- 10 Oct, 2012 40 commits
-
-
Thomas Graf authored
[ Upstream commit 4c3a5bda ] SCTP charges wmem_alloc via sctp_set_owner_w() in sctp_sendmsg() and via skb_set_owner_w() in sctp_packet_transmit(). If a sender runs out of sndbuf it will sleep in sctp_wait_for_sndbuf() and expects to be waken up by __sctp_write_space(). Buffer space charged via sctp_set_owner_w() is released in sctp_wfree() which calls __sctp_write_space() directly. Buffer space charged via skb_set_owner_w() is released via sock_wfree() which calls sk->sk_write_space() _if_ SOCK_USE_WRITE_QUEUE is not set. sctp_endpoint_init() sets SOCK_USE_WRITE_QUEUE on all sockets. Therefore if sctp_packet_transmit() manages to queue up more than sndbuf bytes, sctp_wait_for_sndbuf() will never be woken up again unless it is interrupted by a signal. This could be fixed by clearing the SOCK_USE_WRITE_QUEUE flag but ... Charging for the data twice does not make sense in the first place, it leads to overcharging sndbuf by a factor 2. Therefore this patch only charges a single byte in wmem_alloc when transmitting an SCTP packet to ensure that the socket stays alive until the packet has been released. This means that control chunks are no longer accounted for in wmem_alloc which I believe is not a problem as skb->truesize will typically lead to overcharging anyway and thus compensates for any control overhead. Signed-off-by:
Thomas Graf <tgraf@suug.ch> CC: Vlad Yasevich <vyasevic@redhat.com> CC: Neil Horman <nhorman@tuxdriver.com> CC: David Miller <davem@davemloft.net> Acked-by:
Vlad Yasevich <vyasevich@gmail.com> Signed-off-by:
David S. Miller <davem@davemloft.net> Signed-off-by:
Ben Hutchings <ben@decadent.org.uk>
-
Michal Kubeček authored
[ Upstream commit 15c04175 ] If recv() syscall is called for a TCP socket so that - IOAT DMA is used - MSG_WAITALL flag is used - requested length is bigger than sk_rcvbuf - enough data has already arrived to bring rcv_wnd to zero then when tcp_recvmsg() gets to calling sk_wait_data(), receive window can be still zero while sk_async_wait_queue exhausts enough space to keep it zero. As this queue isn't cleaned until the tcp_service_net_dma() call, sk_wait_data() cannot receive any data and blocks forever. If zero receive window and non-empty sk_async_wait_queue is detected before calling sk_wait_data(), process the queue first. Signed-off-by:
Michal Kubecek <mkubecek@suse.cz> Signed-off-by:
-
Gao feng authored
[ Upstream commit 6825a26c ] as we hold dst_entry before we call __ip6_del_rt, so we should alse call dst_release not only return -ENOENT when the rt6_info is ip6_null_entry. and we already hold the dst entry, so I think it's safe to call dst_release out of the write-read lock. Signed-off-by:
Gao feng <gaofeng@cn.fujitsu.com> Signed-off-by:
-
Antonio Quartulli authored
[ Upstream commit 5316cf9a ] skb_reset_mac_len() relies on the value of the skb->network_header pointer, therefore we must wait for such pointer to be recalculated before computing the new mac_len value. Signed-off-by:
Antonio Quartulli <ordex@autistici.org> Signed-off-by:
-
Lennart Sorensen authored
[ Upstream commit 2120c52d ] I discovered I couldn't get sierra_net to work on a powerpc. Turns out the firmware attribute check assumes the system is little endian and hence fails because the attributes is a 16 bit value. Signed-off-by:
Len Sorensen <lsorense@csclub.uwaterloo.ca> Signed-off-by:
-
Paolo Valente authored
[ Upstream commit 71261956 ] If the old timestamps of a class, say cl, are stale when the class becomes active, then QFQ may assign to cl a much higher start time than the maximum value allowed. This may happen when QFQ assigns to the start time of cl the finish time of a group whose classes are characterized by a higher value of the ratio max_class_pkt/weight_of_the_class with respect to that of cl. Inserting a class with a too high start time into the bucket list corrupts the data structure and may eventually lead to crashes. This patch limits the maximum start time assigned to a class. Signed-off-by:
Paolo Valente <paolo.valente@unimore.it> Signed-off-by:
-
Eric Dumazet authored
[ Upstream commit bdfc87f7 ] Its possible to setup a bad cbq configuration leading to an infinite loop in cbq_classify() DEV_OUT=eth0 ICMP="match ip protocol 1 0xff" U32="protocol ip u32" DST="match ip dst" tc qdisc add dev $DEV_OUT root handle 1: cbq avpkt 1000 \ bandwidth 100mbit tc class add dev $DEV_OUT parent 1: classid 1:1 cbq \ rate 512kbit allot 1500 prio 5 bounded isolated tc filter add dev $DEV_OUT parent 1: prio 3 $U32 \ $ICMP $DST 192.168.3.234 flowid 1: Reported-by:
Denys Fedoryschenko <denys@visp.net.lb> Tested-by:
Eric Dumazet <edumazet@google.com> Signed-off-by:
-
Nikolay Aleksandrov authored
[ Upstream commit e4d1aa40 ] Add a check if pdev->bus->self == NULL (root bus). When attaching a netxen NIC to a VM it can be on the root bus and the guest would crash in netxen_mask_aer_correctable() because of a NULL pointer dereference if CONFIG_PCIEAER is present. Signed-off-by:
Nikolay Aleksandrov <nikolay@redhat.com> Signed-off-by:
-
Florian Fainelli authored
[ Upstream commit 0b836ddd ] Commit 36a12119 (netprio_cgroup.h: dont include module.h from other includes) made the following build error on ixp4xx_hss pop up: CC [M] drivers/net/wan/ixp4xx_hss.o drivers/net/wan/ixp4xx_hss.c:1412:20: error: expected ';', ',' or ')' before string constant drivers/net/wan/ixp4xx_hss.c:1413:25: error: expected ';', ',' or ')' before string constant drivers/net/wan/ixp4xx_hss.c:1414:21: error: expected ';', ',' or ')' before string constant drivers/net/wan/ixp4xx_hss.c:1415:19: error: expected ';', ',' or ')' before string constant make[8]: *** [drivers/net/wan/ixp4xx_hss.o] Error 1 This was previously hidden because ixp4xx_hss includes linux/hdlc.h which includes linux/netdevice.h which includes linux/netprio_cgroup.h which used to include linux/module.h. The real issue was actually present since the initial commit that added this driver since it uses macros from linux/module.h without including this file. Signed-off-by:
Florian Fainelli <florian@openwrt.org> Signed-off-by:
-
htbegin authored
[ Upstream commit ffb5ba90 ] chan->count is used by rx channel. If the desc count is not updated by the clean up loop in cpdma_chan_stop, the value written to the rxfree register in cpdma_chan_start will be incorrect. Signed-off-by:
Tao Hou <hotforest@gmail.com> Signed-off-by:
-
Mathias Krause authored
[ Upstream commit ecd79187 ] The current code fails to ensure that the netlink message actually contains as many bytes as the header indicates. If a user creates a new state or updates an existing one but does not supply the bytes for the whole ESN replay window, the kernel copies random heap bytes into the replay bitmap, the ones happen to follow the XFRMA_REPLAY_ESN_VAL netlink attribute. This leads to following issues: 1. The replay window has random bits set confusing the replay handling code later on. 2. A malicious user could use this flaw to leak up to ~3.5kB of heap memory when she has access to the XFRM netlink interface (requires CAP_NET_ADMIN). Known users of the ESN replay window are strongSwan and Steffen's iproute2 patch (<http://patchwork.ozlabs.org/patch/85962/>). The latter uses the interface with a bitmap supplied while the former does not. strongSwan is therefore prone to run into issue 1. To fix both issues without breaking existing userland allow using the XFRMA_REPLAY_ESN_VAL netlink attribute with either an empty bitmap or a fully specified one. For the former case we initialize the in-kernel bitmap with zero, for the latter we copy the user supplied bitmap. For state updates the full bitmap must be supplied. To prevent overflows in the bitmap length calculation the maximum size of bmp_len is limited to 128 by this patch -- resulting in a maximum replay window of 4096 packets. This should be sufficient for all real life scenarios (RFC 4303 recommends a default replay window size of 64). Cc: Steffen Klassert <steffen.klassert@secunet.com> Cc: Martin Willi <martin@revosec.ch> Cc: Ben Hutchings <bhutchings@solarflare.com> Signed-off-by:
Mathias Krause <minipli@googlemail.com> Signed-off-by:
-
Mathias Krause authored
[ Upstream commit e3ac104d ] The ESN replay window was already fully initialized in xfrm_alloc_replay_state_esn(). No need to copy it again. Cc: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by:
Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by:
-
Mathias Krause authored
[ Upstream commit 1f86840f ] The memory used for the template copy is a local stack variable. As struct xfrm_user_tmpl contains multiple holes added by the compiler for alignment, not initializing the memory will lead to leaking stack bytes to userland. Add an explicit memset(0) to avoid the info leak. Initial version of the patch by Brad Spengler. Cc: Brad Spengler <spender@grsecurity.net> Signed-off-by:
-
Mathias Krause authored
[ Upstream commit 7b789836 ] The memory reserved to dump the xfrm policy includes multiple padding bytes added by the compiler for alignment (padding bytes in struct xfrm_selector and struct xfrm_userpolicy_info). Add an explicit memset(0) before filling the buffer to avoid the heap info leak. Signed-off-by:
-
Mathias Krause authored
[ Upstream commit f778a636 ] The memory reserved to dump the xfrm state includes the padding bytes of struct xfrm_usersa_info added by the compiler for alignment (7 for amd64, 3 for i386). Add an explicit memset(0) before filling the buffer to avoid the info leak. Signed-off-by:
-
Mathias Krause authored
[ Upstream commit 4c87308b ] copy_to_user_auth() fails to initialize the remainder of alg_name and therefore discloses up to 54 bytes of heap memory via netlink to userland. Use strncpy() instead of strcpy() to fill the trailing bytes of alg_name with null bytes. Signed-off-by:
-
Li RongQing authored
[ Upstream commit 433a1954 ] if xfrm_policy_get_afinfo returns 0, it has already released the read lock, xfrm_policy_put_afinfo should not be called again. Signed-off-by:
Li RongQing <roy.qing.li@gmail.com> Signed-off-by:
-
Mathias Krause authored
[ Upstream commit c2546372 ] When dump_one_policy() returns an error, e.g. because of a too small buffer to dump the whole xfrm policy, xfrm_policy_netlink() returns NULL instead of an error pointer. But its caller expects an error pointer and therefore continues to operate on a NULL skbuff. Signed-off-by:
-
Mathias Krause authored
[ Upstream commit 864745d2 ] When dump_one_state() returns an error, e.g. because of a too small buffer to dump the whole xfrm state, xfrm_state_netlink() returns NULL instead of an error pointer. But its callers expect an error pointer and therefore continue to operate on a NULL skbuff. This could lead to a privilege escalation (execution of user code in kernel context) if the attacker has CAP_NET_ADMIN and is able to map address 0. Signed-off-by:
-
Steffen Klassert authored
[ Upstream commit 3b59df46 ] ESN for esp is defined in RFC 4303. This RFC assumes that the sequence number counters are always up to date. However, this is not true if an async crypto algorithm is employed. If the sequence number counters are not up to date on sequence number check, we may incorrectly update the upper 32 bit of the sequence number. This leads to a DOS. We workaround this by comparing the upper sequence number, (used for authentication) with the upper sequence number computed after the async processing. We drop the packet if these numbers are different. To do this, we introduce a recheck function that does this check in the ESN case. Signed-off-by:
Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by:
-
Michal Schmidt authored
[ Upstream commit e488921f ] Commit d6cb3e41 "bnx2x: fix checksum validation" caused a performance regression for IPv6. Rx checksum offload does not work. IPv6 packets are passed to the stack with CHECKSUM_NONE. The hardware obviously cannot perform IP checksum validation for IPv6, because there is no checksum in the IPv6 header. This should not prevent us from setting CHECKSUM_UNNECESSARY. Tested on BCM57711. Signed-off-by:
Michal Schmidt <mschmidt@redhat.com> Acked-by:
Eilon Greenstein <eilong@broadcom.com> Signed-off-by:
-
Bjorn Helgaas authored
commit dfb117b3 upstream. Check whether we evaluated _ADR successfully. Previously we ignored failure, so we would have used garbage data from the stack as the device and function number. We return AE_OK so that we ignore only this slot and continue looking for other slots. Found by Coverity (CID 113981). Signed-off-by:
Bjorn Helgaas <bhelgaas@google.com> Signed-off-by:
-
Ratan Nalumasu authored
commit 4fe9f8e2 upstream. When a device is unplugged, wait for all processes that have opened the device to close before deallocating the device. Signed-off-by:
Ratan Nalumasu <ratan@google.com> Signed-off-by:
Jiri Kosina <jkosina@suse.cz> Signed-off-by:
-
Alexey Khoroshilov authored
commit bcb4a75b upstream. Several improvements in error handling: - do not report success if alloc_chrdev_region() failed - check for error code of cdev_add() - use unregister_chrdev_region() instead of unregister_chrdev() if class_create() failed Found by Linux Driver Verification project (linuxtesting.org). Signed-off-by:
Alexey Khoroshilov <khoroshilov@ispras.ru> Signed-off-by:
-
Matthieu CASTET authored
commit 4c7b417e upstream. If we don't read fast enough hidraw device, hidraw_report_event will cycle and we will leak list->buffer. Also list->buffer are not free on release. After this patch, kmemleak report nothing. Signed-off-by:
Matthieu CASTET <matthieu.castet@parrot.com> Signed-off-by:
-
Jiri Kosina authored
commit d6d7c873 upstream. Commit b6787242 ("HID: hidraw: add proper error handling to raw event reporting") forgot to update the static inline version of hidraw_report_event() for the case when CONFIG_HIDRAW is unset. Fix that up. Reported-by:
Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by:
-
Jiri Kosina authored
commit b6787242 upstream. If kmemdup() in hidraw_report_event() fails, we are not propagating this fact properly. Let hidraw_report_event() and hid_report_raw_event() return an error value to the caller. Reported-by:
Oliver Neukum <oneukum@suse.de> Signed-off-by:
-
Avi Kivity authored
commit cb09cad4 upstream. Probably a leftover from the early days of self-patching, p6nops are marked __initconst_or_module, which causes them to be discarded in a non-modular kernel. If something later triggers patching, it will overwrite kernel code with garbage. Reported-by:
Tomas Racek <tracek@redhat.com> Signed-off-by:
Avi Kivity <avi@redhat.com> Cc: Michael Tokarev <mjt@tls.msk.ru> Cc: Borislav Petkov <borislav.petkov@amd.com> Cc: Marcelo Tosatti <mtosatti@redhat.com> Cc: qemu-devel@nongnu.org Cc: Anthony Liguori <anthony@codemonkey.ws> Cc: H. Peter Anvin <hpa@linux.intel.com> Cc: Alan Cox <alan@lxorguk.ukuu.org.uk> Cc: Alan Cox <alan@linux.intel.com> Link: http://lkml.kernel.org/r/5034AE84.90708@redhat.comSigned-off-by:
Ingo Molnar <mingo@kernel.org> Signed-off-by:
-
Alex Deucher authored
commit 2f1f4d9b upstream. This reverts commit 985f61f7. This commit fixed certain cases, but ended up regressing others due to limitations in the current KMS API. A proper fix is too invasive for 3.6. Push it back to 3.7. Reported-by:
Andres Freund <andres@anarazel.de> Signed-off-by:
Alex Deucher <alexander.deucher@amd.com> [bwh: Backported to 3.2: drop the DCE6 case] Signed-off-by:
-
Srivatsa S. Bhat authored
commit d35be8ba upstream. In the event of CPU hotplug, the kernel modifies the cpusets' cpus_allowed masks as and when necessary to ensure that the tasks belonging to the cpusets have some place (online CPUs) to run on. And regular CPU hotplug is destructive in the sense that the kernel doesn't remember the original cpuset configurations set by the user, across hotplug operations. However, suspend/resume (which uses CPU hotplug) is a special case in which the kernel has the responsibility to restore the system (during resume), to exactly the same state it was in before suspend. In order to achieve that, do the following: 1. Don't modify cpusets during suspend/resume. At all. In particular, don't move the tasks from one cpuset to another, and don't modify any cpuset's cpus_allowed mask. So, simply ignore cpusets during the CPU hotplug operations that are carried out in the suspend/resume path. 2. However, cpusets and sched domains are related. We just want to avoid altering cpusets alone. So, to keep the sched domains updated, build a single sched domain (containing all active cpus) during each of the CPU hotplug operations carried out in s/r path, effectively ignoring the cpusets' cpus_allowed masks. (Since userspace is frozen while doing all this, it will go unnoticed.) 3. During the last CPU online operation during resume, build the sched domains by looking up the (unaltered) cpusets' cpus_allowed masks. That will bring back the system to the same original state as it was in before suspend. Ultimately, this will not only solve the cpuset problem related to suspend resume (ie., restores the cpusets to exactly what it was before suspend, by not touching it at all) but also speeds up suspend/resume because we avoid running cpuset update code for every CPU being offlined/onlined. Signed-off-by:
Srivatsa S. Bhat <srivatsa.bhat@linux.vnet.ibm.com> Signed-off-by:
Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Thomas Gleixner <tglx@linutronix.de> Link: http://lkml.kernel.org/r/20120524141611.3692.20155.stgit@srivatsabhat.in.ibm.com [Preeti U Murthy: Please apply this patch to the stable tree 3.0.y] Signed-off-by:
Preeti U Murthy <preeti@linux.vnet.ibm.com> Signed-off-by:
-
Sebastian Andrzej Siewior authored
commit 1b68a4ca upstream. If USB2 host controller probes fine but USB3 does not then we don't remove the USB controller properly and lock up the system while the HUB code will try to enumerate the USB2 controller and access memory which is no longer available in case the dummy_hcd was compiled as a module. This is a problem since 448b6eb1 ("USB: Make sure to fetch the BOS desc for roothubs.) if used in USB3 mode because dummy does not provide this descriptor and explodes later. Signed-off-by:
Sebastian Andrzej Siewior <sebastian@breakpoint.cc> Signed-off-by:
Felipe Balbi <balbi@ti.com> Signed-off-by:
-
Miklos Szeredi authored
commit 8110e16d upstream. IBM reported a deadlock in select_parent(). This was found to be caused by taking rename_lock when already locked when restarting the tree traversal. There are two cases when the traversal needs to be restarted: 1) concurrent d_move(); this can only happen when not already locked, since taking rename_lock protects against concurrent d_move(). 2) racing with final d_put() on child just at the moment of ascending to parent; rename_lock doesn't protect against this rare race, so it can happen when already locked. Because of case 2, we need to be able to handle restarting the traversal when rename_lock is already held. This patch fixes all three callers of try_to_ascend(). IBM reported that the deadlock is gone with this patch. [ I rewrote the patch to be smaller and just do the "goto again" if the lock was already held, but credit goes to Miklos for the real work. - Linus ] Signed-off-by:
Miklos Szeredi <mszeredi@suse.cz> Cc: Al Viro <viro@ZenIV.linux.org.uk> Signed-off-by:
Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by:
-
Seth Forshee authored
commit 32ab31e0 upstream. The ACPI tables in the Macbook Air 5,1 define a single IOAPIC with id 2, but the only remapping unit described in the DMAR table matches id 0. Interrupt remapping fails as a result, and the kernel panics with the message "timer doesn't work through Interrupt-remapped IO-APIC." To fix this, check each IOAPIC for a corresponding IOMMU. If an IOMMU is not found, do not allow IRQ remapping to be enabled. v2: Move check to parse_ioapics_under_ir(), raise log level to KERN_ERR, and add FW_BUG to the log message v3: Skip check if IOMMU doesn't support interrupt remapping and remove existing check that the IOMMU count equals the IOAPIC count Acked-by:
Suresh Siddha <suresh.b.siddha@intel.com> Signed-off-by:
Seth Forshee <seth.forshee@canonical.com> Acked-by:
Yinghai Lu <yinghai@kernel.org> Signed-off-by:
Joerg Roedel <joerg.roedel@amd.com> [bwh: Backported to 3.2: adjust filename] Signed-off-by:
-
Darren Hart authored
commit 2588aba0 upstream. pch_uart_interrupt() takes priv->port.lock which leads to two recursive spinlock calls if low_latency==1 or CONFIG_PREEMPT_RT_FULL=y (one otherwise): pch_uart_interrupt spin_lock_irqsave(priv->port.lock, flags) case PCH_UART_IID_RDR_TO (data ready) handle_rx_to push_rx tty_port_tty_get spin_lock_irqsave(&port->lock, flags) <--- already hold this lock ... tty_flip_buffer_push ... flush_to_ldisc spin_lock_irqsave(&tty->buf.lock) spin_lock_irqsave(&tty->buf.lock) disc->ops->receive_buf(tty, char_buf) n_tty_receive_buf tty->ops->flush_chars() uart_flush_chars uart_start spin_lock_irqsave(&port->lock) <--- already hold this lock Avoid this by using a dedicated lock to protect the eg20t_port structure and IO access to its membase. This is more consistent with the 8250 driver. Ensure priv->lock is always take prior to priv->port.lock when taken at the same time. V2: Remove inadvertent whitespace change. V3: Account for oops_in_progress for the private lock in pch_console_write(). Note: Like the 8250 driver, if a printk is introduced anywhere inside the pch_console_write() critical section, the kernel will hang on a recursive spinlock on the private lock. The oops case is handled by using a trylock in the oops_in_progress case. Signed-off-by:
Darren Hart <dvhart@linux.intel.com> CC: Tomoya MORINAGA <tomoya.rohm@gmail.com> CC: Feng Tang <feng.tang@intel.com> CC: Alexander Stein <alexander.stein@systec-electronic.com> Acked-by:
Alan Cox <alan@linux.intel.com> Signed-off-by:
Greg Kroah-Hartman <gregkh@linuxfoundation.org> [bwh: Backported to 3.2: - Adjust context - Drop changes to pch_console_write()] Signed-off-by:
-
Vinicius Costa Gomes authored
commit d8343f12 upstream. In the case that the link is already in the connected state and a Pairing request arrives from the mgmt interface, hci_conn_security() would be called but it was not considering LE links. Reported-by:
João Paulo Rechi Vita <jprvita@openbossa.org> Signed-off-by:
Vinicius Costa Gomes <vinicius.gomes@openbossa.org> Signed-off-by:
Gustavo Padovan <gustavo.padovan@collabora.co.uk> Signed-off-by:
-
Vinicius Costa Gomes authored
commit cc110922 upstream. To make it clear that it may be called from contexts that may not have any knowledge of L2CAP, we change the connection parameter, to receive a hci_conn. This also makes it clear that it is checking the security of the link. Signed-off-by:
-
Al Cooper authored
commit 4188bba0 upstream. The driver should not try to switch to 1.8V when the SD 3.0 host controller does not have any UHS capabilities bits set (SDR50, DDR50 or SDR104). See page 72 of "SD Specifications Part A2 SD Host Controller Simplified Specification Version 3.00" under "1.8V Signaling Enable". Instead of setting SDR12 and SDR25 in the host capabilities data structure for all V3.0 host controllers, only set them if SDR104, SDR50 or DDR50 is set in the host capabilities register. This will prevent the switch to 1.8V later. Signed-off-by:
Al Cooper <acooper@gmail.com> Acked-by:
Arindam Nath <arindam.nath@amd.com> Acked-by:
Philip Rakity <prakity@marvell.com> Acked-by:
Girish K S <girish.shivananjappa@linaro.org> Signed-off-by:
Chris Ball <cjb@laptop.org> Signed-off-by:
-
Daniel J Blueman authored
commit c531077f upstream. When using my Seagate FreeAgent GoFlex eSATAp external disk enclosure, interface errors are always seen until 1.5Gbps is negotiated [1]. This occurs using any disk in the enclosure, and when the disk is connected directly with a generic passive eSATAp cable, we see stable 3Gbps operation as expected. Blacklist 3Gbps mode to avoid dataloss and the ~30s delay bus reset and renegotiation incurs. Signed-off-by:
Daniel J Blueman <daniel@quora.org> Signed-off-by:
Jeff Garzik <jgarzik@redhat.com> Signed-off-by:
-
Weiping Pan authored
commit 06b6a1cf upstream. Jay Fenlason (fenlason@redhat.com) found a bug, that recvfrom() on an RDS socket can return the contents of random kernel memory to userspace if it was called with a address length larger than sizeof(struct sockaddr_in). rds_recvmsg() also fails to set the addr_len paramater properly before returning, but that's just a bug. There are also a number of cases wher recvfrom() can return an entirely bogus address. Anything in rds_recvmsg() that returns a non-negative value but does not go through the "sin = (struct sockaddr_in *)msg->msg_name;" code path at the end of the while(1) loop will return up to 128 bytes of kernel memory to userspace. And I write two test programs to reproduce this bug, you will see that in rds_server, fromAddr will be overwritten and the following sock_fd will be destroyed. Yes, it is the programmer's fault to set msg_namelen incorrectly, but it is better to make the kernel copy the real length of address to user space in such case. How to run the test programs ? I test them on 32bit x86 system, 3.5.0-rc7. 1 compile gcc -o rds_client rds_client.c gcc -o rds_server rds_server.c 2 run ./rds_server on one console 3 run ./rds_client on another console 4 you will see something like: server is waiting to receive data... old socket fd=3 server received data from client:data from client msg.msg_namelen=32 new socket fd=-1067277685 sendmsg() : Bad file descriptor /***************** rds_client.c ********************/ int main(void) { int sock_fd; struct sockaddr_in serverAddr; struct sockaddr_in toAddr; char recvBuffer[128] = "data from client"; struct msghdr msg; struct iovec iov; sock_fd = socket(AF_RDS, SOCK_SEQPACKET, 0); if (sock_fd < 0) { perror("create socket error\n"); exit(1); } memset(&serverAddr, 0, sizeof(serverAddr)); serverAddr.sin_family = AF_INET; serverAddr.sin_addr.s_addr = inet_addr("127.0.0.1"); serverAddr.sin_port = htons(4001); if (bind(sock_fd, (struct sockaddr*)&serverAddr, sizeof(serverAddr)) < 0) { perror("bind() error\n"); close(sock_fd); exit(1); } memset(&toAddr, 0, sizeof(toAddr)); toAddr.sin_family = AF_INET; toAddr.sin_addr.s_addr = inet_addr("127.0.0.1"); toAddr.sin_port = htons(4000); msg.msg_name = &toAddr; msg.msg_namelen = sizeof(toAddr); msg.msg_iov = &iov; msg.msg_iovlen = 1; msg.msg_iov->iov_base = recvBuffer; msg.msg_iov->iov_len = strlen(recvBuffer) + 1; msg.msg_control = 0; msg.msg_controllen = 0; msg.msg_flags = 0; if (sendmsg(sock_fd, &msg, 0) == -1) { perror("sendto() error\n"); close(sock_fd); exit(1); } printf("client send data:%s\n", recvBuffer); memset(recvBuffer, '\0', 128); msg.msg_name = &toAddr; msg.msg_namelen = sizeof(toAddr); msg.msg_iov = &iov; msg.msg_iovlen = 1; msg.msg_iov->iov_base = recvBuffer; msg.msg_iov->iov_len = 128; msg.msg_control = 0; msg.msg_controllen = 0; msg.msg_flags = 0; if (recvmsg(sock_fd, &msg, 0) == -1) { perror("recvmsg() error\n"); close(sock_fd); exit(1); } printf("receive data from server:%s\n", recvBuffer); close(sock_fd); return 0; } /***************** rds_server.c ********************/ int main(void) { struct sockaddr_in fromAddr; int sock_fd; struct sockaddr_in serverAddr; unsigned int addrLen; char recvBuffer[128]; struct msghdr msg; struct iovec iov; sock_fd = socket(AF_RDS, SOCK_SEQPACKET, 0); if(sock_fd < 0) { perror("create socket error\n"); exit(0); } memset(&serverAddr, 0, sizeof(serverAddr)); serverAddr.sin_family = AF_INET; serverAddr.sin_addr.s_addr = inet_addr("127.0.0.1"); serverAddr.sin_port = htons(4000); if (bind(sock_fd, (struct sockaddr*)&serverAddr, sizeof(serverAddr)) < 0) { perror("bind error\n"); close(sock_fd); exit(1); } printf("server is waiting to receive data...\n"); msg.msg_name = &fromAddr; /* * I add 16 to sizeof(fromAddr), ie 32, * and pay attention to the definition of fromAddr, * recvmsg() will overwrite sock_fd, * since kernel will copy 32 bytes to userspace. * * If you just use sizeof(fromAddr), it works fine. * */ msg.msg_namelen = sizeof(fromAddr) + 16; /* msg.msg_namelen = sizeof(fromAddr); */ msg.msg_iov = &iov; msg.msg_iovlen = 1; msg.msg_iov->iov_base = recvBuffer; msg.msg_iov->iov_len = 128; msg.msg_control = 0; msg.msg_controllen = 0; msg.msg_flags = 0; while (1) { printf("old socket fd=%d\n", sock_fd); if (recvmsg(sock_fd, &msg, 0) == -1) { perror("recvmsg() error\n"); close(sock_fd); exit(1); } printf("server received data from client:%s\n", recvBuffer); printf("msg.msg_namelen=%d\n", msg.msg_namelen); printf("new socket fd=%d\n", sock_fd); strcat(recvBuffer, "--data from server"); if (sendmsg(sock_fd, &msg, 0) == -1) { perror("sendmsg()\n"); close(sock_fd); exit(1); } } close(sock_fd); return 0; } Signed-off-by:
Weiping Pan <wpan@redhat.com> Signed-off-by:
-
Li Zhong authored
[Fixed upstream by commits 2955b47d and a4683487 from Dan Williams, but they are much more intrusive than this tiny fix, according to Andrew - gregkh] This patch tries to fix a dead loop in async_synchronize_full(), which could be seen when preemption is disabled on a single cpu machine. void async_synchronize_full(void) { do { async_synchronize_cookie(next_cookie); } while (!list_empty(&async_running) || ! list_empty(&async_pending)); } async_synchronize_cookie() calls async_synchronize_cookie_domain() with &async_running as the default domain to synchronize. However, there might be some works in the async_pending list from other domains. On a single cpu system, without preemption, there is no chance for the other works to finish, so async_synchronize_full() enters a dead loop. It seems async_synchronize_full() wants to synchronize all entries in all running lists(domains), so maybe we could just check the entry_count to know whether all works are finished. Currently, async_synchronize_cookie_domain() expects a non-NULL running list ( if NULL, there would be NULL pointer dereference ), so maybe a NULL pointer could be used as an indication for the functions to synchronize all works in all domains. Reported-by:
Paul E. McKenney <paulmck@linux.vnet.ibm.com> Signed-off-by:
Li Zhong <zhong@linux.vnet.ibm.com> Tested-by:
Christian Kujau <lists@nerdbynature.de> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Dan Williams <dan.j.williams@gmail.com> Cc: Christian Kujau <lists@nerdbynature.de> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Cong Wang <xiyou.wangcong@gmail.com> Signed-off-by:
-
