Re: [PATCH 2/2] btrfs-progs: canonicalize dm device name before update kernel

[Anand Jain <Anand.Jain@oracle.com>]

On 15/08/2014 12:30, Eryu Guan wrote:
> On Fri, Aug 15, 2014 at 09:50:34AM +0800, Anand Jain wrote:
>>
>> Eryu,
>>
>>   btrfs dev scan -d option is there for legacy reasons. The new method
>>   is using libblkid to find btrfs devs.
>>
>>     David/Zach, is it time to remove -d option ? or mention deprecated.
>>
>>
>>   But your test case show problem using btrfsck as well. thats nice!
>>   The fix for this is in the kernel, which would return busy
>>   if the device path is being updated when the device is mounted.
>>   Can you try with Chris integration branch ? mainly the patch..
>>
>> -------------
>> commit 4e5c146442b23437d23a2bd81b95f13dfeaffe88
>> Author: Anand Jain <anand.jain@oracle.com>
>> Date:   Thu Jul 3 18:22:05 2014 +0800
>>
>>      Btrfs: device_list_add() should not update list when mounted
>> -------------
>
> Thanks Anand, this patch fixed the issue, btrfsck reports "Device or
> resource busy" now.

  Thanks for testing.

> [root@hp-dl388g8-01 ~]# btrfsck /dev/mapper/rhel_hp--dl388g8--01-btrfs--2
> ERROR: device scan failed '/dev/dm-3' - Device or resource busy
> Checking filesystem on /dev/mapper/rhel_hp--dl388g8--01-btrfs--2
> UUID: 1104d6d6-2653-496b-8d67-184d522dd632
> checking extents
> checking free space cache
> cache and super generation don't match, space cache will be invalidated
> checking fs roots
> checking csums
> checking root refs
> found 114688 bytes used err is 0
> total csum bytes: 0
> total tree bytes: 114688
> total fs tree bytes: 32768
> total extent tree bytes: 16384
> btree space waste bytes: 108436
> file data blocks allocated: 0
>   referenced 0
> Btrfs v3.12
>
> But btrfsck is still scanning unrelated devices when checking a btrfs
> with multiple devices. In my case, I was checking btrfs on lv btrfs-2,
> but btrfs-1(dm-3) was scanned too.

  That's expected, as of now the only way to find a partner device is
  by scanning the available devices.

Anand



> Hope my first patch could fix this issue in an expected way.
>
> Thanks,
> Eryu
>>
>>
>> Thanks, Anand
>>
>>
>> On 14/08/2014 19:40, Eryu Guan wrote:
>>> A btrfsck or btrfs device scan -d operation could change the device
>>> name of other mounted btrfs in kernel, if the other btrfs is on lvm
>>> device.
>>>
>>> Assume that we have two btrfs filesystems, kernel is 3.16.0-rc4+
>>>
>>> [root@hp-dl388eg8-01 btrfs-progs]# btrfs fi show
>>> Label: none  uuid: 1aba7da5-ce2b-4af0-a716-db732abc60b2
>>>          Total devices 1 FS bytes used 384.00KiB
>>>          devid    1 size 15.00GiB used 2.04GiB path /dev/mapper/rhel_hp--dl388eg8--01-testlv1
>>>
>>> Label: none  uuid: 26ff4f12-f6d9-4cbc-aae2-57febeefde37
>>>          Total devices 2 FS bytes used 112.00KiB
>>>          devid    1 size 15.00GiB used 2.03GiB path /dev/mapper/rhel_hp--dl388eg8--01-testlv2
>>>          devid    2 size 15.00GiB used 2.01GiB path /dev/mapper/rhel_hp--dl388eg8--01-testlv3
>>>
>>> Btrfs v3.14.2
>>>
>>> And testlv1 was mounted at /mnt/btrfs
>>>
>>> [root@hp-dl388eg8-01 btrfs-progs]# df -TP /mnt/btrfs
>>> Filesystem                                Type  1024-blocks  Used Available Capacity Mounted on
>>> /dev/mapper/rhel_hp--dl388eg8--01-testlv1 btrfs    15728640   512  13602560       1% /mnt/btrfs
>>>
>>> Now run btrfsck on testlv2 or btrfs device scan -d, which will scan
>>> all btrfs devices and somehow change the device name.
>>>
>>> [root@hp-dl388eg8-01 btrfs-progs]# btrfsck /dev/mapper/rhel_hp--dl388eg8--01-testlv2 >/dev/null 2>&1
>>>
>>> [root@hp-dl388eg8-01 btrfs-progs]# df -TP /mnt/btrfs
>>> Filesystem     Type  1024-blocks  Used Available Capacity Mounted on
>>> /dev/dm-3      btrfs    15728640   512  13602560       1% /mnt/btrfs
>>> [root@hp-dl388eg8-01 btrfs-progs]# btrfs fi show
>>> Label: none  uuid: 1aba7da5-ce2b-4af0-a716-db732abc60b2
>>>          Total devices 1 FS bytes used 384.00KiB
>>>          devid    1 size 15.00GiB used 2.04GiB path /dev/dm-3
>>>
>>> Label: none  uuid: 26ff4f12-f6d9-4cbc-aae2-57febeefde37
>>>          Total devices 2 FS bytes used 112.00KiB
>>>          devid    1 size 15.00GiB used 2.03GiB path /dev/mapper/rhel_hp--dl388eg8--01-testlv2
>>>          devid    2 size 15.00GiB used 2.01GiB path /dev/mapper/rhel_hp--dl388eg8--01-testlv3
>>>
>>> Btrfs v3.14.2
>>>
>>> Now calling btrfs_register_one_device with canonicalized dm device name.
>>>
>>> Signed-off-by: Eryu Guan <guaneryu@gmail.com>
>>> ---
>>>
>>> With patch 1 applied, btrfsck won't change the device name,
>>> but btrfs device scan -d still does.
>>>
>>>   utils.c | 44 +++++++++++++++++++++++++++++++++++++++++---
>>>   1 file changed, 41 insertions(+), 3 deletions(-)
>>>
>>> diff --git a/utils.c b/utils.c
>>> index f54e749..3567094 100644
>>> --- a/utils.c
>>> +++ b/utils.c
>>> @@ -985,6 +985,32 @@ static int blk_file_in_dev_list(struct btrfs_fs_devices* fs_devices,
>>>   }
>>>
>>>   /*
>>> + * Convert dm-N device name to /dev/mapper/name
>>> + */
>>> +static void canonicalize_dm_name(char *devnode, char *path, int len)
>>> +{
>>> +	char *buf = NULL;
>>> +	FILE *sysfsp = NULL;
>>> +
>>> +	buf = malloc(PATH_MAX);
>>> +	if (!buf)
>>> +		return;
>>> +
>>> +	snprintf(buf, PATH_MAX, "/sys/block/%s/dm/name", devnode);
>>> +	sysfsp = fopen(buf, "r");
>>> +	if (!sysfsp)
>>> +		goto out;
>>> +
>>> +	if (fgets(buf, PATH_MAX, sysfsp)) {
>>> +		buf[strlen(buf) - 1] = '\0';
>>> +		snprintf(path, len - 1, "/dev/mapper/%s", buf);
>>> +	}
>>> +	fclose(sysfsp);
>>> +out:
>>> +	free(buf);
>>> +}
>>> +
>>> +/*
>>>    * returns 1 if the device was mounted, < 0 on error or 0 if everything
>>>    * is safe to continue.
>>>    */
>>> @@ -1189,6 +1215,9 @@ again:
>>>   		if (fsid && memcmp(fsid, tmp_devices->fsid, BTRFS_FSID_SIZE))
>>>   			continue;
>>>   		if (run_ioctl > 0) {
>>> +			if (!strncmp(fullpath, "/dev/dm-", strlen("/dev/dm-")))
>>> +				canonicalize_dm_name(dirent->d_name,
>>> +						     fullpath, PATH_MAX);
>>>   			btrfs_register_one_device(fullpath);
>>>   		}
>>>   		close(fd);
>>> @@ -1475,8 +1504,11 @@ int btrfs_scan_block_devices(u8 *fsid, int run_ioctl)
>>>   	int i;
>>>   	char buf[1024];
>>>   	char fullpath[110];
>>> +	char devnode[64];
>>> +	char *dirname = "/dev/";
>>>   	int scans = 0;
>>>   	int special;
>>> +	int is_dm;
>>>
>>>   scan_again:
>>>   	proc_partitions = fopen("/proc/partitions","r");
>>> @@ -1493,9 +1525,11 @@ scan_again:
>>>   			return -ENOENT;
>>>   		}
>>>
>>> -	strcpy(fullpath,"/dev/");
>>> -	while(fgets(buf, 1023, proc_partitions)) {
>>> -		i = sscanf(buf," %*d %*d %*d %99s", fullpath+5);
>>> +	strcpy(fullpath, dirname);
>>> +	while (fgets(buf, sizeof(buf) - 1, proc_partitions)) {
>>> +		i = sscanf(buf, " %*d %*d %*d %99s", devnode);
>>> +		strncpy(fullpath + strlen(dirname), devnode,
>>> +			sizeof(fullpath) - strlen(dirname));
>>>
>>>   		/*
>>>   		 * multipath and MD devices may register as a btrfs filesystem
>>> @@ -1504,6 +1538,7 @@ scan_again:
>>>   		 * This scans the special entries last
>>>   		 */
>>>   		special = strncmp(fullpath, "/dev/dm-", strlen("/dev/dm-")) == 0;
>>> +		is_dm = special;
>>>   		if (!special)
>>>   			special = strncmp(fullpath, "/dev/md", strlen("/dev/md")) == 0;
>>>
>>> @@ -1538,6 +1573,9 @@ scan_again:
>>>   		if (fsid && memcmp(fsid, tmp_devices->fsid, BTRFS_FSID_SIZE))
>>>   			continue;
>>>   		if (run_ioctl > 0) {
>>> +			if (is_dm)
>>> +				canonicalize_dm_name(devnode, fullpath,
>>> +						     sizeof(fullpath));
>>>   			btrfs_register_one_device(fullpath);
>>>   		}
>>>   	}
>>>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>