linux-kernel - Re: NFS client broken in 4.1.0-rc2

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20150520232228.GC2067@n2100.arm.linux.org.uk>
Date:	Thu, 21 May 2015 00:22:28 +0100
From:	Russell King - ARM Linux <linux@....linux.org.uk>
To:	Chuck Lever <chuck.lever@...cle.com>
Cc:	Linux NFS Mailing List <linux-nfs@...r.kernel.org>,
	Trond Myklebust <trond.myklebust@...marydata.com>,
	linux-kernel@...r.kernel.org, linux-fsdevel@...r.kernel.org,
	Anna Schumaker <anna.schumaker@...app.com>,
	linux-arm-kernel@...ts.infradead.org
Subject: Re: NFS client broken in 4.1.0-rc2

On Fri, May 15, 2015 at 03:50:14PM +0100, Russell King - ARM Linux wrote:
> On Fri, May 15, 2015 at 10:26:48AM -0400, Chuck Lever wrote:
> > On May 15, 2015, at 10:24 AM, Russell King - ARM Linux <linux@....linux.org.uk> wrote:
> > > While trying to update a kernel and modules on one of my test systems,
> > > I was greeted by these errors:
> > > 
> > > tar: lib/modules/4.1.0-rc2+/kernel/drivers/media/platform/coda/coda.ko: Cannot utime
> > > tar: lib/modules/4.1.0-rc2+/kernel/drivers/media/dvb-frontends/drx39xyj/drx39xyj.ko: Cannot utime
> > > tar: lib/modules/4.1.0-rc2+/kernel/drivers/media/usb/em28xx/em28xx.ko: Cannot utime
> > > tar: lib/modules/4.1.0-rc2+/kernel/drivers/usb/serial/option.ko: Cannot utime
> > > tar: lib/modules/4.1.0-rc2+/kernel/drivers/usb/serial/ftdi_sio.ko: Cannot utime
> > > tar: lib/modules/4.1.0-rc2+/kernel/drivers/net/wireless/brcm80211/brcmfmac/brcmfmac.ko: Cannot utime
> > > tar: lib/modules/4.1.0-rc2+/kernel/drivers/input/mouse/psmouse.ko: Cannot utime
> > > tar: lib/modules/4.1.0-rc2+/kernel/fs/udf/udf.ko: Cannot utime
> > > tar: lib/modules/4.1.0-rc2+/kernel/fs/fuse/fuse.ko: Cannot utime
> > > tar: lib/modules/4.1.0-rc2+/kernel/fs/nfsd/nfsd.ko: Cannot utime
> > > tar: lib/modules/4.1.0-rc2+/kernel/sound/soc/codecs/snd-soc-wm8962.ko: Cannot utime
> > > tar: lib/modules/4.1.0-rc2+/kernel/net/bluetooth/bluetooth.ko: Cannot utime
> > > tar: lib/modules/4.1.0-rc2+/modules.alias.bin: Cannot utime
> > > tar: lib/modules/4.1.0-rc2+/modules.alias: Cannot utime
> > > tar: Exiting with failure status due to previous errors
> > > 
> > > Searching google wasn't helpful, as all the "Cannot utime" errors that
> > > google could find are followed by an errno string.
> > > 
> > > stracing at first sight didn't seem to be helpful, as no syscalls (apart
> > > from openat() with a pre-existing file) were failing.
> > > 
> > > Having recently updated to fc21 tar generating the archive, I thought
> > > maybe it was a tar format bug between fc21 tar and the target's tar.
> > > That was until I tried to "apt-get source tar" on the target, and was
> > > greeted by the same error.
> > > 
> > > So I then tried untaring the tar source archive onto a ramfs, which
> > > worked without complaint.  The difference being that it's a root NFS
> > > box, and so I was untaring onto NFS.
> > > 
> > > Here's the entry from /proc/mounts:
> > > 
> > > x.y.z.221:/var/boot/ci on / type nfs (rw,nolock,vers=4,addr=x.y.z.221,clientaddr=a.b.c.55)
> > > 
> > > Looking closer at the strace reveals this:
> > > 
> > > openat(AT_FDCWD, "lib/modules/4.1.0-rc2+/kernel/drivers/media/platform/coda/coda.ko", O_WRONLY|O_CREAT|O_EXCL|O_NOCTTY|O_NONBLOCK|O_LARGEFILE|O_CLOEXEC, 0600) = -1 EEXIST (File exists)
> > > unlinkat(AT_FDCWD, "lib/modules/4.1.0-rc2+/kernel/drivers/media/platform/coda/coda.ko", 0) = 0
> > > openat(AT_FDCWD, "lib/modules/4.1.0-rc2+/kernel/drivers/media/platform/coda/coda.ko", O_WRONLY|O_CREAT|O_EXCL|O_NOCTTY|O_NONBLOCK|O_LARGEFILE|O_CLOEXEC, 0600) = 4
> > > write(4, "\177ELF\1\1\1\0\0\0\0\0\0\0\0\0\1\0(\0\1\0\0\0\0\0\0\0\0\0\0\0"..., 4096) = 4096
> > > ...
> > > write(4, "\300H\0\0\34\345\1\0\314H\0\0\34\345\1\0\330H\0\0\34\345\1\0<I\0\0\34\370\1\0"..., 7312) = 7312
> > > dup2(4, 4)                              = 4
> > > fstat64(4, {st_mode=0757221, st_size=13181880119170311768, ...}) = 21
> > > write(2, "tar: ", 5)                    = 5
> > > write(2, "lib/modules/4.1.0-rc2+/kernel/dr"..., 79) = 79
> > > write(2, "\n", 1)                       = 1
> > > fchown32(4, 0, 0)                       = 0
> > > fchmod(4, 0664)                         = 0
> > > close(4)                                = 0
> > > 
> > > Look closely at that fstat64, and you'll notice that it's returning crap.
> > 
> > This is likely fixed by:
> > 
> >   http://marc.info/?l=linux-nfs&m=143095122604344&w=2
> 
> Yes, it appears to be fixed by that change.  I'll add it to my tree so
> my test systems work.

I notice NFS patches were merged by Linus recently, but this patch was not
in the set.  Is this patch scheduled to go in for 4.1?

-- 
FTTC broadband for 0.8mile line: currently at 10.5Mbps down 400kbps up
according to speedtest.net.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/