Re: [PATCH] mount -o loop race condition

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Karel Zak <kzak@xxxxxxxxxx> writes:

> On Tue, Jun 19, 2007 at 12:58:06PM +0200, Matthias Koenig wrote:
>> Hi, 
>> 
>> in addition to the race described recently 
>> (http://article.gmane.org/gmane.linux.utilities.util-linux-ng/241), 
>> there is also a race if you have concurring processes running 
>> mount -o loop. Patch below.
>
>  Good idea! We can also add the same logic to the losetup.c:main().

Ok, I will send a separate patch for this.

>
>  A few suggestions:
>
>> --- util-linux-devel.orig/mount/lomount.c
>> +++ util-linux-devel/mount/lomount.c
>> @@ -341,8 +341,15 @@ set_loop(const char *device, const char 
>>  	}
>>  
>>  	if (ioctl(fd, LOOP_SET_FD, ffd) < 0) {
>> -		perror("ioctl: LOOP_SET_FD");
>> -		return 1;
>> + 		switch (errno) {
>> + 		case EBUSY:
>> + 			if (verbose)
>> + 				perror("ioctl: LOOP_SET_FD");
>                 close (ffd);

Ah yes, and close(fd) is missing too.

>
>> + 			return 2;
>> + 		default:
>> + 			perror("ioctl: LOOP_SET_FD");
>                 close (ffd);
>
>> + 			return 1;
>> + 		}
>>  	}
>>  	close (ffd);
>
>
>> --- util-linux-devel.orig/mount/mount.c
>> +++ util-linux-devel/mount/mount.c
>> @@ -846,20 +846,34 @@ loop_check(const char **spec, const char
>
>
>> +        if (!*loopdev || !**loopdev)
>> +	  *loopdev = find_unused_loop_device();
>> +	if (!*loopdev)
>> +	  return EX_SYSERR;	/* no more loop devices */
>>  	if (verbose)
>> -	  printf(_("mount: failed setting up loop device\n"));
>> -	return EX_FAIL;
>> -      }
>> + 	  printf(_("mount: going to use the loop device %s\n"), *loopdev);
>> +        offset = opt_offset ? strtoull(opt_offset, NULL, 0) : 0;
>> + 	if (res = set_loop (*loopdev, *loopfile, offset,
>> +			    opt_encryption, pfd, &loopro)) {
>> + 	  switch(res) {
>> + 	  case 2:
>> + 		  /* loop dev has been grabbed by some other process,
>> + 		     try again */
>> + 		  if (verbose)
>> + 		    printf("mount: stolen loop=%s ...trying again\n", *loopdev);
>> + 		  *loopdev = NULL;
>               ^^^^^^^^^^^^^^
>   memory leak?
>
>     free(*loopdev);
>     *loopdev = NULL;

Right, find_unused_loop_device is using xstrdup

>
>> + 		  continue;
>               ^^^^^^^^
>
>  Please, don't use the find_unused_loop_device() if user explicitly
>  define a device.
>
>    if (!opt_loop) {
>        if (verbose)
>            printf("mount: stolen loop=%s ...trying again\n", *loopdev);
>        continue;
>    }
>    return EX_FAIL;
>
>  Because:
>
>      mount foo.img /foo -o loop=/dev/loop0

Ok.

Thanks for review, I will send a reworked patch.

Matthias

-
To unsubscribe from this list: send the line "unsubscribe util-linux-ng" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[Index of Archives]     [Netdev]     [Ethernet Bridging]     [Linux Wireless]     [Kernel Newbies]     [Security]     [Linux for Hams]     [Netfilter]     [Bugtraq]     [Yosemite News]     [MIPS Linux]     [ARM Linux]     [Linux RAID]     [Linux Admin]     [Samba]

  Powered by Linux