Date: Tue, 2 Sep 2008 07:41:05 +0100 (BST) From: Robert Watson <rwatson@FreeBSD.org> To: Vladimir Grebenschikov <vova@fbsd.ru> Cc: Attilio Rao <attilio@FreeBSD.org>, cvs-src@FreeBSD.org, src-committers@FreeBSD.org, amistry@am-productions.biz, cvs-all@FreeBSD.org Subject: Re: cvs commit: VOP_ATTRIB ... and sysutils/fusefs-kmod Message-ID: <alpine.BSF.1.10.0809020739390.1150@fledge.watson.org> In-Reply-To: <1220333613.6499.5.camel@localhost> References: <200808281530.m7SFU3h7013986@repoman.freebsd.org> <1220333613.6499.5.camel@localhost>
next in thread | previous in thread | raw e-mail | index | archive | help
On Tue, 2 Sep 2008, Vladimir Grebenschikov wrote:
> On Thu, 2008-08-28 at 15:23 +0000, Attilio Rao wrote:
>> attilio 2008-08-28 15:23:18 UTC
>
> This patch makes CURRENT with Gnome badly crash (just crash while starting
> Gnome).
>
> I've found that it is due to ABI change, and tried to rebuild fusefs-kmod -
> it just fails due to changed interface.
>
> Patch below fixes problem for me. (but I still does not completely sure
> about td = curthread)
>
> Please fix fusefs-kmod port also.
>
> Side question, is there any way to prevent loading of kmod's with broken ABI
> ?
8-CURRENT kernels should reject old modules from loading unless
__FreeBSD_version hasn't been bumped. And indeed, it looks like this change
did not come with a __FreeBSD_version bump, which it should have as a
non-trivial change to the kernel KPI and KBI.
Robert N M Watson
Computer Laboratory
University of Cambridge
>
>> FreeBSD src repository
>>
>> Modified files:
>> share/man/man9 VOP_ATTRIB.9
> ...
>> Log:
>> SVN rev 182371 on 2008-08-28 15:23:18Z by attilio
>>
>> Decontextualize the couplet VOP_GETATTR / VOP_SETATTR as the passed thread
>> was always curthread and totally unuseful.
>>
>> Tested by: Giovanni Trematerra <giovanni dot trematerra at gmail dot com>
>
> --- fuse_module/fuse_vfsops.c.orig 2008-09-02 09:16:27.000000000 +0400
> +++ fuse_module/fuse_vfsops.c 2008-09-02 09:16:50.000000000 +0400
> @@ -668,7 +668,7 @@
> if (vp->v_type == VNON) {
> struct vattr va;
>
> - (void)VOP_GETATTR(vp, &va, td->td_ucred, td);
> + (void)VOP_GETATTR(vp, &va, td->td_ucred);
> }
> *vpp = vp;
> #if _DEBUG2G
> --- fuse_module/fuse_vnops.c.orig 2008-09-02 09:17:32.000000000 +0400
> +++ fuse_module/fuse_vnops.c 2008-09-02 09:21:02.000000000 +0400
> @@ -799,7 +799,7 @@
> struct vnode *vp = ap->a_vp;
> struct vattr *vap = ap->a_vap;
> struct ucred *cred = ap->a_cred;
> - struct thread *td = ap->a_td;
> + struct thread *td = curthread;
>
> struct fuse_dispatcher fdi;
> struct timespec uptsp;
> @@ -946,7 +946,7 @@
> /* We are to do the check in-kernel */
>
> if (! (facp->facc_flags & FACCESS_VA_VALID)) {
> - err = VOP_GETATTR(vp, VTOVA(vp), cred, td);
> + err = VOP_GETATTR(vp, VTOVA(vp), cred);
> if (err)
> return (err);
> facp->facc_flags |= FACCESS_VA_VALID;
> @@ -3005,7 +3005,7 @@
> struct vattr *vap = ap->a_vap;
> struct vnode *vp = ap->a_vp;
> struct ucred *cred = ap->a_cred;
> - struct thread *td = ap->a_td;
> + struct thread *td = curthread;
>
> int err = 0;
> struct fuse_dispatcher fdi;
> --- fuse_module/fuse_io.c.orig 2008-09-02 09:21:25.000000000 +0400
> +++ fuse_module/fuse_io.c 2008-09-02 09:21:49.000000000 +0400
> @@ -157,7 +157,7 @@
> goto out;
>
> if (uio->uio_rw == UIO_WRITE && fp->f_flag & O_APPEND) {
> - if ((err = VOP_GETATTR(vp, &va, cred, td)))
> + if ((err = VOP_GETATTR(vp, &va, cred)))
> goto out;
> uio->uio_offset = va.va_size;
> } else if ((flags & FOF_OFFSET) == 0)
> @@ -823,7 +823,7 @@
> #if FUSELIB_CONFORM_BIOREAD
> struct vattr va;
>
> - if ((err = VOP_GETATTR(vp, &va, cred, curthread)))
> + if ((err = VOP_GETATTR(vp, &va, cred)))
> goto out;
> #endif
>
>
>
> --
> Vladimir B. Grebenschikov
> vova@fbsd.ru
>
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?alpine.BSF.1.10.0809020739390.1150>