Date: Wed, 15 Nov 2006 08:03:35 -0600 From: Eric Anderson <anderson@centtech.com> To: Shane Adams <adamsch1@yahoo.com> Cc: freebsd-geom@freebsd.org Subject: Re: g_access questions (Out on a limb here) Message-ID: <455B1E37.3000201@centtech.com> In-Reply-To: <20061114230032.45377.qmail@web31801.mail.mud.yahoo.com> References: <20061114230032.45377.qmail@web31801.mail.mud.yahoo.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On 11/14/06 17:00, Shane Adams wrote:
> Hello,
>
> I have a question regarding g_access. I'm mounting a custom filesytem twice against the same device. The first time I'm mounting it with read/write access, I then mount the same device to a different mount point. After I do this, the first mount starts getting write errors, so somehow the second mount is screwing up access rights to the device. For what its worth, I'm mounting a ram backed mdconfig device.
>
> I'm not sure if this is a geom related issue, or related to the fact that you cannot mount a device more than twice in vfs?
>
> Anyway - the first time I mount (R/W) I do:
>
> g_access(cp, 1,1, 0)
>
> I then mount the device again as read only to a different mount point with
>
> g_access(cp,1,0,0)
>
> The code I used is taken from g_vfs_open with wr flag set depending if write access is required:
>
> pp = g_dev_getprovider(vp->v_rdev);
> if (pp == NULL)
> return (ENOENT);
> gp = g_new_geomf(&g_vfs_class2, "%s.%s", fsname, pp->name);
> cp = g_new_consumer(gp);
> g_attach(cp, pp);
> error = g_access(cp, 1, wr, 0);
> if (error) {
> g_wither_geom(gp, ENXIO);
> return (error);
> }
Interesting.. I'm mounting a custom file system also, and I'm doing it
much differently:
[...]
struct vnode *devvp; /* vnode of the mount device */
struct g_consumer *cp;
[...]
DROP_GIANT();
g_topology_lock();
error = g_vfs_open(devvp, &cp, "tarfs", 0);
g_topology_unlock();
PICKUP_GIANT();
VOP_UNLOCK(devvp, 0, td);
[...]
Eric
--
------------------------------------------------------------------------
Eric Anderson Sr. Systems Administrator Centaur Technology
Anything that works is better than anything that doesn't.
------------------------------------------------------------------------
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?455B1E37.3000201>