linux.git/drivers/misc, branch v3.16.81

VMCI: Release resource if the work is already queued

2019-11-22T15:57:24+00:00

commit ba03a9bbd17b149c373c0ea44017f35fc2cd0f28 upstream.

Francois reported that VMware balloon gets stuck after a balloon reset,
when the VMCI doorbell is removed. A similar error can occur when the
balloon driver is removed with the following splat:

[ 1088.622000] INFO: task modprobe:3565 blocked for more than 120 seconds.
[ 1088.622035]       Tainted: G        W         5.2.0 #4
[ 1088.622087] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
[ 1088.622205] modprobe        D    0  3565   1450 0x00000000
[ 1088.622210] Call Trace:
[ 1088.622246]  __schedule+0x2a8/0x690
[ 1088.622248]  schedule+0x2d/0x90
[ 1088.622250]  schedule_timeout+0x1d3/0x2f0
[ 1088.622252]  wait_for_completion+0xba/0x140
[ 1088.622320]  ? wake_up_q+0x80/0x80
[ 1088.622370]  vmci_resource_remove+0xb9/0xc0 [vmw_vmci]
[ 1088.622373]  vmci_doorbell_destroy+0x9e/0xd0 [vmw_vmci]
[ 1088.622379]  vmballoon_vmci_cleanup+0x6e/0xf0 [vmw_balloon]
[ 1088.622381]  vmballoon_exit+0x18/0xcc8 [vmw_balloon]
[ 1088.622394]  __x64_sys_delete_module+0x146/0x280
[ 1088.622408]  do_syscall_64+0x5a/0x130
[ 1088.622410]  entry_SYSCALL_64_after_hwframe+0x44/0xa9
[ 1088.622415] RIP: 0033:0x7f54f62791b7
[ 1088.622421] Code: Bad RIP value.
[ 1088.622421] RSP: 002b:00007fff2a949008 EFLAGS: 00000206 ORIG_RAX: 00000000000000b0
[ 1088.622426] RAX: ffffffffffffffda RBX: 000055dff8b55d00 RCX: 00007f54f62791b7
[ 1088.622426] RDX: 0000000000000000 RSI: 0000000000000800 RDI: 000055dff8b55d68
[ 1088.622427] RBP: 000055dff8b55d00 R08: 00007fff2a947fb1 R09: 0000000000000000
[ 1088.622427] R10: 00007f54f62f5cc0 R11: 0000000000000206 R12: 000055dff8b55d68
[ 1088.622428] R13: 0000000000000001 R14: 000055dff8b55d68 R15: 00007fff2a94a3f0

The cause for the bug is that when the "delayed" doorbell is invoked, it
takes a reference on the doorbell entry and schedules work that is
supposed to run the appropriate code and drop the doorbell entry
reference. The code ignores the fact that if the work is already queued,
it will not be scheduled to run one more time. As a result one of the
references would not be dropped. When the code waits for the reference
to get to zero, during balloon reset or module removal, it gets stuck.

Fix it. Drop the reference if schedule_work() indicates that the work is
already queued.

Note that this bug got more apparent (or apparent at all) due to
commit ce664331b248 ("vmw_balloon: VMCI_DOORBELL_SET does not check status").

Fixes: 83e2ec765be03 ("VMCI: doorbell implementation.")
Reported-by: Francois Rigault 
Cc: Jorgen Hansen 
Cc: Adit Ranadive 
Cc: Alexios Zavras 
Cc: Vishnu DASA 
Signed-off-by: Nadav Amit 
Reviewed-by: Vishnu Dasa 
Link: https://lore.kernel.org/r/20190820202638.49003-1-namit@vmware.com
Signed-off-by: Greg Kroah-Hartman 
Signed-off-by: Ben Hutchings

VMCI: Fix integer overflow in VMCI handle arrays

2019-10-31T22:14:57+00:00

commit 1c2eb5b2853c9f513690ba6b71072d8eb65da16a upstream.

The VMCI handle array has an integer overflow in
vmci_handle_arr_append_entry when it tries to expand the array. This can be
triggered from a guest, since the doorbell link hypercall doesn't impose a
limit on the number of doorbell handles that a VM can create in the
hypervisor, and these handles are stored in a handle array.

In this change, we introduce a mandatory max capacity for handle
arrays/lists to avoid excessive memory usage.

Signed-off-by: Vishnu Dasa 
Reviewed-by: Adit Ranadive 
Reviewed-by: Jorgen Hansen 
Signed-off-by: Greg Kroah-Hartman 
Signed-off-by: Ben Hutchings

genwqe: Prevent an integer overflow in the ioctl

2019-10-05T15:19:46+00:00

commit 110080cea0d0e4dfdb0b536e7f8a5633ead6a781 upstream.

There are a couple potential integer overflows here.

	round_up(m->size + (m->addr & ~PAGE_MASK), PAGE_SIZE);

The first thing is that the "m->size + (...)" addition could overflow,
and the second is that round_up() overflows to zero if the result is
within PAGE_SIZE of the type max.

In this code, the "m->size" variable is an u64 but we're saving the
result in "map_size" which is an unsigned long and genwqe_user_vmap()
takes an unsigned long as well.  So I have used ULONG_MAX as the upper
bound.  From a practical perspective unsigned long is fine/better than
trying to change all the types to u64.

Fixes: eaf4722d4645 ("GenWQE Character device and DDCB queue")
Signed-off-by: Dan Carpenter 
Signed-off-by: Greg Kroah-Hartman 
Signed-off-by: Ben Hutchings

genwqe: Fix size check

2019-04-04T15:13:58+00:00

commit fdd669684655c07dacbdb0d753fd13833de69a33 upstream.

Calling the test program genwqe_cksum with the default buffer size of
2MB triggers the following kernel warning on s390:

WARNING: CPU: 30 PID: 9311 at mm/page_alloc.c:3189 __alloc_pages_nodemask+0x45c/0xbe0
CPU: 30 PID: 9311 Comm: genwqe_cksum Kdump: loaded Not tainted 3.10.0-957.el7.s390x #1
task: 00000005e5d13980 ti: 00000005e7c6c000 task.ti: 00000005e7c6c000
Krnl PSW : 0704c00180000000 00000000002780ac (__alloc_pages_nodemask+0x45c/0xbe0)
           R:0 T:1 IO:1 EX:1 Key:0 M:1 W:0 P:0 AS:3 CC:0 PM:0 EA:3
Krnl GPRS: 00000000002932b8 0000000000b73d7c 0000000000000010 0000000000000009
           0000000000000041 00000005e7c6f9b8 0000000000000001 00000000000080d0
           0000000000000000 0000000000b70500 0000000000000001 0000000000000000
           0000000000b70528 00000000007682c0 0000000000277df2 00000005e7c6f9a0
Krnl Code: 000000000027809e: de7195001000	ed	1280(114,%r9),0(%r1)
	   00000000002780a4: a774fead		brc	7,277dfe
	  #00000000002780a8: a7f40001		brc	15,2780aa
	  >00000000002780ac: 92011000		mvi	0(%r1),1
	   00000000002780b0: a7f4fea7		brc	15,277dfe
	   00000000002780b4: 9101c6b6		tm	1718(%r12),1
	   00000000002780b8: a784ff3a		brc	8,277f2c
	   00000000002780bc: a7f4fe2e		brc	15,277d18
Call Trace:
([<0000000000277df2>] __alloc_pages_nodemask+0x1a2/0xbe0)
 [<000000000013afae>] s390_dma_alloc+0xfe/0x310
 [<000003ff8065f362>] __genwqe_alloc_consistent+0xfa/0x148 [genwqe_card]
 [<000003ff80658f7a>] genwqe_mmap+0xca/0x248 [genwqe_card]
 [<00000000002b2712>] mmap_region+0x4e2/0x778
 [<00000000002b2c54>] do_mmap+0x2ac/0x3e0
 [<0000000000292d7e>] vm_mmap_pgoff+0xd6/0x118
 [<00000000002b081c>] SyS_mmap_pgoff+0xdc/0x268
 [<00000000002b0a34>] SyS_old_mmap+0x8c/0xb0
 [<000000000074e518>] sysc_tracego+0x14/0x1e
 [<000003ffacf87dc6>] 0x3ffacf87dc6

turns out the check in __genwqe_alloc_consistent uses "> MAX_ORDER"
while the mm code uses ">= MAX_ORDER". Fix genwqe.

Signed-off-by: Christian Borntraeger 
Signed-off-by: Frank Haverkamp 
Signed-off-by: Greg Kroah-Hartman 
Signed-off-by: Ben Hutchings

misc: vexpress: Off by one in vexpress_syscfg_exec()

2019-04-04T15:13:47+00:00

commit f8a70d8b889f180e6860cb1f85fed43d37844c5a upstream.

The > comparison should be >= to prevent reading beyond the end of the
func->template[] array.

(The func->template array is allocated in vexpress_syscfg_regmap_init()
and it has func->num_templates elements.)

Fixes: 974cc7b93441 ("mfd: vexpress: Define the device as MFD cells")
Signed-off-by: Dan Carpenter 
Acked-by: Sudeep Holla 
Signed-off-by: Greg Kroah-Hartman 
Signed-off-by: Ben Hutchings

altera-stapl: check for a null key before strcasecmp'ing it

2019-04-04T15:13:45+00:00

commit 9ccb645683ef46e3c52c12c088a368baa58447d4 upstream.

Currently the null check on key is occurring after the strcasecmp on
the key, hence there is a potential null pointer dereference on key.
Fix this by checking if key is null first. Also replace the == 0
check on strcasecmp with just the ! operator.

Detected by CoverityScan, CID#1248787 ("Dereference before null check")

Fixes: fa766c9be58b ("[media] Altera FPGA firmware download module")
Signed-off-by: Colin Ian King 
Signed-off-by: Greg Kroah-Hartman 
[bwh: Backported to 3.16: old code used strnicmp()]
Signed-off-by: Ben Hutchings

misc: atmel-ssc: Fix section annotation on atmel_ssc_get_driver_data

2019-02-11T17:53:59+00:00

commit 7c97301285b62a41d6bceded7d964085fc8cc50f upstream.

After building the kernel with Clang, the following section mismatch
warning appears:

WARNING: vmlinux.o(.text+0x3bf19a6): Section mismatch in reference from
the function ssc_probe() to the function
.init.text:atmel_ssc_get_driver_data()
The function ssc_probe() references
the function __init atmel_ssc_get_driver_data().
This is often because ssc_probe lacks a __init
annotation or the annotation of atmel_ssc_get_driver_data is wrong.

Remove __init from atmel_ssc_get_driver_data to get rid of the mismatch.

Signed-off-by: Nathan Chancellor 
Signed-off-by: Greg Kroah-Hartman 
Signed-off-by: Ben Hutchings

drivers/misc/sgi-gru: fix Spectre v1 vulnerability

2019-02-11T17:53:59+00:00

commit fee05f455ceb5c670cbe48e2f9454ebc4a388554 upstream.

req.gid can be indirectly controlled by user-space, hence leading to
a potential exploitation of the Spectre variant 1 vulnerability.

This issue was detected with the help of Smatch:

vers/misc/sgi-gru/grukdump.c:200 gru_dump_chiplet_request() warn:
potential spectre issue 'gru_base' [w]

Fix this by sanitizing req.gid before calling macro GID_TO_GRU, which
uses it to index gru_base.

Notice that given that speculation windows are large, the policy is
to kill the speculation on the first load and not worry if it can be
completed with a dependent load/store [1].

[1] https://marc.info/?l=linux-kernel&m=152449131114778&w=2

Signed-off-by: Gustavo A. R. Silva 
Signed-off-by: Greg Kroah-Hartman 
Signed-off-by: Ben Hutchings

VMCI: Resource wildcard match fixed

2019-02-11T17:53:25+00:00

commit 11924ba5e671d6caef1516923e2bd8c72929a3fe upstream.

When adding a VMCI resource, the check for an existing entry
would ignore that the new entry could be a wildcard. This could
result in multiple resource entries that would match a given
handle. One disastrous outcome of this is that the
refcounting used to ensure that delayed callbacks for VMCI
datagrams have run before the datagram is destroyed can be
wrong, since the refcount could be increased on the duplicate
entry. This in turn leads to a use after free bug. This issue
was discovered by Hangbin Liu using KASAN and syzkaller.

Fixes: bc63dedb7d46 ("VMCI: resource object implementation")
Reported-by: Hangbin Liu 
Reviewed-by: Adit Ranadive 
Reviewed-by: Vishnu Dasa 
Signed-off-by: Jorgen Hansen 
Signed-off-by: Greg Kroah-Hartman 
[bwh: Backported to 3.16: Drop the version change.]
Signed-off-by: Ben Hutchings

signal/GenWQE: Fix sending of SIGKILL

2019-02-11T17:53:18+00:00

commit 0ab93e9c99f8208c0a1a7b7170c827936268c996 upstream.

The genweq_add_file and genwqe_del_file by caching current without
using reference counting embed the assumption that a file descriptor
will never be passed from one process to another.  It even embeds the
assumption that the the thread that opened the file will be in
existence when the process terminates.   Neither of which are
guaranteed to be true.

Therefore replace caching the task_struct of the opener with
pid of the openers thread group id.  All the knowledge of the
opener is used for is as the target of SIGKILL and a SIGKILL
will kill the entire process group.

Rename genwqe_force_sig to genwqe_terminate, remove it's unncessary
signal argument, update it's ownly caller, and use kill_pid
instead of force_sig.

The work force_sig does in changing signal handling state is not
relevant to SIGKILL sent as SEND_SIG_PRIV.  The exact same processess
will be killed just with less work, and less confusion.  The work done
by force_sig is really only needed for handling syncrhonous
exceptions.

It will still be possible to cause genwqe_device_remove to wait
8 seconds by passing a file descriptor to another process but
the possible user after free is fixed.

Fixes: eaf4722d4645 ("GenWQE Character device and DDCB queue")
Cc: Greg Kroah-Hartman 
Cc: Frank Haverkamp 
Cc: Joerg-Stephan Vogt 
Cc: Michael Jung 
Cc: Michael Ruettger 
Cc: Kleber Sacilotto de Souza 
Cc: Sebastian Ott 
Cc: Eberhard S. Amann 
Cc: Gabriel Krisman Bertazi 
Cc: Guilherme G. Piccoli 
Signed-off-by: "Eric W. Biederman" 
Signed-off-by: Ben Hutchings