t pid;
+	struct k_sigaction sa;
+
+	/* Install a handler: if SIGCLD isn't handled sys_wait4 won't
+	 * populate the status, but will return -ECHILD. */
+	sa.sa.sa_handler = SIG_IGN;
+	sa.sa.sa_flags = 0;
+	siginitset(&sa.sa.sa_mask, sigmask(SIGCHLD));
+	do_sigaction(SIGCHLD, &sa, (struct k_sigaction *)0);
+	allow_signal(SIGCHLD);
 
-	sub_info->retval = 0;
 	pid = kernel_thread(____call_usermodehelper, sub_info, SIGCHLD);
 	if (pid < 0)
 		sub_info->retval = pid;
 	else
-		/* We don't have a SIGCHLD signal handler, so this
-		 * always returns -ECHILD, but the important thing is
-		 * that it blocks. */
-		sys_wait4(pid, NULL, 0, NULL);
+		sys_wait4(pid, &sub_info->retval, 0, NULL);
 
 	complete(sub_info->complete);
 	return 0;

--
Anyone who quotes me in their sig is an idiot. -- Rusty Russell.

#1 lore raw
Quoted 7 lines · Rusty Russell

The provides backwards compat for old char and block aliases.

MODULE_ALIAS_BLOCK() and MODULE_ALIAS_CHAR() define aliases of form "XXX-<major>-<minor>", so we should probe for modules using this form. Unfortunately in 2.4, block aliases were "XXX-<major>" and char aliases were of both forms.

Quoted 5 lines

Ideally, all modules would now be using MODULE_ALIAS() macros to define their aliases, and the old configuration files wouldn't matter as much. Unfortunately, this hasn't happened, so we make request_module() return the exit status of modprobe, and then do fallback when probing for char and block devices.

Umm.. Although I may be mis-understanding this problem, is the following scripts the not enough?

This does

	block-major-1 -> block-major-1-*
generate-modprobe.conf renamed from generate-modprobe.conf.orig +2 -1
@@ -59,8 +59,9 @@
 parse_alias()
 {
     PA_ALIAS=`resolve_alias $1 $3`
+    NAME=`echo $2|sed -e 's/\(block\|char\)-major-\([0-9]\+\)$/\1-major-\2-*/'`
 
-    echo "alias $2 $PA_ALIAS"
+    echo "alias $NAME $PA_ALIAS"
 }
 
 # Parse options: args modulename aliasto.
Signature · 1 line
OGAWA Hirofumi <[email protected]>
author Applied
#2 lore raw

In message <[email protected]> you write:

Quoted 8 lines

Rusty Russell <[email protected]> writes:

Umm.. Although I may be mis-understanding this problem, is the following scripts the not enough?

This does

	block-major-1 -> block-major-1-*

I've applied this, too, thanks.

But I still think we'll want the fallback for people who don't realize the change or used older generate-modprobe.conf versions.

Thanks!
Rusty.
--
Anyone who quotes me in their sig is an idiot. -- Rusty Russell.