linux-stable.git/drivers/infiniband, branch v3.18.32

IB/srpt: Simplify srpt_handle_tsk_mgmt()

2016-04-18T12:49:20+00:00

[ Upstream commit 51093254bf879bc9ce96590400a87897c7498463 ]

Let the target core check task existence instead of the SRP target
driver. Additionally, let the target core check the validity of the
task management request instead of the ib_srpt driver.

This patch fixes the following kernel crash:

BUG: unable to handle kernel NULL pointer dereference at 0000000000000001
IP: [] srpt_handle_new_iu+0x6d7/0x790 [ib_srpt]
Oops: 0002 [#1] SMP
Call Trace:
 [] srpt_process_completion+0xde/0x570 [ib_srpt]
 [] srpt_compl_thread+0x13f/0x160 [ib_srpt]
 [] kthread+0xcf/0xe0
 [] ret_from_fork+0x7c/0xb0

Signed-off-by: Bart Van Assche 
Fixes: 3e4f574857ee ("ib_srpt: Convert TMR path to target_submit_tmr")
Tested-by: Alex Estrin 
Reviewed-by: Christoph Hellwig 
Cc: Nicholas Bellinger 
Cc: Sagi Grimberg 
Cc: stable 
Signed-off-by: Doug Ledford 
Signed-off-by: Sasha Levin

iw_cxgb3: Fix incorrectly returning error on success

2016-04-14T00:44:36+00:00

[ Upstream commit 67f1aee6f45059fd6b0f5b0ecb2c97ad0451f6b3 ]

The cxgb3_*_send() functions return NET_XMIT_ values, which are
positive integers values. So don't treat positive return values
as an error.

Signed-off-by: Steve Wise 
Signed-off-by: Hariprasad Shenai 
Signed-off-by: Doug Ledford 
Signed-off-by: Sasha Levin

target: Remove first argument of target_{get,put}_sess_cmd()

2016-03-02T20:18:56+00:00

[ Upstream commit afc16604c06414223478df3e42301ab630b9960a ]

The first argument of these two functions is always identical
to se_cmd->se_sess. Hence remove the first argument.

Signed-off-by: Bart Van Assche 
Reviewed-by: Sagi Grimberg 
Reviewed-by: Christoph Hellwig 
Cc: Andy Grover 
Cc: 
Cc: Felipe Balbi 
Cc: Michael S. Tsirkin 
Signed-off-by: Nicholas Bellinger 
Signed-off-by: Sasha Levin

IB/qib: Support creating qps with GFP_NOIO flag

2016-02-10T03:57:35+00:00

[ Upstream commit fbbeb8632bf0b46ab44cfcedc4654cd7831b7161 ]

The current code is problematic when the QP creation and ipoib is used to
support NFS and NFS desires to do IO for paging purposes. In that case, the
GFP_KERNEL allocation in qib_qp.c causes a deadlock in tight memory
situations.

This fix adds support to create queue pair with GFP_NOIO flag for connected
mode only to cleanly fail the create queue pair in those situations.

Cc:  # 3.16+
Reviewed-by: Mike Marciniszyn 
Signed-off-by: Vinit Agnihotri 
Signed-off-by: Doug Ledford 
Signed-off-by: Sasha Levin

IB/qib: fix mcast detach when qp not attached

2016-02-10T03:57:35+00:00

[ Upstream commit 09dc9cd6528f5b52bcbd3292a6312e762c85260f ]

The code produces the following trace:

[1750924.419007] general protection fault: 0000 [#3] SMP
[1750924.420364] Modules linked in: nfnetlink autofs4 rpcsec_gss_krb5 nfsv4
dcdbas rfcomm bnep bluetooth nfsd auth_rpcgss nfs_acl dm_multipath nfs lockd
scsi_dh sunrpc fscache radeon ttm drm_kms_helper drm serio_raw parport_pc
ppdev i2c_algo_bit lpc_ich ipmi_si ib_mthca ib_qib dca lp parport ib_ipoib
mac_hid ib_cm i3000_edac ib_sa ib_uverbs edac_core ib_umad ib_mad ib_core
ib_addr tg3 ptp dm_mirror dm_region_hash dm_log psmouse pps_core
[1750924.420364] CPU: 1 PID: 8401 Comm: python Tainted: G D
3.13.0-39-generic #66-Ubuntu
[1750924.420364] Hardware name: Dell Computer Corporation PowerEdge
860/0XM089, BIOS A04 07/24/2007
[1750924.420364] task: ffff8800366a9800 ti: ffff88007af1c000 task.ti:
ffff88007af1c000
[1750924.420364] RIP: 0010:[] []
qib_mcast_qp_free+0x11/0x50 [ib_qib]
[1750924.420364] RSP: 0018:ffff88007af1dd70  EFLAGS: 00010246
[1750924.420364] RAX: 0000000000000001 RBX: ffff88007b822688 RCX:
000000000000000f
[1750924.420364] RDX: ffff88007b822688 RSI: ffff8800366c15a0 RDI:
6764697200000000
[1750924.420364] RBP: ffff88007af1dd78 R08: 0000000000000001 R09:
0000000000000000
[1750924.420364] R10: 0000000000000011 R11: 0000000000000246 R12:
ffff88007baa1d98
[1750924.420364] R13: ffff88003ecab000 R14: ffff88007b822660 R15:
0000000000000000
[1750924.420364] FS:  00007ffff7fd8740(0000) GS:ffff88007fc80000(0000)
knlGS:0000000000000000
[1750924.420364] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[1750924.420364] CR2: 00007ffff597c750 CR3: 000000006860b000 CR4:
00000000000007e0
[1750924.420364] Stack:
[1750924.420364]  ffff88007b822688 ffff88007af1ddf0 ffffffffa0132429
000000007af1de20
[1750924.420364]  ffff88007baa1dc8 ffff88007baa0000 ffff88007af1de70
ffffffffa00cb313
[1750924.420364]  00007fffffffde88 0000000000000000 0000000000000008
ffff88003ecab000
[1750924.420364] Call Trace:
[1750924.420364]  [] qib_multicast_detach+0x1e9/0x350
[ib_qib]
[1750924.568035]  [] ? ib_uverbs_modify_qp+0x323/0x3d0
[ib_uverbs]
[1750924.568035]  [] ib_detach_mcast+0x31/0x50 [ib_core]
[1750924.568035]  [] ib_uverbs_detach_mcast+0x93/0x170
[ib_uverbs]
[1750924.568035]  [] ib_uverbs_write+0xc6/0x2c0 [ib_uverbs]
[1750924.568035]  [] ? apparmor_file_permission+0x18/0x20
[1750924.568035]  [] ? security_file_permission+0x23/0xa0
[1750924.568035]  [] vfs_write+0xb4/0x1f0
[1750924.568035]  [] SyS_write+0x49/0xa0
[1750924.568035]  [] system_call_fastpath+0x1a/0x1f
[1750924.568035] Code: 66 2e 0f 1f 84 00 00 00 00 00 31 c0 5d c3 66 2e 0f 1f
84 00 00 00 00 00 66 90 0f 1f 44 00 00 55 48 89 e5 53 48 89 fb 48 8b 7f 10
 ff 8f 40 01 00 00 74 0e 48 89 df e8 8e f8 06 e1 5b 5d c3 0f
[1750924.568035] RIP  [] qib_mcast_qp_free+0x11/0x50
[ib_qib]
[1750924.568035]  RSP 
[1750924.650439] ---[ end trace 73d5d4b3f8ad4851 ]

The fix is to note the qib_mcast_qp that was found.   If none is found, then
return EINVAL indicating the error.

Cc: 
Reviewed-by: Dennis Dalessandro 
Reported-by: Jason Gunthorpe 
Signed-off-by: Mike Marciniszyn 
Signed-off-by: Doug Ledford 
Signed-off-by: Sasha Levin

IB/cm: Fix rb-tree duplicate free and use-after-free

2015-11-15T17:51:53+00:00

[ Upstream commit 0ca81a2840f77855bbad1b9f172c545c4dc9e6a4 ]

ib_send_cm_sidr_rep could sometimes erase the node from the sidr
(depending on errors in the process). Since ib_send_cm_sidr_rep is
called both from cm_sidr_req_handler and cm_destroy_id, cm_id_priv
could be either erased from the rb_tree twice or not erased at all.
Fixing that by making sure it's erased only once before freeing
cm_id_priv.

Fixes: a977049dacde ('[PATCH] IB: Add the kernel CM implementation')
Signed-off-by: Doron Tsur 
Signed-off-by: Matan Barak 
Signed-off-by: Doug Ledford 
Signed-off-by: Sasha Levin

iser-target: remove command with state ISTATE_REMOVE

2015-10-28T02:12:54+00:00

[ Upstream commit a4c15cd957cbd728f685645de7a150df5912591a ]

As documented in iscsit_sequence_cmd:
/*
 * Existing callers for iscsit_sequence_cmd() will silently
 * ignore commands with CMDSN_LOWER_THAN_EXP, so force this
 * return for CMDSN_MAXCMDSN_OVERRUN as well..
 */

We need to silently finish a command when it's in ISTATE_REMOVE.
This fixes an teardown hang we were seeing where a mis-behaved
initiator (triggered by allocation error injections) sent us a
cmdsn which was lower than expected.

Signed-off-by: Jenny Derzhavetz 
Signed-off-by: Sagi Grimberg 
Cc:  # v3.10+
Signed-off-by: Nicholas Bellinger 
Signed-off-by: Sasha Levin

IB/mlx4: Use correct SL on AH query under RoCE

2015-10-07T14:23:10+00:00

[ Upstream commit 5e99b139f1b68acd65e36515ca347b03856dfb5a ]

The mlx4 IB driver implementation for ib_query_ah used a wrong offset
(28 instead of 29) when link type is Ethernet. Fixed to use the correct one.

Fixes: fa417f7b520e ('IB/mlx4: Add support for IBoE')
Signed-off-by: Shani Michaeli 
Signed-off-by: Noa Osherovich 
Signed-off-by: Or Gerlitz 
Signed-off-by: Doug Ledford 
Signed-off-by: Sasha Levin

IB/mlx4: Forbid using sysfs to change RoCE pkeys

2015-10-07T14:22:45+00:00

[ Upstream commit 2b135db3e81301d0452e6aa107349abe67b097d6 ]

The pkey mapping for RoCE must remain the default mapping:
VFs:
  virtual index 0 = mapped to real index 0 (0xFFFF)
  All others indices: mapped to a real pkey index containing an
                      invalid pkey.
PF:
  virtual index i = real index i.

Don't allow users to change these mappings using files found in
sysfs.

Fixes: c1e7e466120b ('IB/mlx4: Add iov directory in sysfs under the ib device')
Signed-off-by: Jack Morgenstein 
Signed-off-by: Or Gerlitz 
Signed-off-by: Doug Ledford 
Signed-off-by: Sasha Levin

IB/uverbs: Fix race between ib_uverbs_open and remove_one

2015-10-07T14:22:05+00:00

[ Upstream commit 35d4a0b63dc0c6d1177d4f532a9deae958f0662c ]

Fixes: 2a72f212263701b927559f6850446421d5906c41 ("IB/uverbs: Remove dev_table")

Before this commit there was a device look-up table that was protected
by a spin_lock used by ib_uverbs_open and by ib_uverbs_remove_one. When
it was dropped and container_of was used instead, it enabled the race
with remove_one as dev might be freed just after:
dev = container_of(inode->i_cdev, struct ib_uverbs_device, cdev) but
before the kref_get.

In addition, this buggy patch added some dead code as
container_of(x,y,z) can never be NULL and so dev can never be NULL.
As a result the comment above ib_uverbs_open saying "the open method
will either immediately run -ENXIO" is wrong as it can never happen.

The solution follows Jason Gunthorpe suggestion from below URL:
https://www.mail-archive.com/linux-rdma@vger.kernel.org/msg25692.html

cdev will hold a kref on the parent (the containing structure,
ib_uverbs_device) and only when that kref is released it is
guaranteed that open will never be called again.

In addition, fixes the active count scheme to use an atomic
not a kref to prevent WARN_ON as pointed by above comment
from Jason.

Signed-off-by: Yishai Hadas 
Signed-off-by: Shachar Raindel 
Reviewed-by: Jason Gunthorpe 
Signed-off-by: Doug Ledford 
Signed-off-by: Sasha Levin