[Bugme-new] [Bug 4203] New: coda interacts badly with mount-program

[bugme-daemon at osdl.org]

http://bugme.osdl.org/show_bug.cgi?id=4203

           Summary: coda interacts badly with mount-program
    Kernel Version: 2.6.8 with Debian-patches
            Status: NEW
          Severity: normal
             Owner: fs_other at kernel-bugs.osdl.org
         Submitter: werner.baumann at onlinehome.de


Distribution: Debian sarge
Hardware Environment: amd athlon, via chipset
Software Environment: gcc 3.3.5, davfs2 0.2.3 modified for user-mount
Problem Description:
I use davfs, which uses the coda kernel module.
To enable user mounts without running setuid, I replaced the mount code from
davfs and exec the normal mount-programm instead. This works fine as long as the
device special file has minor number 0. It fails with any other minor number.
The reason is that coda does not get the information from the mount programm, it
wants, and in this case allways uses minor number 0. In file inode.c:

static int get_device_index(struct coda_mount_data *data)
{
	struct file *file;
	struct inode *inode;
	int idx;

	if(data == NULL) {
		printk("coda_read_super: Bad mount data\n");
		return -1;
	}

	if(data->version != CODA_MOUNT_VERSION) {
		printk("coda_read_super: Bad mount version\n");
		return -1;
	}

At this point it fails as mount seems not to deliver the desired version. It
returns to coda_fill_super() where idx is set to 0.
My questions:
- as it works fine with minor number 0, is this really necessary (for some
coda-venus related reasons)?
- in case, the data are not delivered by mount. Can't coda just try the device,
named by mount?
- is there some patched or newer mount-version that satisfies coda?
Steps to reproduce:
I might send my modified version of davfs. But I think it's the intended
behavior of coda, so there is no need to reproduce. My question is, whether it
is possible to change this behavior, or whether there would be bad effects if I
tried to change it.

Greetings
Werner

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.