| CVE |
Vendors |
Products |
Updated |
CVSS v3.1 |
| In the Linux kernel, the following vulnerability has been resolved:
RDMA/irdma: Fix KASAN issue with tasklet
KASAN testing revealed the following issue assocated with freeing an IRQ.
[50006.466686] Call Trace:
[50006.466691] <IRQ>
[50006.489538] dump_stack+0x5c/0x80
[50006.493475] print_address_description.constprop.6+0x1a/0x150
[50006.499872] ? irdma_sc_process_ceq+0x483/0x790 [irdma]
[50006.505742] ? irdma_sc_process_ceq+0x483/0x790 [irdma]
[50006.511644] kasan_report.cold.11+0x7f/0x118
[50006.516572] ? irdma_sc_process_ceq+0x483/0x790 [irdma]
[50006.522473] irdma_sc_process_ceq+0x483/0x790 [irdma]
[50006.528232] irdma_process_ceq+0xb2/0x400 [irdma]
[50006.533601] ? irdma_hw_flush_wqes_callback+0x370/0x370 [irdma]
[50006.540298] irdma_ceq_dpc+0x44/0x100 [irdma]
[50006.545306] tasklet_action_common.isra.14+0x148/0x2c0
[50006.551096] __do_softirq+0x1d0/0xaf8
[50006.555396] irq_exit_rcu+0x219/0x260
[50006.559670] irq_exit+0xa/0x20
[50006.563320] smp_apic_timer_interrupt+0x1bf/0x690
[50006.568645] apic_timer_interrupt+0xf/0x20
[50006.573341] </IRQ>
The issue is that a tasklet could be pending on another core racing
the delete of the irq.
Fix by insuring any scheduled tasklet is killed after deleting the
irq. |
| In the Linux kernel, the following vulnerability has been resolved:
net: bridge: switchdev: Skip MDB replays of deferred events on offload
Before this change, generation of the list of MDB events to replay
would race against the creation of new group memberships, either from
the IGMP/MLD snooping logic or from user configuration.
While new memberships are immediately visible to walkers of
br->mdb_list, the notification of their existence to switchdev event
subscribers is deferred until a later point in time. So if a replay
list was generated during a time that overlapped with such a window,
it would also contain a replay of the not-yet-delivered event.
The driver would thus receive two copies of what the bridge internally
considered to be one single event. On destruction of the bridge, only
a single membership deletion event was therefore sent. As a
consequence of this, drivers which reference count memberships (at
least DSA), would be left with orphan groups in their hardware
database when the bridge was destroyed.
This is only an issue when replaying additions. While deletion events
may still be pending on the deferred queue, they will already have
been removed from br->mdb_list, so no duplicates can be generated in
that scenario.
To a user this meant that old group memberships, from a bridge in
which a port was previously attached, could be reanimated (in
hardware) when the port joined a new bridge, without the new bridge's
knowledge.
For example, on an mv88e6xxx system, create a snooping bridge and
immediately add a port to it:
root@infix-06-0b-00:~$ ip link add dev br0 up type bridge mcast_snooping 1 && \
> ip link set dev x3 up master br0
And then destroy the bridge:
root@infix-06-0b-00:~$ ip link del dev br0
root@infix-06-0b-00:~$ mvls atu
ADDRESS FID STATE Q F 0 1 2 3 4 5 6 7 8 9 a
DEV:0 Marvell 88E6393X
33:33:00:00:00:6a 1 static - - 0 . . . . . . . . . .
33:33:ff:87:e4:3f 1 static - - 0 . . . . . . . . . .
ff:ff:ff:ff:ff:ff 1 static - - 0 1 2 3 4 5 6 7 8 9 a
root@infix-06-0b-00:~$
The two IPv6 groups remain in the hardware database because the
port (x3) is notified of the host's membership twice: once via the
original event and once via a replay. Since only a single delete
notification is sent, the count remains at 1 when the bridge is
destroyed.
Then add the same port (or another port belonging to the same hardware
domain) to a new bridge, this time with snooping disabled:
root@infix-06-0b-00:~$ ip link add dev br1 up type bridge mcast_snooping 0 && \
> ip link set dev x3 up master br1
All multicast, including the two IPv6 groups from br0, should now be
flooded, according to the policy of br1. But instead the old
memberships are still active in the hardware database, causing the
switch to only forward traffic to those groups towards the CPU (port
0).
Eliminate the race in two steps:
1. Grab the write-side lock of the MDB while generating the replay
list.
This prevents new memberships from showing up while we are generating
the replay list. But it leaves the scenario in which a deferred event
was already generated, but not delivered, before we grabbed the
lock. Therefore:
2. Make sure that no deferred version of a replay event is already
enqueued to the switchdev deferred queue, before adding it to the
replay list, when replaying additions. |
| In the Linux kernel, the following vulnerability has been resolved:
netfilter: nf_tables: set dormant flag on hook register failure
We need to set the dormant flag again if we fail to register
the hooks.
During memory pressure hook registration can fail and we end up
with a table marked as active but no registered hooks.
On table/base chain deletion, nf_tables will attempt to unregister
the hook again which yields a warn splat from the nftables core. |
| In the Linux kernel, the following vulnerability has been resolved:
i40e: Do not allow untrusted VF to remove administratively set MAC
Currently when PF administratively sets VF's MAC address and the VF
is put down (VF tries to delete all MACs) then the MAC is removed
from MAC filters and primary VF MAC is zeroed.
Do not allow untrusted VF to remove primary MAC when it was set
administratively by PF.
Reproducer:
1) Create VF
2) Set VF interface up
3) Administratively set the VF's MAC
4) Put VF interface down
[root@host ~]# echo 1 > /sys/class/net/enp2s0f0/device/sriov_numvfs
[root@host ~]# ip link set enp2s0f0v0 up
[root@host ~]# ip link set enp2s0f0 vf 0 mac fe:6c:b5:da:c7:7d
[root@host ~]# ip link show enp2s0f0
23: enp2s0f0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP mode DEFAULT group default qlen 1000
link/ether 3c:ec:ef:b7:dd:04 brd ff:ff:ff:ff:ff:ff
vf 0 link/ether fe:6c:b5:da:c7:7d brd ff:ff:ff:ff:ff:ff, spoof checking on, link-state auto, trust off
[root@host ~]# ip link set enp2s0f0v0 down
[root@host ~]# ip link show enp2s0f0
23: enp2s0f0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP mode DEFAULT group default qlen 1000
link/ether 3c:ec:ef:b7:dd:04 brd ff:ff:ff:ff:ff:ff
vf 0 link/ether 00:00:00:00:00:00 brd ff:ff:ff:ff:ff:ff, spoof checking on, link-state auto, trust off |
| In the Linux kernel, the following vulnerability has been resolved:
cifs: fix underflow in parse_server_interfaces()
In this loop, we step through the buffer and after each item we check
if the size_left is greater than the minimum size we need. However,
the problem is that "bytes_left" is type ssize_t while sizeof() is type
size_t. That means that because of type promotion, the comparison is
done as an unsigned and if we have negative bytes left the loop
continues instead of ending. |
| In the Linux kernel, the following vulnerability has been resolved:
mptcp: fix data re-injection from stale subflow
When the MPTCP PM detects that a subflow is stale, all the packet
scheduler must re-inject all the mptcp-level unacked data. To avoid
acquiring unneeded locks, it first try to check if any unacked data
is present at all in the RTX queue, but such check is currently
broken, as it uses TCP-specific helper on an MPTCP socket.
Funnily enough fuzzers and static checkers are happy, as the accessed
memory still belongs to the mptcp_sock struct, and even from a
functional perspective the recovery completed successfully, as
the short-cut test always failed.
A recent unrelated TCP change - commit d5fed5addb2b ("tcp: reorganize
tcp_sock fast path variables") - exposed the issue, as the tcp field
reorganization makes the mptcp code always skip the re-inection.
Fix the issue dropping the bogus call: we are on a slow path, the early
optimization proved once again to be evil. |
| In the Linux kernel, the following vulnerability has been resolved:
net/sched: taprio: proper TCA_TAPRIO_TC_ENTRY_INDEX check
taprio_parse_tc_entry() is not correctly checking
TCA_TAPRIO_TC_ENTRY_INDEX attribute:
int tc; // Signed value
tc = nla_get_u32(tb[TCA_TAPRIO_TC_ENTRY_INDEX]);
if (tc >= TC_QOPT_MAX_QUEUE) {
NL_SET_ERR_MSG_MOD(extack, "TC entry index out of range");
return -ERANGE;
}
syzbot reported that it could fed arbitary negative values:
UBSAN: shift-out-of-bounds in net/sched/sch_taprio.c:1722:18
shift exponent -2147418108 is negative
CPU: 0 PID: 5066 Comm: syz-executor367 Not tainted 6.8.0-rc7-syzkaller-00136-gc8a5c731fd12 #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 02/29/2024
Call Trace:
<TASK>
__dump_stack lib/dump_stack.c:88 [inline]
dump_stack_lvl+0x1e7/0x2e0 lib/dump_stack.c:106
ubsan_epilogue lib/ubsan.c:217 [inline]
__ubsan_handle_shift_out_of_bounds+0x3c7/0x420 lib/ubsan.c:386
taprio_parse_tc_entry net/sched/sch_taprio.c:1722 [inline]
taprio_parse_tc_entries net/sched/sch_taprio.c:1768 [inline]
taprio_change+0xb87/0x57d0 net/sched/sch_taprio.c:1877
taprio_init+0x9da/0xc80 net/sched/sch_taprio.c:2134
qdisc_create+0x9d4/0x1190 net/sched/sch_api.c:1355
tc_modify_qdisc+0xa26/0x1e40 net/sched/sch_api.c:1776
rtnetlink_rcv_msg+0x885/0x1040 net/core/rtnetlink.c:6617
netlink_rcv_skb+0x1e3/0x430 net/netlink/af_netlink.c:2543
netlink_unicast_kernel net/netlink/af_netlink.c:1341 [inline]
netlink_unicast+0x7ea/0x980 net/netlink/af_netlink.c:1367
netlink_sendmsg+0xa3b/0xd70 net/netlink/af_netlink.c:1908
sock_sendmsg_nosec net/socket.c:730 [inline]
__sock_sendmsg+0x221/0x270 net/socket.c:745
____sys_sendmsg+0x525/0x7d0 net/socket.c:2584
___sys_sendmsg net/socket.c:2638 [inline]
__sys_sendmsg+0x2b0/0x3a0 net/socket.c:2667
do_syscall_64+0xf9/0x240
entry_SYSCALL_64_after_hwframe+0x6f/0x77
RIP: 0033:0x7f1b2dea3759
Code: 48 83 c4 28 c3 e8 d7 19 00 00 0f 1f 80 00 00 00 00 48 89 f8 48 89 f7 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 c7 c1 b8 ff ff ff f7 d8 64 89 01 48
RSP: 002b:00007ffd4de452f8 EFLAGS: 00000246 ORIG_RAX: 000000000000002e
RAX: ffffffffffffffda RBX: 00007f1b2def0390 RCX: 00007f1b2dea3759
RDX: 0000000000000000 RSI: 00000000200007c0 RDI: 0000000000000004
RBP: 0000000000000003 R08: 0000555500000000 R09: 0000555500000000
R10: 0000555500000000 R11: 0000000000000246 R12: 00007ffd4de45340
R13: 00007ffd4de45310 R14: 0000000000000001 R15: 00007ffd4de45340 |
| In the Linux kernel, the following vulnerability has been resolved:
vfio/pci: Create persistent INTx handler
A vulnerability exists where the eventfd for INTx signaling can be
deconfigured, which unregisters the IRQ handler but still allows
eventfds to be signaled with a NULL context through the SET_IRQS ioctl
or through unmask irqfd if the device interrupt is pending.
Ideally this could be solved with some additional locking; the igate
mutex serializes the ioctl and config space accesses, and the interrupt
handler is unregistered relative to the trigger, but the irqfd path
runs asynchronous to those. The igate mutex cannot be acquired from the
atomic context of the eventfd wake function. Disabling the irqfd
relative to the eventfd registration is potentially incompatible with
existing userspace.
As a result, the solution implemented here moves configuration of the
INTx interrupt handler to track the lifetime of the INTx context object
and irq_type configuration, rather than registration of a particular
trigger eventfd. Synchronization is added between the ioctl path and
eventfd_signal() wrapper such that the eventfd trigger can be
dynamically updated relative to in-flight interrupts or irqfd callbacks. |
| In the Linux kernel, the following vulnerability has been resolved:
vfio/pci: Lock external INTx masking ops
Mask operations through config space changes to DisINTx may race INTx
configuration changes via ioctl. Create wrappers that add locking for
paths outside of the core interrupt code.
In particular, irq_type is updated holding igate, therefore testing
is_intx() requires holding igate. For example clearing DisINTx from
config space can otherwise race changes of the interrupt configuration.
This aligns interfaces which may trigger the INTx eventfd into two
camps, one side serialized by igate and the other only enabled while
INTx is configured. A subsequent patch introduces synchronization for
the latter flows. |
| In the Linux kernel, the following vulnerability has been resolved:
netfilter: nft_chain_filter: handle NETDEV_UNREGISTER for inet/ingress basechain
Remove netdevice from inet/ingress basechain in case NETDEV_UNREGISTER
event is reported, otherwise a stale reference to netdevice remains in
the hook list. |
| In the Linux kernel, the following vulnerability has been resolved:
net: veth: clear GRO when clearing XDP even when down
veth sets NETIF_F_GRO automatically when XDP is enabled,
because both features use the same NAPI machinery.
The logic to clear NETIF_F_GRO sits in veth_disable_xdp() which
is called both on ndo_stop and when XDP is turned off.
To avoid the flag from being cleared when the device is brought
down, the clearing is skipped when IFF_UP is not set.
Bringing the device down should indeed not modify its features.
Unfortunately, this means that clearing is also skipped when
XDP is disabled _while_ the device is down. And there's nothing
on the open path to bring the device features back into sync.
IOW if user enables XDP, disables it and then brings the device
up we'll end up with a stray GRO flag set but no NAPI instances.
We don't depend on the GRO flag on the datapath, so the datapath
won't crash. We will crash (or hang), however, next time features
are sync'ed (either by user via ethtool or peer changing its config).
The GRO flag will go away, and veth will try to disable the NAPIs.
But the open path never created them since XDP was off, the GRO flag
was a stray. If NAPI was initialized before we'll hang in napi_disable().
If it never was we'll crash trying to stop uninitialized hrtimer.
Move the GRO flag updates to the XDP enable / disable paths,
instead of mixing them with the ndo_open / ndo_close paths. |
| In the Linux kernel, the following vulnerability has been resolved:
Bluetooth: Avoid potential use-after-free in hci_error_reset
While handling the HCI_EV_HARDWARE_ERROR event, if the underlying
BT controller is not responding, the GPIO reset mechanism would
free the hci_dev and lead to a use-after-free in hci_error_reset.
Here's the call trace observed on a ChromeOS device with Intel AX201:
queue_work_on+0x3e/0x6c
__hci_cmd_sync_sk+0x2ee/0x4c0 [bluetooth <HASH:3b4a6>]
? init_wait_entry+0x31/0x31
__hci_cmd_sync+0x16/0x20 [bluetooth <HASH:3b4a 6>]
hci_error_reset+0x4f/0xa4 [bluetooth <HASH:3b4a 6>]
process_one_work+0x1d8/0x33f
worker_thread+0x21b/0x373
kthread+0x13a/0x152
? pr_cont_work+0x54/0x54
? kthread_blkcg+0x31/0x31
ret_from_fork+0x1f/0x30
This patch holds the reference count on the hci_dev while processing
a HCI_EV_HARDWARE_ERROR event to avoid potential crash. |
| In the Linux kernel, the following vulnerability has been resolved:
iommufd: Fix iopt_access_list_id overwrite bug
Syzkaller reported the following WARN_ON:
WARNING: CPU: 1 PID: 4738 at drivers/iommu/iommufd/io_pagetable.c:1360
Call Trace:
iommufd_access_change_ioas+0x2fe/0x4e0
iommufd_access_destroy_object+0x50/0xb0
iommufd_object_remove+0x2a3/0x490
iommufd_object_destroy_user
iommufd_access_destroy+0x71/0xb0
iommufd_test_staccess_release+0x89/0xd0
__fput+0x272/0xb50
__fput_sync+0x4b/0x60
__do_sys_close
__se_sys_close
__x64_sys_close+0x8b/0x110
do_syscall_x64
The mismatch between the access pointer in the list and the passed-in
pointer is resulting from an overwrite of access->iopt_access_list_id, in
iopt_add_access(). Called from iommufd_access_change_ioas() when
xa_alloc() succeeds but iopt_calculate_iova_alignment() fails.
Add a new_id in iopt_add_access() and only update iopt_access_list_id when
returning successfully. |
| In the Linux kernel, the following vulnerability has been resolved:
iommufd: Fix protection fault in iommufd_test_syz_conv_iova
Syzkaller reported the following bug:
general protection fault, probably for non-canonical address 0xdffffc0000000038: 0000 [#1] SMP KASAN
KASAN: null-ptr-deref in range [0x00000000000001c0-0x00000000000001c7]
Call Trace:
lock_acquire
lock_acquire+0x1ce/0x4f0
down_read+0x93/0x4a0
iommufd_test_syz_conv_iova+0x56/0x1f0
iommufd_test_access_rw.isra.0+0x2ec/0x390
iommufd_test+0x1058/0x1e30
iommufd_fops_ioctl+0x381/0x510
vfs_ioctl
__do_sys_ioctl
__se_sys_ioctl
__x64_sys_ioctl+0x170/0x1e0
do_syscall_x64
do_syscall_64+0x71/0x140
This is because the new iommufd_access_change_ioas() sets access->ioas to
NULL during its process, so the lock might be gone in a concurrent racing
context.
Fix this by doing the same access->ioas sanity as iommufd_access_rw() and
iommufd_access_pin_pages() functions do. |
| In the Linux kernel, the following vulnerability has been resolved:
mptcp: fix double-free on socket dismantle
when MPTCP server accepts an incoming connection, it clones its listener
socket. However, the pointer to 'inet_opt' for the new socket has the same
value as the original one: as a consequence, on program exit it's possible
to observe the following splat:
BUG: KASAN: double-free in inet_sock_destruct+0x54f/0x8b0
Free of addr ffff888485950880 by task swapper/25/0
CPU: 25 PID: 0 Comm: swapper/25 Kdump: loaded Not tainted 6.8.0-rc1+ #609
Hardware name: Supermicro SYS-6027R-72RF/X9DRH-7TF/7F/iTF/iF, BIOS 3.0 07/26/2013
Call Trace:
<IRQ>
dump_stack_lvl+0x32/0x50
print_report+0xca/0x620
kasan_report_invalid_free+0x64/0x90
__kasan_slab_free+0x1aa/0x1f0
kfree+0xed/0x2e0
inet_sock_destruct+0x54f/0x8b0
__sk_destruct+0x48/0x5b0
rcu_do_batch+0x34e/0xd90
rcu_core+0x559/0xac0
__do_softirq+0x183/0x5a4
irq_exit_rcu+0x12d/0x170
sysvec_apic_timer_interrupt+0x6b/0x80
</IRQ>
<TASK>
asm_sysvec_apic_timer_interrupt+0x16/0x20
RIP: 0010:cpuidle_enter_state+0x175/0x300
Code: 30 00 0f 84 1f 01 00 00 83 e8 01 83 f8 ff 75 e5 48 83 c4 18 44 89 e8 5b 5d 41 5c 41 5d 41 5e 41 5f c3 cc cc cc cc fb 45 85 ed <0f> 89 60 ff ff ff 48 c1 e5 06 48 c7 43 18 00 00 00 00 48 83 44 2b
RSP: 0018:ffff888481cf7d90 EFLAGS: 00000202
RAX: 0000000000000000 RBX: ffff88887facddc8 RCX: 0000000000000000
RDX: 1ffff1110ff588b1 RSI: 0000000000000019 RDI: ffff88887fac4588
RBP: 0000000000000004 R08: 0000000000000002 R09: 0000000000043080
R10: 0009b02ea273363f R11: ffff88887fabf42b R12: ffffffff932592e0
R13: 0000000000000004 R14: 0000000000000000 R15: 00000022c880ec80
cpuidle_enter+0x4a/0xa0
do_idle+0x310/0x410
cpu_startup_entry+0x51/0x60
start_secondary+0x211/0x270
secondary_startup_64_no_verify+0x184/0x18b
</TASK>
Allocated by task 6853:
kasan_save_stack+0x1c/0x40
kasan_save_track+0x10/0x30
__kasan_kmalloc+0xa6/0xb0
__kmalloc+0x1eb/0x450
cipso_v4_sock_setattr+0x96/0x360
netlbl_sock_setattr+0x132/0x1f0
selinux_netlbl_socket_post_create+0x6c/0x110
selinux_socket_post_create+0x37b/0x7f0
security_socket_post_create+0x63/0xb0
__sock_create+0x305/0x450
__sys_socket_create.part.23+0xbd/0x130
__sys_socket+0x37/0xb0
__x64_sys_socket+0x6f/0xb0
do_syscall_64+0x83/0x160
entry_SYSCALL_64_after_hwframe+0x6e/0x76
Freed by task 6858:
kasan_save_stack+0x1c/0x40
kasan_save_track+0x10/0x30
kasan_save_free_info+0x3b/0x60
__kasan_slab_free+0x12c/0x1f0
kfree+0xed/0x2e0
inet_sock_destruct+0x54f/0x8b0
__sk_destruct+0x48/0x5b0
subflow_ulp_release+0x1f0/0x250
tcp_cleanup_ulp+0x6e/0x110
tcp_v4_destroy_sock+0x5a/0x3a0
inet_csk_destroy_sock+0x135/0x390
tcp_fin+0x416/0x5c0
tcp_data_queue+0x1bc8/0x4310
tcp_rcv_state_process+0x15a3/0x47b0
tcp_v4_do_rcv+0x2c1/0x990
tcp_v4_rcv+0x41fb/0x5ed0
ip_protocol_deliver_rcu+0x6d/0x9f0
ip_local_deliver_finish+0x278/0x360
ip_local_deliver+0x182/0x2c0
ip_rcv+0xb5/0x1c0
__netif_receive_skb_one_core+0x16e/0x1b0
process_backlog+0x1e3/0x650
__napi_poll+0xa6/0x500
net_rx_action+0x740/0xbb0
__do_softirq+0x183/0x5a4
The buggy address belongs to the object at ffff888485950880
which belongs to the cache kmalloc-64 of size 64
The buggy address is located 0 bytes inside of
64-byte region [ffff888485950880, ffff8884859508c0)
The buggy address belongs to the physical page:
page:0000000056d1e95e refcount:1 mapcount:0 mapping:0000000000000000 index:0xffff888485950700 pfn:0x485950
flags: 0x57ffffc0000800(slab|node=1|zone=2|lastcpupid=0x1fffff)
page_type: 0xffffffff()
raw: 0057ffffc0000800 ffff88810004c640 ffffea00121b8ac0 dead000000000006
raw: ffff888485950700 0000000000200019 00000001ffffffff 0000000000000000
page dumped because: kasan: bad access detected
Memory state around the buggy address:
ffff888485950780: fa fb fb
---truncated--- |
| In the Linux kernel, the following vulnerability has been resolved:
cxl/pci: Fix disabling memory if DVSEC CXL Range does not match a CFMWS window
The Linux CXL subsystem is built on the assumption that HPA == SPA.
That is, the host physical address (HPA) the HDM decoder registers are
programmed with are system physical addresses (SPA).
During HDM decoder setup, the DVSEC CXL range registers (cxl-3.1,
8.1.3.8) are checked if the memory is enabled and the CXL range is in
a HPA window that is described in a CFMWS structure of the CXL host
bridge (cxl-3.1, 9.18.1.3).
Now, if the HPA is not an SPA, the CXL range does not match a CFMWS
window and the CXL memory range will be disabled then. The HDM decoder
stops working which causes system memory being disabled and further a
system hang during HDM decoder initialization, typically when a CXL
enabled kernel boots.
Prevent a system hang and do not disable the HDM decoder if the
decoder's CXL range is not found in a CFMWS window.
Note the change only fixes a hardware hang, but does not implement
HPA/SPA translation. Support for this can be added in a follow on
patch series. |
| In the Linux kernel, the following vulnerability has been resolved:
mm/swap: fix race when skipping swapcache
When skipping swapcache for SWP_SYNCHRONOUS_IO, if two or more threads
swapin the same entry at the same time, they get different pages (A, B).
Before one thread (T0) finishes the swapin and installs page (A) to the
PTE, another thread (T1) could finish swapin of page (B), swap_free the
entry, then swap out the possibly modified page reusing the same entry.
It breaks the pte_same check in (T0) because PTE value is unchanged,
causing ABA problem. Thread (T0) will install a stalled page (A) into the
PTE and cause data corruption.
One possible callstack is like this:
CPU0 CPU1
---- ----
do_swap_page() do_swap_page() with same entry
<direct swapin path> <direct swapin path>
<alloc page A> <alloc page B>
swap_read_folio() <- read to page A swap_read_folio() <- read to page B
<slow on later locks or interrupt> <finished swapin first>
... set_pte_at()
swap_free() <- entry is free
<write to page B, now page A stalled>
<swap out page B to same swap entry>
pte_same() <- Check pass, PTE seems
unchanged, but page A
is stalled!
swap_free() <- page B content lost!
set_pte_at() <- staled page A installed!
And besides, for ZRAM, swap_free() allows the swap device to discard the
entry content, so even if page (B) is not modified, if swap_read_folio()
on CPU0 happens later than swap_free() on CPU1, it may also cause data
loss.
To fix this, reuse swapcache_prepare which will pin the swap entry using
the cache flag, and allow only one thread to swap it in, also prevent any
parallel code from putting the entry in the cache. Release the pin after
PT unlocked.
Racers just loop and wait since it's a rare and very short event. A
schedule_timeout_uninterruptible(1) call is added to avoid repeated page
faults wasting too much CPU, causing livelock or adding too much noise to
perf statistics. A similar livelock issue was described in commit
029c4628b2eb ("mm: swap: get rid of livelock in swapin readahead")
Reproducer:
This race issue can be triggered easily using a well constructed
reproducer and patched brd (with a delay in read path) [1]:
With latest 6.8 mainline, race caused data loss can be observed easily:
$ gcc -g -lpthread test-thread-swap-race.c && ./a.out
Polulating 32MB of memory region...
Keep swapping out...
Starting round 0...
Spawning 65536 workers...
32746 workers spawned, wait for done...
Round 0: Error on 0x5aa00, expected 32746, got 32743, 3 data loss!
Round 0: Error on 0x395200, expected 32746, got 32743, 3 data loss!
Round 0: Error on 0x3fd000, expected 32746, got 32737, 9 data loss!
Round 0 Failed, 15 data loss!
This reproducer spawns multiple threads sharing the same memory region
using a small swap device. Every two threads updates mapped pages one by
one in opposite direction trying to create a race, with one dedicated
thread keep swapping out the data out using madvise.
The reproducer created a reproduce rate of about once every 5 minutes, so
the race should be totally possible in production.
After this patch, I ran the reproducer for over a few hundred rounds and
no data loss observed.
Performance overhead is minimal, microbenchmark swapin 10G from 32G
zram:
Before: 10934698 us
After: 11157121 us
Cached: 13155355 us (Dropping SWP_SYNCHRONOUS_IO flag)
[kasong@tencent.com: v4] |
| In the Linux kernel, the following vulnerability has been resolved:
md: Don't ignore suspended array in md_check_recovery()
mddev_suspend() never stop sync_thread, hence it doesn't make sense to
ignore suspended array in md_check_recovery(), which might cause
sync_thread can't be unregistered.
After commit f52f5c71f3d4 ("md: fix stopping sync thread"), following
hang can be triggered by test shell/integrity-caching.sh:
1) suspend the array:
raid_postsuspend
mddev_suspend
2) stop the array:
raid_dtr
md_stop
__md_stop_writes
stop_sync_thread
set_bit(MD_RECOVERY_INTR, &mddev->recovery);
md_wakeup_thread_directly(mddev->sync_thread);
wait_event(..., !test_bit(MD_RECOVERY_RUNNING, &mddev->recovery))
3) sync thread done:
md_do_sync
set_bit(MD_RECOVERY_DONE, &mddev->recovery);
md_wakeup_thread(mddev->thread);
4) daemon thread can't unregister sync thread:
md_check_recovery
if (mddev->suspended)
return; -> return directly
md_read_sync_thread
clear_bit(MD_RECOVERY_RUNNING, &mddev->recovery);
-> MD_RECOVERY_RUNNING can't be cleared, hence step 2 hang;
This problem is not just related to dm-raid, fix it by ignoring
suspended array in md_check_recovery(). And follow up patches will
improve dm-raid better to frozen sync thread during suspend. |
| In the Linux kernel, the following vulnerability has been resolved:
md: Don't ignore read-only array in md_check_recovery()
Usually if the array is not read-write, md_check_recovery() won't
register new sync_thread in the first place. And if the array is
read-write and sync_thread is registered, md_set_readonly() will
unregister sync_thread before setting the array read-only. md/raid
follow this behavior hence there is no problem.
After commit f52f5c71f3d4 ("md: fix stopping sync thread"), following
hang can be triggered by test shell/integrity-caching.sh:
1) array is read-only. dm-raid update super block:
rs_update_sbs
ro = mddev->ro
mddev->ro = 0
-> set array read-write
md_update_sb
2) register new sync thread concurrently.
3) dm-raid set array back to read-only:
rs_update_sbs
mddev->ro = ro
4) stop the array:
raid_dtr
md_stop
stop_sync_thread
set_bit(MD_RECOVERY_INTR, &mddev->recovery);
md_wakeup_thread_directly(mddev->sync_thread);
wait_event(..., !test_bit(MD_RECOVERY_RUNNING, &mddev->recovery))
5) sync thread done:
md_do_sync
set_bit(MD_RECOVERY_DONE, &mddev->recovery);
md_wakeup_thread(mddev->thread);
6) daemon thread can't unregister sync thread:
md_check_recovery
if (!md_is_rdwr(mddev) &&
!test_bit(MD_RECOVERY_NEEDED, &mddev->recovery))
return;
-> -> MD_RECOVERY_RUNNING can't be cleared, hence step 4 hang;
The root cause is that dm-raid manipulate 'mddev->ro' by itself,
however, dm-raid really should stop sync thread before setting the
array read-only. Unfortunately, I need to read more code before I
can refacter the handler of 'mddev->ro' in dm-raid, hence let's fix
the problem the easy way for now to prevent dm-raid regression. |
| In the Linux kernel, the following vulnerability has been resolved:
dmaengine: idxd: Ensure safe user copy of completion record
If CONFIG_HARDENED_USERCOPY is enabled, copying completion record from
event log cache to user triggers a kernel bug.
[ 1987.159822] usercopy: Kernel memory exposure attempt detected from SLUB object 'dsa0' (offset 74, size 31)!
[ 1987.170845] ------------[ cut here ]------------
[ 1987.176086] kernel BUG at mm/usercopy.c:102!
[ 1987.180946] invalid opcode: 0000 [#1] PREEMPT SMP NOPTI
[ 1987.186866] CPU: 17 PID: 528 Comm: kworker/17:1 Not tainted 6.8.0-rc2+ #5
[ 1987.194537] Hardware name: Intel Corporation AvenueCity/AvenueCity, BIOS BHSDCRB1.86B.2492.D03.2307181620 07/18/2023
[ 1987.206405] Workqueue: wq0.0 idxd_evl_fault_work [idxd]
[ 1987.212338] RIP: 0010:usercopy_abort+0x72/0x90
[ 1987.217381] Code: 58 65 9c 50 48 c7 c2 17 85 61 9c 57 48 c7 c7 98 fd 6b 9c 48 0f 44 d6 48 c7 c6 b3 08 62 9c 4c 89 d1 49 0f 44 f3 e8 1e 2e d5 ff <0f> 0b 49 c7 c1 9e 42 61 9c 4c 89 cf 4d 89 c8 eb a9 66 66 2e 0f 1f
[ 1987.238505] RSP: 0018:ff62f5cf20607d60 EFLAGS: 00010246
[ 1987.244423] RAX: 000000000000005f RBX: 000000000000001f RCX: 0000000000000000
[ 1987.252480] RDX: 0000000000000000 RSI: ffffffff9c61429e RDI: 00000000ffffffff
[ 1987.260538] RBP: ff62f5cf20607d78 R08: ff2a6a89ef3fffe8 R09: 00000000fffeffff
[ 1987.268595] R10: ff2a6a89eed00000 R11: 0000000000000003 R12: ff2a66934849c89a
[ 1987.276652] R13: 0000000000000001 R14: ff2a66934849c8b9 R15: ff2a66934849c899
[ 1987.284710] FS: 0000000000000000(0000) GS:ff2a66b22fe40000(0000) knlGS:0000000000000000
[ 1987.293850] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 1987.300355] CR2: 00007fe291a37000 CR3: 000000010fbd4005 CR4: 0000000000f71ef0
[ 1987.308413] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[ 1987.316470] DR3: 0000000000000000 DR6: 00000000fffe07f0 DR7: 0000000000000400
[ 1987.324527] PKRU: 55555554
[ 1987.327622] Call Trace:
[ 1987.330424] <TASK>
[ 1987.332826] ? show_regs+0x6e/0x80
[ 1987.336703] ? die+0x3c/0xa0
[ 1987.339988] ? do_trap+0xd4/0xf0
[ 1987.343662] ? do_error_trap+0x75/0xa0
[ 1987.347922] ? usercopy_abort+0x72/0x90
[ 1987.352277] ? exc_invalid_op+0x57/0x80
[ 1987.356634] ? usercopy_abort+0x72/0x90
[ 1987.360988] ? asm_exc_invalid_op+0x1f/0x30
[ 1987.365734] ? usercopy_abort+0x72/0x90
[ 1987.370088] __check_heap_object+0xb7/0xd0
[ 1987.374739] __check_object_size+0x175/0x2d0
[ 1987.379588] idxd_copy_cr+0xa9/0x130 [idxd]
[ 1987.384341] idxd_evl_fault_work+0x127/0x390 [idxd]
[ 1987.389878] process_one_work+0x13e/0x300
[ 1987.394435] ? __pfx_worker_thread+0x10/0x10
[ 1987.399284] worker_thread+0x2f7/0x420
[ 1987.403544] ? _raw_spin_unlock_irqrestore+0x2b/0x50
[ 1987.409171] ? __pfx_worker_thread+0x10/0x10
[ 1987.414019] kthread+0x107/0x140
[ 1987.417693] ? __pfx_kthread+0x10/0x10
[ 1987.421954] ret_from_fork+0x3d/0x60
[ 1987.426019] ? __pfx_kthread+0x10/0x10
[ 1987.430281] ret_from_fork_asm+0x1b/0x30
[ 1987.434744] </TASK>
The issue arises because event log cache is created using
kmem_cache_create() which is not suitable for user copy.
Fix the issue by creating event log cache with
kmem_cache_create_usercopy(), ensuring safe user copy. |
