[Crash-utility] PATCH 00/10] teach crash to work with "live" ramdump

[Oleg Nesterov]

On 04/26, Dave Anderson wrote:
>
> > Even if this module was loaded and /dev/crash existed before I start /bin/crash?
> >
> > 	# ll /dev/crash
...
> > 	# crash ../VMLINUX
...
> > 	# ll /dev/crash
> > 	ls: cannot access /dev/crash: No such file or directory
> >
> > doesn't look friendly.
> >
> > And I can't use /bin/crash without root even if I do "chmod a+r /dev/crash" on my
> > machine.
> >
> > Is it all intentional?
>
> Yes.  The whole /dev/crash driver bullshit was put in place because of CONFIG_STRICT_DEVMEM.
> So we pretty much had to work around it by creating the read-only /dev/crash driver.
> Anyway, in the original discussions, the security folks didn't want /dev/crash hanging
> around unless it was actively being used by a root-only live crash session.  Later on it
> was decided that the driver should be built into the kernel.

I understand the problems with /dev/crash. But I don't think /bin/crash should remove
the file created by me.

> Please, just leave it be...   ;-)

OK.

But why /bin/crash REMOVES /dev/crash even if this file was NOT created at startup time?
Why get_live_memory_source() doesn't even check file_exists("/dev/crash") if
/lib/modules/%s/kernel/drivers/char/crash.ko exists?

Why I can't use /bin/crash on MY machine as non-root user even if I have the permission
to read /dev/crash?

OK, I'll probably keep the patch below for myself. It doesn't really fix the problems,
but it allows me to use "crash /dev/crash" on my machine without root.

Oleg.


--- x/filesys.c
+++ x/filesys.c
@@ -3859,8 +3859,7 @@ create_memory_device(dev_t dev)
 		/*
 		 *  It already exists -- just use it.
 		 */
-		if ((stat.st_mode == MEMORY_DRIVER_DEVICE_MODE) && 
-		    (stat.st_rdev == dev))
+		if (S_ISCHR(stat.st_mode) && (stat.st_rdev == dev))
 			return TRUE;
 
 		/*