On Wed, 6 Apr 2011 00:26:48 +0200
Luk Claes <luk@xxxxxxxxxx> wrote:
> When dev_name is constructed by combining slashes with the different
> parsed_mount_info fields, there can be an unnecessary appended slash
> compared to the original device string. So only convert backslashes
> to slashes to get the device name from the original device string.
>
> Signed-off-by: Luk Claes <luk@xxxxxxxxxx>
> ---
> mount.cifs.c | 15 +++++----------
> 1 files changed, 5 insertions(+), 10 deletions(-)
>
> diff --git a/mount.cifs.c b/mount.cifs.c
> index 8e1e32b..2cec7e4 100644
> --- a/mount.cifs.c
> +++ b/mount.cifs.c
> @@ -1824,10 +1824,7 @@ int main(int argc, char **argv)
> }
>
> /* lengths of different strings + slashes + trailing \0 */
> - dev_len = strnlen(parsed_info->host, sizeof(parsed_info->host)) +
> - strnlen(parsed_info->share, sizeof(parsed_info->share)) +
> - strnlen(parsed_info->prefix, sizeof(parsed_info->prefix)) +
> - 2 + 1 + 1 + 1;
> + dev_len = strlen(orig_dev) + 1;
> dev_name = calloc(dev_len, 1);
> if (!dev_name) {
> rc = EX_SYSERR;
> @@ -1835,12 +1832,10 @@ int main(int argc, char **argv)
> }
>
> /* rebuild device name with forward slashes */
> - strlcpy(dev_name, "//", dev_len);
> - strlcat(dev_name, parsed_info->host, dev_len);
> - strlcat(dev_name, "/", dev_len);
> - strlcat(dev_name, parsed_info->share, dev_len);
> - strlcat(dev_name, "/", dev_len);
> - strlcat(dev_name, parsed_info->prefix, dev_len);
> + strlcpy(dev_name, orig_dev, dev_len);
> + while (c = strcspn(dev_name, "\\") < dev_len - 1) {
> + dev_name[c] = '/';
> + }
>
> currentaddress = parsed_info->addrlist;
> nextaddress = strchr(currentaddress, ',');
Uh, no. That doesn't solve your problem. Imagine your fstab contains a
UNC line this:
\\\\server\\share
Then you're back to the same issue -- the devname in /proc/mounts
doesn't match the one in fstab.
I think you just want to pass in orig_dev to mount() as-is (and maybe
get rid of dev_name altogether). The fact that the kernel doesn't print
'\\' sanely in /proc/mounts is a separate (but related) issue, but
passing in the raw devname directly is the most future-proof solution
and shouldn't make the current situation any worse.
--
Jeff Layton <jlayton@xxxxxxxxx>
--
To unsubscribe from this list: send the line "unsubscribe linux-cifs" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
