From: Willy Tarreau <w@1wt.eu>
To: linux-kernel@vger.kernel.org, stable@vger.kernel.org
Cc: linux-scsi@vger.kernel.org, Jens Axboe <axboe@kernel.dk>,
James Bottomley <JBottomley@parallels.com>,
Paolo Bonzini <pbonzini@redhat.com>,
Linus Torvalds <torvalds@linux-foundation.org>,
Ben Hutchings <ben@decadent.org.uk>,
Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Subject: [PATCH 90/91] block: fail SCSI passthrough ioctls on partition devices
Date: Sun, 05 Feb 2012 23:11:19 +0100 [thread overview]
Message-ID: <20120205220953.312427771@pcw.home.local> (raw)
In-Reply-To: <0635750f5f06ed2ca212b91fcb5c4483@local>
2.6.27-longterm review patch. If anyone has any objections, please let us know.
------------------
commit 0bfc96cb77224736dfa35c3c555d37b3646ef35e upstream.
[ Changes with respect to 3.3: return -ENOTTY from scsi_verify_blk_ioctl
and -ENOIOCTLCMD from sd_compat_ioctl. ]
Linux allows executing the SG_IO ioctl on a partition or LVM volume, and
will pass the command to the underlying block device. This is
well-known, but it is also a large security problem when (via Unix
permissions, ACLs, SELinux or a combination thereof) a program or user
needs to be granted access only to part of the disk.
This patch lets partitions forward a small set of harmless ioctls;
others are logged with printk so that we can see which ioctls are
actually sent. In my tests only CDROM_GET_CAPABILITY actually occurred.
Of course it was being sent to a (partition on a) hard disk, so it would
have failed with ENOTTY and the patch isn't changing anything in
practice. Still, I'm treating it specially to avoid spamming the logs.
In principle, this restriction should include programs running with
CAP_SYS_RAWIO. If for example I let a program access /dev/sda2 and
/dev/sdb, it still should not be able to read/write outside the
boundaries of /dev/sda2 independent of the capabilities. However, for
now programs with CAP_SYS_RAWIO will still be allowed to send the
ioctls. Their actions will still be logged.
This patch does not affect the non-libata IDE driver. That driver
however already tests for bd != bd->bd_contains before issuing some
ioctl; it could be restricted further to forbid these ioctls even for
programs running with CAP_SYS_ADMIN/CAP_SYS_RAWIO.
Cc: linux-scsi@vger.kernel.org
Cc: Jens Axboe <axboe@kernel.dk>
Cc: James Bottomley <JBottomley@parallels.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
[ Make it also print the command name when warning - Linus ]
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
[bwh: Backport to 2.6.32 - ENOIOCTLCMD does not get converted to
ENOTTY, so we must return ENOTTY directly]
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[wt: no scsi_cmd_blk_ioctl in 2.6.27, change callers instead. cciss is OK,
ub,virtio_blk,ide-floppy,sd need fixing, cdrom&st can be ignored ]
---
block/scsi_ioctl.c | 45 +++++++++++++++++++++++++++++++++++++++++++++
drivers/scsi/sd.c | 11 +++++++++--
include/linux/blkdev.h | 1 +
3 files changed, 55 insertions(+), 2 deletions(-)
Index: longterm-2.6.27/block/scsi_ioctl.c
===================================================================
--- longterm-2.6.27.orig/block/scsi_ioctl.c 2012-02-05 22:34:32.181915026 +0100
+++ longterm-2.6.27/block/scsi_ioctl.c 2012-02-05 22:39:32.090915159 +0100
@@ -653,3 +653,43 @@
}
EXPORT_SYMBOL(scsi_cmd_ioctl);
+
+int scsi_verify_blk_ioctl(struct block_device *bd, unsigned int cmd)
+{
+ if (bd && bd == bd->bd_contains)
+ return 0;
+
+ /* Actually none of these is particularly useful on a partition,
+ * but they are safe.
+ */
+ switch (cmd) {
+ case SCSI_IOCTL_GET_IDLUN:
+ case SCSI_IOCTL_GET_BUS_NUMBER:
+ case SCSI_IOCTL_GET_PCI:
+ case SCSI_IOCTL_PROBE_HOST:
+ case SG_GET_VERSION_NUM:
+ case SG_SET_TIMEOUT:
+ case SG_GET_TIMEOUT:
+ case SG_GET_RESERVED_SIZE:
+ case SG_SET_RESERVED_SIZE:
+ case SG_EMULATED_HOST:
+ return 0;
+ case CDROM_GET_CAPABILITY:
+ /* Keep this until we remove the printk below. udev sends it
+ * and we do not want to spam dmesg about it. CD-ROMs do
+ * not have partitions, so we get here only for disks.
+ */
+ return -ENOTTY;
+ default:
+ break;
+ }
+
+ /* In particular, rule out all resets and host-specific ioctls. */
+ if (printk_ratelimit())
+ printk(KERN_WARNING
+ "%s: sending ioctl %x to a partition!\n", current->comm, cmd);
+
+ return capable(CAP_SYS_RAWIO) ? 0 : -ENOTTY;
+}
+EXPORT_SYMBOL(scsi_verify_blk_ioctl);
+
Index: longterm-2.6.27/drivers/scsi/sd.c
===================================================================
--- longterm-2.6.27.orig/drivers/scsi/sd.c 2012-02-05 22:34:32.191915216 +0100
+++ longterm-2.6.27/drivers/scsi/sd.c 2012-02-05 22:34:47.717916292 +0100
@@ -749,6 +749,10 @@
SCSI_LOG_IOCTL(1, printk("sd_ioctl: disk=%s, cmd=0x%x\n",
disk->disk_name, cmd));
+ error = scsi_verify_blk_ioctl(bdev, cmd);
+ if (error < 0)
+ return error;
+
/*
* If we are in the middle of error recovery, don't let anyone
* else try and use this device. Also, if error recovery fails, it
@@ -927,6 +931,11 @@
struct block_device *bdev = file->f_path.dentry->d_inode->i_bdev;
struct gendisk *disk = bdev->bd_disk;
struct scsi_device *sdev = scsi_disk(disk)->device;
+ int ret;
+
+ ret = scsi_verify_blk_ioctl(bdev, cmd);
+ if (ret < 0)
+ return -ENOIOCTLCMD;
/*
* If we are in the middle of error recovery, don't let anyone
@@ -938,8 +947,6 @@
return -ENODEV;
if (sdev->host->hostt->compat_ioctl) {
- int ret;
-
ret = sdev->host->hostt->compat_ioctl(sdev, cmd, (void __user *)arg);
return ret;
Index: longterm-2.6.27/include/linux/blkdev.h
===================================================================
--- longterm-2.6.27.orig/include/linux/blkdev.h 2012-02-05 22:34:32.200915634 +0100
+++ longterm-2.6.27/include/linux/blkdev.h 2012-02-05 22:34:47.724916417 +0100
@@ -669,6 +669,7 @@
extern void blk_plug_device_unlocked(struct request_queue *);
extern int blk_remove_plug(struct request_queue *);
extern void blk_recount_segments(struct request_queue *, struct bio *);
+extern int scsi_verify_blk_ioctl(struct block_device *, unsigned int);
extern int scsi_cmd_ioctl(struct file *, struct request_queue *,
struct gendisk *, unsigned int, void __user *);
extern int sg_scsi_ioctl(struct file *, struct request_queue *,
Index: longterm-2.6.27/drivers/block/ub.c
===================================================================
--- longterm-2.6.27.orig/drivers/block/ub.c 2012-02-05 22:34:32.188915006 +0100
+++ longterm-2.6.27/drivers/block/ub.c 2012-02-05 22:34:47.731916501 +0100
@@ -1731,7 +1731,11 @@
{
struct gendisk *disk = inode->i_bdev->bd_disk;
void __user *usermem = (void __user *) arg;
+ int ret;
+ ret = scsi_verify_blk_ioctl(inode->i_bdev, cmd);
+ if (ret < 0)
+ return ret;
return scsi_cmd_ioctl(filp, disk->queue, disk, cmd, usermem);
}
Index: longterm-2.6.27/drivers/block/virtio_blk.c
===================================================================
--- longterm-2.6.27.orig/drivers/block/virtio_blk.c 2012-02-05 22:34:32.185914704 +0100
+++ longterm-2.6.27/drivers/block/virtio_blk.c 2012-02-05 22:34:47.735917527 +0100
@@ -149,6 +149,12 @@
static int virtblk_ioctl(struct inode *inode, struct file *filp,
unsigned cmd, unsigned long data)
{
+ int ret;
+
+ ret = scsi_verify_blk_ioctl(inode->i_bdev, cmd);
+ if (ret < 0)
+ return ret;
+
return scsi_cmd_ioctl(filp, inode->i_bdev->bd_disk->queue,
inode->i_bdev->bd_disk, cmd,
(void __user *)data);
Index: longterm-2.6.27/drivers/ide/ide-floppy.c
===================================================================
--- longterm-2.6.27.orig/drivers/ide/ide-floppy.c 2012-02-05 22:34:32.195915402 +0100
+++ longterm-2.6.27/drivers/ide/ide-floppy.c 2012-02-05 22:34:47.741917013 +0100
@@ -1336,9 +1336,12 @@
* skip SCSI_IOCTL_SEND_COMMAND (deprecated)
* and CDROM_SEND_PACKET (legacy) ioctls
*/
- if (cmd != CDROM_SEND_PACKET && cmd != SCSI_IOCTL_SEND_COMMAND)
- err = scsi_cmd_ioctl(file, bdev->bd_disk->queue,
- bdev->bd_disk, cmd, argp);
+ if (cmd != CDROM_SEND_PACKET && cmd != SCSI_IOCTL_SEND_COMMAND) {
+ err = scsi_verify_blk_ioctl(bdev, cmd);
+ if (err == 0)
+ err = scsi_cmd_ioctl(file, bdev->bd_disk->queue,
+ bdev->bd_disk, cmd, argp);
+ }
else
err = -ENOTTY;
next prev parent reply other threads:[~2012-02-05 22:28 UTC|newest]
Thread overview: 106+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <0635750f5f06ed2ca212b91fcb5c4483@local>
2012-02-05 22:09 ` [PATCH 00/91] 2.6.27.60-longterm review Willy Tarreau
2012-02-05 22:09 ` [PATCH 01/91] UBIFS: fix master node recovery Willy Tarreau
2012-02-05 22:09 ` [PATCH 02/91] slub: fix panic with DISCONTIGMEM Willy Tarreau
2012-02-06 21:58 ` David Rientjes
2012-02-07 6:13 ` Willy Tarreau
2012-02-05 22:09 ` [PATCH 03/91] set memory ranges in N_NORMAL_MEMORY when onlined Willy Tarreau
2012-02-05 22:09 ` [PATCH 04/91] agp: fix arbitrary kernel memory writes Willy Tarreau
2012-02-05 22:09 ` [PATCH 05/91] agp: fix OOM and buffer overflow Willy Tarreau
2012-02-05 22:09 ` [PATCH 06/91] put stricter guards on queue dead checks Willy Tarreau
2012-02-05 22:09 ` [PATCH 07/91] mmc: sdhci-pci: Fix error case in sdhci_pci_probe_slot() Willy Tarreau
2012-02-05 22:09 ` [PATCH 08/91] mmc: sdhci: Check mrq->cmd in sdhci_tasklet_finish Willy Tarreau
2012-02-05 22:09 ` [PATCH 09/91] mmc: sdhci: Check mrq != NULL " Willy Tarreau
2012-02-05 22:09 ` [PATCH 10/91] af_unix: Only allow recv on connected seqpacket sockets Willy Tarreau
2012-02-05 22:10 ` [PATCH 11/91] ARM: 6891/1: prevent heap corruption in OABI semtimedop Willy Tarreau
2012-02-05 22:10 ` [PATCH 12/91] Open with O_CREAT flag set fails to open existing files on non writable directories Willy Tarreau
2012-02-05 22:10 ` [PATCH 13/91] fs/partitions/ldm.c: fix oops caused by corrupted partition table Willy Tarreau
2012-02-05 22:10 ` [PATCH 14/91] SUNRPC: fix NFS client over TCP hangs due to packet loss (Bug 16494) Willy Tarreau
2012-02-05 22:10 ` [PATCH 15/91] Fix corrupted OSF partition table parsing Willy Tarreau
2012-02-05 22:10 ` [PATCH 16/91] sata_via: Delay on vt6420 when starting ATAPI DMA write Willy Tarreau
2012-02-05 22:10 ` [PATCH 17/91] libata: set queue DMA alignment to sector size for ATAPI too Willy Tarreau
2012-02-05 22:10 ` [PATCH 18/91] usb: musb: core: set has_tt flag Willy Tarreau
2012-02-05 22:10 ` [PATCH 19/91] Validate size of EFI GUID partition entries Willy Tarreau
2012-02-05 22:10 ` [PATCH 20/91] libertas: fix cmdpendingq locking Willy Tarreau
2012-02-05 22:10 ` [PATCH 21/91] powerpc/oprofile: Handle events that raise an exception without overflowing Willy Tarreau
2012-02-05 22:10 ` [PATCH 22/91] ext3: Fix fs corruption when make_indexed_dir() fails Willy Tarreau
2012-02-05 22:10 ` [PATCH 23/91] Fix for buffer overflow in ldm_frag_add not sufficient Willy Tarreau
2012-02-05 22:10 ` [PATCH 24/91] seqlock: Dont smp_rmb in seqlock reader spin loop Willy Tarreau
2012-02-05 22:10 ` [PATCH 25/91] x86/amd-iommu: Fix 3 possible endless loops Willy Tarreau
2012-02-05 22:10 ` [PATCH 26/91] md: check ->hot_remove_disk when removing disk Willy Tarreau
2012-02-05 22:10 ` [PATCH 27/91] uvcvideo: Remove buffers from the queues when freeing Willy Tarreau
2012-02-05 22:10 ` [PATCH 28/91] cfq-iosched: fix locking around ioc->ioc_data assignment Willy Tarreau
2012-02-05 22:10 ` [PATCH 29/91] cfq-iosched: fix a rcu warning Willy Tarreau
2012-02-05 22:10 ` [PATCH 30/91] SUNRPC: Fix use of static variable in rpcb_getport_async Willy Tarreau
2012-02-05 22:10 ` [PATCH 31/91] x86: Make Dell Latitude E5420 use reboot=pci Willy Tarreau
2012-02-05 22:10 ` [PATCH 32/91] libsas: remove expander from dev list on error Willy Tarreau
2012-02-05 23:48 ` Luben Tuikov
2012-02-06 0:52 ` Wanlong Gao
2012-02-06 1:14 ` Luben Tuikov
2012-02-06 6:25 ` Willy Tarreau
2012-02-05 22:10 ` [PATCH 33/91] powerpc/kdump: Fix timeout in crash_kexec_wait_realmode Willy Tarreau
2012-02-05 22:10 ` [PATCH 34/91] ext3: Fix oops in ext3_try_to_allocate_with_rsv() Willy Tarreau
2012-02-05 22:10 ` [PATCH 35/91] svcrpc: fix list-corrupting race on nfsd shutdown Willy Tarreau
2012-02-05 22:10 ` [PATCH 36/91] powerpc/pseries/hvconsole: Fix dropped console output Willy Tarreau
2012-02-05 22:10 ` [PATCH 37/91] alpha: fix several security issues Willy Tarreau
2012-02-05 22:10 ` [PATCH 38/91] ALSA: timer - Fix Oops at closing slave timer Willy Tarreau
2012-02-05 22:10 ` [PATCH 39/91] powerpc: Fix device tree claim code Willy Tarreau
2012-02-05 22:10 ` [PATCH 40/91] powerpc: pseries: Fix kexec on machines with more than 4TB of RAM Willy Tarreau
2012-02-05 22:10 ` [PATCH 41/91] xen/smp: Warn user why they keel over - nosmp or noapic and what to use instead Willy Tarreau
2012-02-06 16:50 ` Konrad Rzeszutek Wilk
2012-02-06 18:30 ` Willy Tarreau
2012-02-05 22:10 ` [PATCH 42/91] cifs: fix possible memory corruption in CIFSFindNext Willy Tarreau
2012-02-05 22:10 ` [PATCH 43/91] TPM: Call tpm_transmit with correct size Willy Tarreau
2012-02-05 22:10 ` [PATCH 44/91] TPM: Zero buffer after copying to userspace Willy Tarreau
2012-02-05 22:10 ` [PATCH 45/91] aacraid: reset should disable MSI interrupt Willy Tarreau
2012-02-05 22:10 ` [PATCH 46/91] libsas: fix panic when single phy is disabled on a wide port Willy Tarreau
2012-02-05 22:10 ` [PATCH 47/91] KVM: s390: check cpu_id prior to using it Willy Tarreau
2012-02-05 22:10 ` [PATCH 48/91] carminefb: Fix module parameters permissions Willy Tarreau
2012-02-05 22:10 ` [PATCH 49/91] um: fix ubd cow size Willy Tarreau
2012-02-05 22:10 ` [PATCH 50/91] NLM: Dont hang forever on NLM unlock requests Willy Tarreau
2012-02-05 22:10 ` [PATCH 51/91] Bluetooth: Prevent buffer overflow in l2cap config request Willy Tarreau
2012-02-05 22:10 ` [PATCH 52/91] net_sched: Fix qdisc_notify() Willy Tarreau
2012-02-05 22:10 ` [PATCH 53/91] ext4: fix BUG_ON() in ext4_ext_insert_extent() Willy Tarreau
2012-02-05 22:10 ` [PATCH 54/91] drivers/net/rionet.c: fix ethernet address macros for LE platforms Willy Tarreau
2012-02-05 22:10 ` [PATCH 55/91] Make scsi_free_queue() kill pending SCSI commands Willy Tarreau
2012-02-06 7:28 ` Bart Van Assche
2012-02-06 7:37 ` Willy Tarreau
2012-02-05 22:10 ` [PATCH 56/91] hfs: add sanity check for file name length Willy Tarreau
2012-02-05 22:10 ` [PATCH 57/91] USB: Fix Corruption issue in USB ftdi driver ftdi_sio.c Willy Tarreau
2012-02-05 22:10 ` [PATCH 58/91] oprofile, x86: Fix crash when unloading module (nmi timer mode) Willy Tarreau
2012-02-05 22:10 ` [PATCH 59/91] jbd/jbd2: validate sb->s_first in journal_get_superblock() Willy Tarreau
2012-02-05 22:10 ` [PATCH 60/91] Make TASKSTATS require root access Willy Tarreau
2012-02-05 22:10 ` [PATCH 61/91] hfs: fix hfs_find_init() sb->ext_tree NULL ptr oops Willy Tarreau
2012-02-05 22:10 ` [PATCH 62/91] [PATCH] x86, mm: Add __get_user_pages_fast() Willy Tarreau
2012-02-05 22:10 ` [PATCH 63/91] export __get_user_pages_fast() function Willy Tarreau
2012-02-05 22:10 ` [PATCH 64/91] oprofile, x86: Fix nmi-unsafe callgraph support Willy Tarreau
2012-02-05 22:10 ` [PATCH 65/91] ext4: avoid hangs in ext4_da_should_update_i_disksize() Willy Tarreau
2012-02-05 22:10 ` [PATCH 66/91] offb: Fix setting of the pseudo-palette for >8bpp Willy Tarreau
2012-02-05 22:10 ` [PATCH 67/91] offb: Fix bug in calculating requested vram size Willy Tarreau
2012-02-05 22:10 ` [PATCH 68/91] usb: usb-storage doesnt support dynamic id currently, the patch disables the feature to fix an oops Willy Tarreau
2012-02-05 22:10 ` [PATCH 69/91] SCSI: scsi_dh: check queuedata pointer before proceeding further Willy Tarreau
2012-02-05 22:10 ` [PATCH 70/91] ALSA: ice1724 - Check for ac97 to avoid kernel oops Willy Tarreau
2012-02-05 22:11 ` [PATCH 71/91] UBI: fix nameless volumes handling Willy Tarreau
2012-02-05 22:11 ` [PATCH 72/91] svcrpc: fix double-free on shutdown of nfsd after changing pool mode Willy Tarreau
2012-02-05 22:11 ` [PATCH 73/91] nfsd: Fix oops when parsing a 0 length export Willy Tarreau
2012-02-05 22:11 ` [PATCH 74/91] sym53c8xx: Fix NULL pointer dereference in slave_destroy Willy Tarreau
2012-02-05 22:11 ` [PATCH 75/91] [PATCH] bonding: correctly process non-linear skbs Willy Tarreau
2012-02-05 22:11 ` [PATCH 76/91] bonding: Ensure that we unshare skbs prior to calling pskb_may_pull Willy Tarreau
2012-02-05 22:11 ` [PATCH 77/91] block: add proper state guards to __elv_next_request Willy Tarreau
2012-02-05 22:11 ` [PATCH 78/91] x86, 64-bit: Fix copy_[to/from]_user() checks for the userspace address limit Willy Tarreau
2012-02-05 22:11 ` [PATCH 79/91] SCSI: scsi_lib: fix potential NULL dereference Willy Tarreau
2012-02-05 22:11 ` [PATCH 80/91] MAINTAINERS: stable: Update address Willy Tarreau
2012-02-05 22:11 ` [PATCH 81/91] af_packet: prevent information leak Willy Tarreau
2012-02-05 22:11 ` [PATCH 82/91] Fix time() inconsistencies caused by intermediate xtime_cache values being read Willy Tarreau
2012-02-05 22:11 ` [PATCH 83/91] net/ipv4: Check for mistakenly passed in non-IPv4 address Willy Tarreau
2012-02-05 22:11 ` [PATCH 84/91] x86: Fix mmap random address range Willy Tarreau
2012-02-05 22:11 ` [PATCH 85/91] i8k: Tell gcc that *regs gets clobbered Willy Tarreau
2012-02-05 22:11 ` [PATCH 86/91] Fix gcc 4.5.1 miscompiling drivers/char/i8k.c (again) Willy Tarreau
2012-02-05 22:11 ` [PATCH 87/91] kbuild: Disable -Wunused-but-set-variable for gcc 4.6.0 Willy Tarreau
2012-02-05 22:11 ` [PATCH 88/91] kbuild: Fix passing -Wno-* options to gcc 4.4+ Willy Tarreau
2012-02-05 22:11 ` [PATCH 89/91] i8k: Avoid lahf in 64-bit code Willy Tarreau
2012-02-05 22:11 ` Willy Tarreau [this message]
2012-02-05 22:44 ` [PATCH 90/91] block: fail SCSI passthrough ioctls on partition devices Paolo Bonzini
2012-02-05 22:53 ` Willy Tarreau
2012-02-07 10:03 ` Paolo Bonzini
2012-02-07 10:21 ` Willy Tarreau
2012-02-05 22:11 ` [PATCH 91/91] dm: do not forward ioctls from logical volumes to the underlying device Willy Tarreau
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20120205220953.312427771@pcw.home.local \
--to=w@1wt.eu \
--cc=JBottomley@parallels.com \
--cc=axboe@kernel.dk \
--cc=ben@decadent.org.uk \
--cc=gregkh@linuxfoundation.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-scsi@vger.kernel.org \
--cc=pbonzini@redhat.com \
--cc=stable@vger.kernel.org \
--cc=torvalds@linux-foundation.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®