- 12 Mar, 2017 8 commits
-
-
Paul Burton authored
commit b6c7a324 upstream. get_frame_info() is meant to iterate over up to the first 128 instructions within a function, but for microMIPS kernels it will not reach that many instructions unless the function is 512 bytes long since we calculate the maximum number of instructions to check by dividing the function length by the 4 byte size of a union mips_instruction. In microMIPS kernels this won't do since instructions are variable length. Fix this by instead checking whether the pointer to the current instruction has reached the end of the function, and use max_insns as a simple constant to check the number of iterations against. Signed-off-by:
Paul Burton <paul.burton@imgtec.com> Fixes: 34c2f668 ("MIPS: microMIPS: Add unaligned access support.") Cc: Leonid Yegoshin <leonid.yegoshin@imgtec.com> Cc: linux-mips@linux-mips.org Patchwork: https://patchwork.linux-mips.org/patch/14530/Signed-off-by:
Ralf Baechle <ralf@linux-mips.org> Signed-off-by:
Greg Kroah-Hartman <gregkh@linuxfoundation.org>
-
Paul Burton authored
commit a3552dac upstream. During stack unwinding we call a number of functions to determine what type of instruction we're looking at. The union mips_instruction pointer provided to them may be pointing at a 2 byte, but not 4 byte, aligned address & we thus cannot directly access the 4 byte wide members of the union mips_instruction. To avoid this is_ra_save_ins() copies the required half-words of the microMIPS instruction to a correctly aligned union mips_instruction on the stack, which it can then access safely. The is_jump_ins() & is_sp_move_ins() functions do not correctly perform this temporary copy, and instead attempt to directly dereference 4 byte fields which may be misaligned and lead to an address exception. Fix this by copying the instruction halfwords to a temporary union mips_instruction in get_frame_info() such that we can provide a 4 byte aligned union mips_instruction to the is_*_ins() functions and they do not need to deal with misalignment themselves. Signed-off-by:
-
Paul Burton authored
commit ccaf7caf upstream. get_frame_info() can be called in microMIPS kernels with the ISA bit already clear. For example this happens when unwind_stack_by_address() is called because we begin with a PC that has the ISA bit set & subtract the (odd) offset from the preceding symbol (which does not have the ISA bit set). Since get_frame_info() unconditionally subtracts 1 from the PC in microMIPS kernels it incorrectly misaligns the address it then attempts to access code at, leading to an address error exception. Fix this by using msk_isa16_mode() to clear the ISA bit, which allows get_frame_info() to function regardless of whether it is provided with a PC that has the ISA bit set or not. Signed-off-by:
-
Felix Fietkau authored
commit 774f0c64 upstream. Disabling ethernet during reboot (only to enable it again when the ethernet driver attaches) can put the chip into a faulty state where it corrupts the header of all incoming packets. This happens if packets arrive during the time window where the core is disabled, and it can be easily reproduced by rebooting while sending a flood ping to the broadcast address. Fixes: 95135bfa ("MIPS: Lantiq: Deactivate most of the devices by default") Signed-off-by:
Felix Fietkau <nbd@nbd.name> Acked-by:
John Crispin <john@phrozen.org> Cc: hauke.mehrtens@lantiq.com Cc: linux-mips@linux-mips.org Patchwork: https://patchwork.linux-mips.org/patch/15078/Signed-off-by:
James Hogan <james.hogan@imgtec.com> Signed-off-by:
-
James Cowgill authored
commit 884b4269 upstream. If copy_from_user is called with a large buffer (>= 128 bytes) and the userspace buffer refers partially to unreadable memory, then it is possible for Octeon's copy_from_user to report the wrong number of bytes have been copied. In the case where the buffer size is an exact multiple of 128 and the fault occurs in the last 64 bytes, copy_from_user will report that all the bytes were copied successfully but leave some garbage in the destination buffer. The bug is in the main __copy_user_common loop in octeon-memcpy.S where in the middle of the loop, src and dst are incremented by 128 bytes. The l_exc_copy fault handler is used after this but that assumes that "src < THREAD_BUADDR($28)". This is not the case if src has already been incremented. Fix by adding an extra fault handler which rewinds the src and dst pointers 128 bytes before falling though to l_exc_copy. Thanks to the pwritev test from the strace test suite for originally highlighting this bug! Fixes: 5b3b1688 ("MIPS: Add Cavium OCTEON processor support ...") Signed-off-by:
James Cowgill <James.Cowgill@imgtec.com> Acked-by:
David Daney <david.daney@cavium.com> Reviewed-by:
-
Mirko Parthey authored
commit bdfdaf1a upstream. The Asus WL-500W buttons are active high, but the software treats them as active low. Fix the inverted logic. Fixes: 3be97255 ("MIPS: BCM47XX: Import buttons database from OpenWrt") Signed-off-by:
Mirko Parthey <mirko.parthey@web.de> Acked-by:
Rafał Miłecki <rafal@milecki.pl> Cc: Hauke Mehrtens <hauke@hauke-m.de> Cc: linux-mips@linux-mips.org Patchwork: https://patchwork.linux-mips.org/patch/15295/Signed-off-by:
-
Ralf Baechle authored
commit 66fd848c upstream. For certain arguments such as saddr = 0xc0a8fd60, daddr = 0xc0a8fda1, len = 80, proto = 17, sum = 0x7eae049d there will be a carry when folding the intermediate 64 bit checksum to 32 bit but the code doesn't add the carry back to the one's complement sum, thus an incorrect result will be generated. Reported-by:
Mark Zhang <bomb.zhang@gmail.com> Signed-off-by:
-
Purna Chandra Mandal authored
commit a726f1d2 upstream. Early clock API pic32_get_pbclk() is defined in early_clk.c and used by time.c and early_console.c. When CONFIG_EARLY_PRINTK isn't set, early_clk.c isn't compiled and time.c fails to link. Fix it by compiling early_clk.c always. Also sort files in alphabetical order. Fixes: 6e4ad1b4 ("MIPS: pic32mzda: fix getting timer clock rate.") Reported-by:
Harvey Hunt <harvey.hunt@imgtec.com> Signed-off-by:
Purna Chandra Mandal <purna.mandal@microchip.com> Reviewed-by:
-
- 26 Feb, 2017 32 commits
-
-
Greg Kroah-Hartman authored
-
Brian Foster authored
commit fa7f138a upstream. The buffered write failure handling code in xfs_file_iomap_end_delalloc() has a couple minor problems. First, if written == 0, start_fsb is not rounded down and it fails to kill off a delalloc block if the start offset is block unaligned. This results in a lingering delalloc block and broken delalloc block accounting detected at unmount time. Fix this by rounding down start_fsb in the unlikely event that written == 0. Second, it is possible for a failed overwrite of a delalloc extent to leave dirty pagecache around over a hole in the file. This is because is possible to hit ->iomap_end() on write failure before the iomap code has attempted to allocate pagecache, and thus has no need to clean it up. If the targeted delalloc extent was successfully written by a previous write, however, then it does still have dirty pages when ->iomap_end() punches out the underlying blocks. This ultimately results in writeback over a hole. To fix this problem, unconditionally punch out the pagecache from XFS before the associated delalloc range. Signed-off-by:
Brian Foster <bfoster@redhat.com> Reviewed-by:
Christoph Hellwig <hch@lst.de> Reviewed-by:
Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by:
-
Michael Schenk authored
commit 575ddce0 upstream. In the function rtl_usb_start we pre-allocate a certain number of urbs for RX path but they will not be freed when calling rtl_usb_stop. This results in leaking urbs when doing ifconfig up and down. Eventually, the system has no available urbs. Signed-off-by:
Michael Schenk <michael.schenk@albis-elcon.com> Signed-off-by:
Larry Finger <Larry.Finger@lwfinger.net> Signed-off-by:
Kalle Valo <kvalo@codeaurora.org> Signed-off-by:
-
Tejun Heo authored
commit 5f478e4e upstream. When !CONFIG_CGROUP_WRITEBACK, bdi has single bdi_writeback_congested at bdi->wb_congested. cgwb_bdi_init() allocates it with kzalloc() and doesn't do further initialization. This usually works fine as the reference count gets bumped to 1 by wb_init() and the put from wb_exit() releases it. However, when wb_init() fails, it puts the wb base ref automatically freeing the wb and the explicit kfree() in cgwb_bdi_init() error path ends up trying to free the same pointer the second time causing a double-free. Fix it by explicitly initilizing the refcnt to 1 and putting the base ref from cgwb_bdi_destroy(). Signed-off-by:
Tejun Heo <tj@kernel.org> Reported-by:
Dmitry Vyukov <dvyukov@google.com> Fixes: a13f35e8 ("writeback: don't embed root bdi_writeback_congested in bdi_writeback") Signed-off-by:
Jens Axboe <axboe@fb.com> Signed-off-by:
-
Lv Zheng authored
commit ffab9188 upstream. ACPICA commit b59347d0b8b676cb555fe8da5cad08fcd4eeb0d3 The following commit cleans up compiler specific inclusions: Commit: 9fa1cebd Subject: ACPICA: OSL: Cleanup the inclusion order of the compiler-specific headers But breaks one thing due to the following old issue: Buidling Linux kernel with Intel compiler originally depends on acgcc.h not acintel.h. So after making Intel compiler build working in ACPICA upstream by correctly using acintel.h, it becomes unable to build Linux kernel using Intel compiler as there is no acintel.h in the kernel source tree. This patch releases acintel.h to Linux kernel and fixes its inclusion in acenv.h. Fixes: 9fa1cebd (ACPICA: OSL: Cleanup the inclusion order of the compiler-specific headers) Link: https://github.com/acpica/acpica/commit/b59347d0Tested-by:
Stepan M Mishura <stepan.m.mishura@intel.com> Signed-off-by:
Lv Zheng <lv.zheng@intel.com> Signed-off-by:
Rafael J. Wysocki <rafael.j.wysocki@intel.com> Signed-off-by:
-
Jiri Kosina authored
commit dfe75ff8 upstream. Commit 3bb398d9 ("netfilter: nf_ct_helper: disable automatic helper assignment") is causing behavior regressions in firewalls, as traffic handled by conntrack helpers is now by default not passed through even though it was before due to missing CT targets (which were not necessary before this commit). The default had to be switched off due to security reasons [1] [2] and therefore should stay the way it is, but let's be friendly to firewall admins and issue a warning the first time we're in situation where packet would be likely passed through with the old default but we're likely going to drop it on the floor now. Rewrite the code a little bit as suggested by Linus, so that we avoid spaghettiing the code even more -- namely the whole decision making process regarding helper selection (either automatic or not) is being separated, so that the whole logic can be simplified and code (condition) duplication reduced. [1] https://cansecwest.com/csw12/conntrack-attack.pdf [2] https://home.regit.org/netfilter-en/secure-use-of-helpers/Signed-off-by:
Jiri Kosina <jkosina@suse.cz> Signed-off-by:
Pablo Neira Ayuso <pablo@netfilter.org> Signed-off-by:
-
Thomas Gleixner authored
commit 6cf18e69 upstream. This interrupt handler is broken in several ways: - It loops forever when the op code is not decodeable - It never returns IRQ_HANDLED because the only way to exit the loop returns IRQ_NONE unconditionally. The whole concept of this is broken. Creating devices in an interrupt handler is beyond any point of sanity. Make it at least behave halfways sane so accidental users do not have to deal with a hard to debug lockup. Fixes: e809c22b ("goldfish: add the goldfish virtual bus") Reported-by:
Gabriel C <nix.or.die@gmail.com> Signed-off-by:
Thomas Gleixner <tglx@linutronix.de> Acked-by:
Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by:
-
Thomas Gleixner authored
commit 47512cfd upstream. The goldfish platform code registers the platform device unconditionally which causes havoc in several ways if the goldfish_pdev_bus driver is enabled: - Access to the hardcoded physical memory region, which is either not available or contains stuff which is completely unrelated. - Prevents that the interrupt of the serial port can be requested - In case of a spurious interrupt it goes into a infinite loop in the interrupt handler of the pdev_bus driver (which needs to be fixed seperately). Add a 'goldfish' command line option to make the registration opt-in when the platform is compiled in. I'm seriously grumpy about this engineering trainwreck, which has seven SOBs from Intel developers for 50 lines of code. And none of them figured out that this is broken. Impressive fail! Fixes: ddd70cf9 ("goldfish: platform device for x86") Reported-by:
-
Johan Hovold authored
commit 14816b16 upstream. Since commit 4a510969 ("tty: Make tty_files_lock per-tty") a new tty_struct spin lock is taken in the tty release path, but the USB-serial-console hack was never updated hence leaving the lock of its "fake" tty uninitialised. This was eventually detected by lockdep. Make sure to initialise the new lock also for the fake tty to address this regression. Yes, this code is a mess, but cleaning it up is left for another day. Fixes: 4a510969 ("tty: Make tty_files_lock per-tty") Reviewed-by:
Johan Hovold <johan@kernel.org> Signed-off-by:
-
Johan Hovold authored
commit 9fef37d7 upstream. The current implementation failed to detect short transfers, something which could lead to bits of the uninitialised heap transfer buffer leaking to user space. Fixes: 149fc791 ("USB: ark3116: Setup some basic infrastructure for new ark3116 driver.") Fixes: f4c1e8d5 ("USB: ark3116: Make existing functions 16450-aware and add close and release functions.") Reviewed-by:
-
Johan Hovold authored
commit 2eee0502 upstream. The opticon driver used a control request at open to trigger a CTS status notification to be sent over the bulk-in pipe. When the driver was converted to using the generic read implementation, an inverted test prevented this request from being sent, something which could lead to TIOCMGET reporting an incorrect CTS state. Reported-by:
Dan Carpenter <dan.carpenter@oracle.com> Fixes: 7a6ee2b0 ("USB: opticon: switch to generic read implementation") Reviewed-by:
-
Johan Hovold authored
commit 5ed8d410 upstream. Make sure to detect short control transfers and return zero on success when retrieving the modem status. This fixes the TIOCMGET implementation which since e1ed212d ("USB: spcp8x5: add proper modem-status support") has returned TIOCM_LE on successful retrieval, and avoids leaking bits from the stack on short transfers. This also fixes the carrier-detect implementation which since the above mentioned commit unconditionally has returned true. Fixes: e1ed212d ("USB: spcp8x5: add proper modem-status support") Reviewed-by:
-
Johan Hovold authored
commit a6bb1e17 upstream. FTDI devices use a receive latency timer to periodically empty the receive buffer and report modem and line status (also when the buffer is empty). When a break or error condition is detected the corresponding status flags will be set on a packet with nonzero data payload and the flags are not updated until the break is over or further characters are received. In order to avoid over-reporting break and error conditions, these flags must therefore only be processed for packets with payload. This specifically fixes the case where after an overrun, the error condition is continuously reported and NULL-characters inserted until further data is received. Reported-by:
Michael Walle <michael@walle.cc> Fixes: 72fda3ca ("USB: serial: ftd_sio: implement sysrq handling on break") Fixes: 166ceb69 ("USB: ftdi_sio: clean up line-status handling") Signed-off-by:
-
Johan Hovold authored
commit c6dce262 upstream. Since commit 557aaa7f ("ft232: support the ASYNC_LOW_LATENCY flag") the FTDI driver has been using a receive latency-timer value of 1 ms instead of the device default of 16 ms. The latency timer is used to periodically empty a non-full receive buffer, but a status header is always sent when the timer expires including when the buffer is empty. This means that a two-byte bulk message is received every millisecond also for an otherwise idle port as long as it is open. Let's restore the pre-2009 behaviour which reduces the rate of the status messages to 1/16th (e.g. interrupt frequency drops from 1 kHz to 62.5 Hz) by not setting ASYNC_LOW_LATENCY by default. Anyone willing to pay the price for the minimum-latency behaviour should set the flag explicitly instead using the TIOCSSERIAL ioctl or a tool such as setserial (e.g. setserial /dev/ttyUSB0 low_latency). Note that since commit 0cbd81a9 ("USB: ftdi_sio: remove tty->low_latency") the ASYNC_LOW_LATENCY flag has no other effects but to set a minimal latency timer. Reported-by:
Antoine Aubert <a.aubert@overkiz.com> Fixes: 557aaa7f ("ft232: support the ASYNC_LOW_LATENCY flag") Signed-off-by:
-
Johan Hovold authored
commit 427c3a95 upstream. Make sure to detect short responses when fetching the modem status in order to avoid parsing uninitialised buffer data and having bits of it leak to user space. Note that we still allow for short 1-byte responses. Fixes: 1da177e4 ("Linux-2.6.12-rc2") Reviewed-by:
-
Ken Lin authored
commit 9a593656 upstream. Add new USB IDs for cp2104/5 devices on Bx50v3 boards due to the design change. Signed-off-by:
Ken Lin <yungching0725@gmail.com> Signed-off-by:
-
Johan Hovold authored
commit 5182c2cf upstream. Fix another NULL-pointer dereference at open should a malicious device lack an interrupt-in endpoint. Note that the driver has a broken check for an interrupt-in endpoint which means that an interrupt URB has never even been submitted. Fixes: 3f542974 ("USB: Moschip 7840 USB-Serial Driver") Reviewed-by:
-
Javier Martinez Canillas authored
commit abe81f3b upstream. If the driver is built as a module, autoload won't work because the module alias information is not filled. So user-space can't match the registered device with the corresponding module. Export the module alias information using the MODULE_DEVICE_TABLE() macro. Before this patch: $ modinfo drivers/tty/serial/msm_serial.ko | grep alias $ After this patch: $ modinfo drivers/tty/serial/msm_serial.ko | grep alias alias: of:N*T*Cqcom,msm-uartdmC* alias: of:N*T*Cqcom,msm-uartdm alias: of:N*T*Cqcom,msm-uartC* alias: of:N*T*Cqcom,msm-uart Signed-off-by:
Javier Martinez Canillas <javier@osg.samsung.com> Acked-by:
Bjorn Andersson <bjorn.andersson@linaro.org> Cc: stable <stable@vger.kernel.org> Signed-off-by:
-
Maxime Jayat authored
[ Upstream commit e623a9e9 ] Commit 34b88a68 ("net: Fix use after free in the recvmmsg exit path"), changed the exit path of recvmmsg to always return the datagrams variable and modified the error paths to set the variable to the error code returned by recvmsg if necessary. However in the case sock_error returned an error, the error code was then ignored, and recvmmsg returned 0. Change the error path of recvmmsg to correctly return the error code of sock_error. The bug was triggered by using recvmmsg on a CAN interface which was not up. Linux 4.6 and later return 0 in this case while earlier releases returned -ENETDOWN. Fixes: 34b88a68 ("net: Fix use after free in the recvmmsg exit path") Signed-off-by:
Maxime Jayat <maxime.jayat@mobile-devices.fr> Signed-off-by:
David S. Miller <davem@davemloft.net> Signed-off-by:
-
Paolo Abeni authored
[ Upstream commit ca4ef457 ] The skbs processed by ip_cmsg_recv() are not guaranteed to be linear e.g. when sending UDP packets over loopback with MSGMORE. Using csum_partial() on [potentially] the whole skb len is dangerous; instead be on the safe side and use skb_checksum(). Thanks to syzkaller team to detect the issue and provide the reproducer. v1 -> v2: - move the variable declaration in a tighter scope Fixes: ad6f939a ("ip: Add offset parameter to ip_cmsg_recv") Reported-by:
Andrey Konovalov <andreyknvl@google.com> Signed-off-by:
Paolo Abeni <pabeni@redhat.com> Acked-by:
Eric Dumazet <edumazet@google.com> Signed-off-by:
-
Michael S. Tsirkin authored
[ Upstream commit e7169530 ] Resizing currently drops consumer lock. This can cause entries to be reordered, which isn't good in itself. More importantly, consumer can detect a false ring empty condition and block forever. Further, nesting of consumer within producer lock is problematic for tun, since it produces entries in a BH, which causes a lock order reversal: CPU0 CPU1 ---- ---- consume: lock(&(&r->consumer_lock)->rlock); resize: local_irq_disable(); lock(&(&r->producer_lock)->rlock); lock(&(&r->consumer_lock)->rlock); <Interrupt> produce: lock(&(&r->producer_lock)->rlock); To fix, nest producer lock within consumer lock during resize, and keep consumer lock during the whole swap operation. Reported-by:
Jason Wang <jasowang@redhat.com> Signed-off-by:
Michael S. Tsirkin <mst@redhat.com> Signed-off-by:
-
David S. Miller authored
[ Upstream commit 4c03b862 ] A nested lock depth was added to the hasbin_delete() code but it doesn't actually work some well and results in tons of lockdep splats. Fix the code instead to properly drop the lock around the operation and just keep peeking the head of the hashbin queue. Reported-by:
-
Paolo Abeni authored
[ Upstream commit 22f0708a ] Since the commit 0c1d70af ("net: use dst_cache for vxlan device") vxlan_fill_metadata_dst() calls vxlan_get_route() passing a NULL dst_cache pointer, so the latter should explicitly check for valid dst_cache ptr. Unfortunately the commit d71785ff ("net: add dst_cache to ovs vxlan lwtunnel") removed said check. As a result is possible to trigger a null pointer access calling vxlan_fill_metadata_dst(), e.g. with: ovs-vsctl add-br ovs-br0 ovs-vsctl add-port ovs-br0 vxlan0 -- set interface vxlan0 \ type=vxlan options:remote_ip=192.168.1.1 \ options:key=1234 options:dst_port=4789 ofport_request=10 ip address add dev ovs-br0 172.16.1.2/24 ovs-vsctl set Bridge ovs-br0 ipfix=@i -- --id=@i create IPFIX \ targets=\"172.16.1.1:1234\" sampling=1 iperf -c 172.16.1.1 -u -l 1000 -b 10M -t 1 -p 1234 This commit addresses the issue passing to vxlan_get_route() the dst_cache already available into the lwt info processed by vxlan_fill_metadata_dst(). Fixes: d71785ff ("net: add dst_cache to ovs vxlan lwtunnel") Signed-off-by:
Jiri Benc <jbenc@redhat.com> Signed-off-by:
-
Andrey Konovalov authored
[ Upstream commit 5edabca9 ] In the current DCCP implementation an skb for a DCCP_PKT_REQUEST packet is forcibly freed via __kfree_skb in dccp_rcv_state_process if dccp_v6_conn_request successfully returns. However, if IPV6_RECVPKTINFO is set on a socket, the address of the skb is saved to ireq->pktopts and the ref count for skb is incremented in dccp_v6_conn_request, so skb is still in use. Nevertheless, it gets freed in dccp_rcv_state_process. Fix by calling consume_skb instead of doing goto discard and therefore calling __kfree_skb. Similar fixes for TCP: fb7e2399 [TCP]: skb is unexpectedly freed. 0aea76d3 tcp: SYN packets are now simply consumed Signed-off-by:
-
Marcus Huewe authored
[ Upstream commit 7627ae60 ] When setting a neigh related sysctl parameter, we always send a NETEVENT_DELAY_PROBE_TIME_UPDATE netevent. For instance, when executing sysctl net.ipv6.neigh.wlp3s0.retrans_time_ms=2000 a NETEVENT_DELAY_PROBE_TIME_UPDATE netevent is generated. This is caused by commit 2a4501ae ("neigh: Send a notification when DELAY_PROBE_TIME changes"). According to the commit's description, it was intended to generate such an event when setting the "delay_first_probe_time" sysctl parameter. In order to fix this, only generate this event when actually setting the "delay_first_probe_time" sysctl parameter. This fix should not have any unintended side-effects, because all but one registered netevent callbacks check for other netevent event types (the registered callbacks were obtained by grepping for "register_netevent_notifier"). The only callback that uses the NETEVENT_DELAY_PROBE_TIME_UPDATE event is mlxsw_sp_router_netevent_event() (in drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c): in case of this event, it only accesses the DELAY_PROBE_TIME of the passed neigh_parms. Fixes: 2a4501ae ("neigh: Send a notification when DELAY_PROBE_TIME changes") Signed-off-by:
Marcus Huewe <suse-tux@gmx.de> Reviewed-by:
Ido Schimmel <idosch@mellanox.com> Signed-off-by:
-
Anoob Soman authored
[ Upstream commit 2bd624b4 ] Commit 66644982 ("packet: call fanout_release, while UNREGISTERING a netdev"), unfortunately, introduced the following issues. 1. calling mutex_lock(&fanout_mutex) (fanout_release()) from inside rcu_read-side critical section. rcu_read_lock disables preemption, most often, which prohibits calling sleeping functions. [ ] include/linux/rcupdate.h:560 Illegal context switch in RCU read-side critical section! [ ] [ ] rcu_scheduler_active = 1, debug_locks = 0 [ ] 4 locks held by ovs-vswitchd/1969: [ ] #0: (cb_lock){++++++}, at: [<ffffffff8158a6c9>] genl_rcv+0x19/0x40 [ ] #1: (ovs_mutex){+.+.+.}, at: [<ffffffffa04878ca>] ovs_vport_cmd_del+0x4a/0x100 [openvswitch] [ ] #2: (rtnl_mutex){+.+.+.}, at: [<ffffffff81564157>] rtnl_lock+0x17/0x20 [ ] #3: (rcu_read_lock){......}, at: [<ffffffff81614165>] packet_notifier+0x5/0x3f0 [ ] [ ] Call Trace: [ ] [<ffffffff813770c1>] dump_stack+0x85/0xc4 [ ] [<ffffffff810c9077>] lockdep_rcu_suspicious+0x107/0x110 [ ] [<ffffffff810a2da7>] ___might_sleep+0x57/0x210 [ ] [<ffffffff810a2fd0>] __might_sleep+0x70/0x90 [ ] [<ffffffff8162e80c>] mutex_lock_nested+0x3c/0x3a0 [ ] [<ffffffff810de93f>] ? vprintk_default+0x1f/0x30 [ ] [<ffffffff81186e88>] ? printk+0x4d/0x4f [ ] [<ffffffff816106dd>] fanout_release+0x1d/0xe0 [ ] [<ffffffff81614459>] packet_notifier+0x2f9/0x3f0 2. calling mutex_lock(&fanout_mutex) inside spin_lock(&po->bind_lock). "sleeping function called from invalid context" [ ] BUG: sleeping function called from invalid context at kernel/locking/mutex.c:620 [ ] in_atomic(): 1, irqs_disabled(): 0, pid: 1969, name: ovs-vswitchd [ ] INFO: lockdep is turned off. [ ] Call Trace: [ ] [<ffffffff813770c1>] dump_stack+0x85/0xc4 [ ] [<ffffffff810a2f52>] ___might_sleep+0x202/0x210 [ ] [<ffffffff810a2fd0>] __might_sleep+0x70/0x90 [ ] [<ffffffff8162e80c>] mutex_lock_nested+0x3c/0x3a0 [ ] [<ffffffff816106dd>] fanout_release+0x1d/0xe0 [ ] [<ffffffff81614459>] packet_notifier+0x2f9/0x3f0 3. calling dev_remove_pack(&fanout->prot_hook), from inside spin_lock(&po->bind_lock) or rcu_read-side critical-section. dev_remove_pack() -> synchronize_net(), which might sleep. [ ] BUG: scheduling while atomic: ovs-vswitchd/1969/0x00000002 [ ] INFO: lockdep is turned off. [ ] Call Trace: [ ] [<ffffffff813770c1>] dump_stack+0x85/0xc4 [ ] [<ffffffff81186274>] __schedule_bug+0x64/0x73 [ ] [<ffffffff8162b8cb>] __schedule+0x6b/0xd10 [ ] [<ffffffff8162c5db>] schedule+0x6b/0x80 [ ] [<ffffffff81630b1d>] schedule_timeout+0x38d/0x410 [ ] [<ffffffff810ea3fd>] synchronize_sched_expedited+0x53d/0x810 [ ] [<ffffffff810ea6de>] synchronize_rcu_expedited+0xe/0x10 [ ] [<ffffffff8154eab5>] synchronize_net+0x35/0x50 [ ] [<ffffffff8154eae3>] dev_remove_pack+0x13/0x20 [ ] [<ffffffff8161077e>] fanout_release+0xbe/0xe0 [ ] [<ffffffff81614459>] packet_notifier+0x2f9/0x3f0 4. fanout_release() races with calls from different CPU. To fix the above problems, remove the call to fanout_release() under rcu_read_lock(). Instead, call __dev_remove_pack(&fanout->prot_hook) and netdev_run_todo will be happy that &dev->ptype_specific list is empty. In order to achieve this, I moved dev_{add,remove}_pack() out of fanout_{add,release} to __fanout_{link,unlink}. So, call to {,__}unregister_prot_hook() will make sure fanout->prot_hook is removed as well. Fixes: 66644982 ("packet: call fanout_release, while UNREGISTERING a netdev") Reported-by:
Anoob Soman <anoob.soman@citrix.com> Acked-by:
-
Eric Dumazet authored
[ Upstream commit d199fab6 ] Multiple threads can call fanout_add() at the same time. We need to grab fanout_mutex earlier to avoid races that could lead to one thread freeing po->rollover that was set by another thread. Do the same in fanout_release(), for peace of mind, and to help us finding lockdep issues earlier. Fixes: dc99f600 ("packet: Add fanout support.") Fixes: 0648ab70 ("packet: rollover prepare: per-socket state") Signed-off-by:
-
Ivan Khoronzhuk authored
[ Upstream commit a60ced99 ] There is a copy-paste error, which hides breaking of resume for CPSW driver: there was replaced netdev_priv() to ndev_to_cpsw(ndev) in suspend, but left it unchanged in resume. Fixes: 606f3993 (ti: cpsw: move platform data and slaves info to cpsw_common) Reported-by:
Alexey Starikovskiy <AStarikovskiy@topcon.com> Signed-off-by:
Ivan Khoronzhuk <ivan.khoronzhuk@linaro.org> Signed-off-by:
-
Eric Dumazet authored
[ Upstream commit 8b74d439 ] It seems nobody used LLC since linux-3.12. Fortunately fuzzers like syzkaller still know how to run this code, otherwise it would be no fun. Setting skb->sk without skb->destructor leads to all kinds of bugs, we now prefer to be very strict about it. Ideally here we would use skb_set_owner() but this helper does not exist yet, only CAN seems to have a private helper for that. Fixes: 376c7311 ("net: add a temporary sanity check in skb_orphan()") Signed-off-by:
-
Or Gerlitz authored
[ Upstream commit fed06ee8 ] When called by HW offloading drivers, the TC action (e.g net/sched/act_mirred.c) code uses this_cpu logic, e.g _bstats_cpu_update(this_cpu_ptr(a->cpu_bstats), bytes, packets) per the kernel documention, preemption should be disabled, add that. Before the fix, when running with CONFIG_PREEMPT set, we get a BUG: using smp_processor_id() in preemptible [00000000] code: tc/3793 asserion from the TC action (mirred) stats_update callback. Fixes: aad7e08d ('net/mlx5e: Hardware offloaded flower filter statistics support') Signed-off-by:
Or Gerlitz <ogerlitz@mellanox.com> Signed-off-by:
-
WANG Cong authored
[ Upstream commit cd27b96b ] In commit 98e3862c ("kcm: fix 0-length case for kcm_sendmsg()") I tried to avoid skb allocation for 0-length case, but missed a check for NULL pointer in the non EOR case. Fixes: 98e3862c ("kcm: fix 0-length case for kcm_sendmsg()") Reported-by:
Cong Wang <xiyou.wangcong@gmail.com> Acked-by:
Tom Herbert <tom@herbertland.com> Signed-off-by:
-
WANG Cong authored
[ Upstream commit 98e3862c ] Dmitry reported a kernel warning: WARNING: CPU: 3 PID: 2936 at net/kcm/kcmsock.c:627 kcm_write_msgs+0x12e3/0x1b90 net/kcm/kcmsock.c:627 CPU: 3 PID: 2936 Comm: a.out Not tainted 4.10.0-rc6+ #209 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011 Call Trace: __dump_stack lib/dump_stack.c:15 [inline] dump_stack+0x2ee/0x3ef lib/dump_stack.c:51 panic+0x1fb/0x412 kernel/panic.c:179 __warn+0x1c4/0x1e0 kernel/panic.c:539 warn_slowpath_null+0x2c/0x40 kernel/panic.c:582 kcm_write_msgs+0x12e3/0x1b90 net/kcm/kcmsock.c:627 kcm_sendmsg+0x163a/0x2200 net/kcm/kcmsock.c:1029 sock_sendmsg_nosec net/socket.c:635 [inline] sock_sendmsg+0xca/0x110 net/socket.c:645 sock_write_iter+0x326/0x600 net/socket.c:848 new_sync_write fs/read_write.c:499 [inline] __vfs_write+0x483/0x740 fs/read_write.c:512 vfs_write+0x187/0x530 fs/read_write.c:560 SYSC_write fs/read_write.c:607 [inline] SyS_write+0xfb/0x230 fs/read_write.c:599 entry_SYSCALL_64_fastpath+0x1f/0xc2 when calling syscall(__NR_write, sock2, 0x208aaf27ul, 0x0ul) on a KCM seqpacket socket. It appears that kcm_sendmsg() does not handle len==0 case correctly, which causes an empty skb is allocated and queued. Fix this by skipping the skb allocation for len==0 case. Reported-by:
-
