aboutsummaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAge
* Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6Linus Torvalds2008-02-24
|\ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | * git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6: (37 commits) [NETFILTER]: fix ebtable targets return [IP_TUNNEL]: Don't limit the number of tunnels with generic name explicitly. [NET]: Restore sanity wrt. print_mac(). [NEIGH]: Fix race between neighbor lookup and table's hash_rnd update. [RTNL]: Validate hardware and broadcast address attribute for RTM_NEWLINK tg3: ethtool phys_id default [BNX2]: Update version to 1.7.4. [BNX2]: Disable parallel detect on an HP blade. [BNX2]: More 5706S link down workaround. ssb: Fix support for PCI devices behind a SSB->PCI bridge zd1211rw: fix sparse warnings rtl818x: fix sparse warnings ssb: Fix pcicore cardbus mode ssb: Make the GPIO API reentrancy safe ssb: Fix the GPIO API ssb: Fix watchdog access for devices without a chipcommon ssb: Fix serial console on new bcm47xx devices ath5k: Fix build warnings on some 64-bit platforms. WDEV, ath5k, don't return int from bool function WDEV: ath5k, fix lock imbalance ...
| * [NETFILTER]: fix ebtable targets returnJoonwoo Park2008-02-23
| | | | | | | | | | | | | | | | The function ebt_do_table doesn't take NF_DROP as a verdict from the targets. Signed-off-by: Joonwoo Park <joonwpark81@gmail.com> Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
| * [IP_TUNNEL]: Don't limit the number of tunnels with generic name explicitly.Pavel Emelyanov2008-02-23
| | | | | | | | | | | | | | | | | | | | | | | | | | | | Use the added dev_alloc_name() call to create tunnel device name, rather than iterate in a hand-made loop with an artificial limit. Thanks Patrick for noticing this. [ The way this works is, when the device is actually registered, the generic code noticed the '%' in the name and invokes dev_alloc_name() to fully resolve the name. -DaveM ] Signed-off-by: Pavel Emelyanov <xemul@openvz.org> Signed-off-by: David S. Miller <davem@davemloft.net>
| * [NET]: Restore sanity wrt. print_mac().David S. Miller2008-02-23
| | | | | | | | | | | | | | | | | | | | | | | | | | | | MAC_FMT had only one user and we tried to get rid of that, but this created more problems than it solved. As a result, this reverts three commits: 235365f3aaaa10b7056293877c0ead50425f25c7 ("net/8021q/vlan_dev.c: Use print_mac."), fea5fa875eb235dc186b1f5184eb36abc63e26cc ("[NET]: Remove MAC_FMT"), and 8f789c48448aed74fe1c07af76de8f04adacec7d ("[NET]: Elminate spurious print_mac() calls.") Signed-off-by: David S. Miller <davem@davemloft.net>
| * [NEIGH]: Fix race between neighbor lookup and table's hash_rnd update.Pavel Emelyanov2008-02-23
| | | | | | | | | | | | | | | | | | | | | | | | | | The neigh_hash_grow() may update the tbl->hash_rnd value, which is used in all tbl->hash callbacks to calculate the hashval. Two lookup routines may race with this, since they call the ->hash callback without the tbl->lock held. Since the hash_rnd is changed with this lock write-locked moving the calls to ->hash under this lock read-locked closes this gap. Signed-off-by: Pavel Emelyanov <xemul@openvz.org> Signed-off-by: David S. Miller <davem@davemloft.net>
| * [RTNL]: Validate hardware and broadcast address attribute for RTM_NEWLINKThomas Graf2008-02-23
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | RTM_NEWLINK allows for already existing links to be modified. For this purpose do_setlink() is called which expects address attributes with a payload length of at least dev->addr_len. This patch adds the necessary validation for the RTM_NEWLINK case. The address length for links to be created is not checked for now as the actual attribute length is used when copying the address to the netdevice structure. It might make sense to report an error if less than addr_len bytes are provided but enforcing this might break drivers trying to be smart with not transmitting all zero addresses. Signed-off-by: Thomas Graf <tgraf@suug.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
| * tg3: ethtool phys_id defaultStephen Hemminger2008-02-23
| | | | | | | | | | | | | | | | | | | | | | | | | | When asked to blink LEDs the tg3 driver behaves when using: ethtool -p ethX The default value for data is zero, and other drivers interpret this as blink forever (or at least a really long time). The tg3 driver interprets this as blink once. All drivers should have the same behaviour. Signed-off-by: Stephen Hemminger <shemminger@vyatta.com> Acked-by: Michael Chan <mchan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
| * [BNX2]: Update version to 1.7.4.Michael Chan2008-02-23
| | | | | | | | | | Signed-off-by: Michael Chan <mchan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
| * [BNX2]: Disable parallel detect on an HP blade.Michael Chan2008-02-23
| | | | | | | | | | | | | | | | | | Because of some board issues, we need to disable parallel detect on an HP blade. Without this patch, the link state can become stuck when it goes into parallel detect mode. Signed-off-by: Michael Chan <mchan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
| * [BNX2]: More 5706S link down workaround.Michael Chan2008-02-23
| | | | | | | | | | | | | | | | | | The previous patches to workaround the 5706S on an HP blade were not sufficient. The link state still does not change properly in some cases. This patch adds polling to make it completely reliable. Signed-off-by: Michael Chan <mchan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
| * ssb: Fix support for PCI devices behind a SSB->PCI bridgeMichael Buesch2008-02-20
| | | | | | | | | | | | | | | | We must pin all resources and make sure the PCI subsystem won't relocate us, as the addresses are hardwired into hardware. Signed-off-by: Michael Buesch <mb@bu3sch.de> Signed-off-by: John W. Linville <linville@tuxdriver.com>
| * zd1211rw: fix sparse warningsJohannes Berg2008-02-20
| | | | | | | | | | | | | | This silences sparse when run on zd1211rw. Signed-off-by: Johannes Berg <johannes@sipsolutions.net> Signed-off-by: John W. Linville <linville@tuxdriver.com>
| * rtl818x: fix sparse warningsJohannes Berg2008-02-20
| | | | | | | | | | | | | | | | This silences a few sparse warnings. There are two more where I can't follow the code. Signed-off-by: Johannes Berg <johannes@sipsolutions.net> Signed-off-by: John W. Linville <linville@tuxdriver.com>
| * ssb: Fix pcicore cardbus modeMichael Buesch2008-02-20
| | | | | | | | | | | | | | | | This fixes the pcicore driver to not die a horrible crash death when inserting a cardbus card. Signed-off-by: Michael Buesch <mb@bu3sch.de> Signed-off-by: John W. Linville <linville@tuxdriver.com>
| * ssb: Make the GPIO API reentrancy safeMichael Buesch2008-02-20
| | | | | | | | | | | | | | This fixes the GPIO API to be reentrancy safe. Signed-off-by: Michael Buesch <mb@bu3sch.de> Signed-off-by: John W. Linville <linville@tuxdriver.com>
| * ssb: Fix the GPIO APIMichael Buesch2008-02-20
| | | | | | | | | | | | | | This fixes the GPIO API to be usable. Signed-off-by: Michael Buesch <mb@bu3sch.de> Signed-off-by: John W. Linville <linville@tuxdriver.com>
| * ssb: Fix watchdog access for devices without a chipcommonMichael Buesch2008-02-20
| | | | | | | | | | | | | | | | This fixes the SSB watchdog access for devices without a chipcommon. These devices have the watchdog on the extif. Signed-off-by: Michael Buesch <mb@bu3sch.de> Signed-off-by: John W. Linville <linville@tuxdriver.com>
| * ssb: Fix serial console on new bcm47xx devicesMichael Buesch2008-02-20
| | | | | | | | | | | | | | | | This fixes the baud settings for new devices like the Linksys WRT350n. Signed-off-by: Michael Buesch <mb@bu3sch.de> Signed-off-by: John W. Linville <linville@tuxdriver.com>
| * ath5k: Fix build warnings on some 64-bit platforms.David Miller2008-02-20
| | | | | | | | | | | | | | | | | | | | | | | | | | 'u64' is not necessarily 'unsigned long long' drivers/net/wireless/ath5k/base.c: In function 'ath5k_beacon_update_timers': drivers/net/wireless/ath5k/base.c:2130: warning: format '%llx' expects type 'long long unsigned int', but argument 4 has type 'u64' drivers/net/wireless/ath5k/base.c:2130: warning: format '%llx' expects type 'long long unsigned int', but argument 5 has type 'u64' drivers/net/wireless/ath5k/base.c: In function 'ath5k_intr': drivers/net/wireless/ath5k/base.c:2391: warning: format '%llx' expects type 'long long unsigned int', but argument 6 has type 'u64' Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: John W. Linville <linville@tuxdriver.com>
| * WDEV, ath5k, don't return int from bool functionJiri Slaby2008-02-20
| | | | | | | | | | | | | | | | | | | | | | sparse sees int -> bool cast as an error: hw.c:3754:10: warning: cast truncates bits from constant value (ffffffea becomes 0) Fix it by converting the rettype to int and check appropriately. Signed-off-by: Jiri Slaby <jirislaby@gmail.com> Cc: Nick Kossifidis <mickflemm@gmail.com> Cc: Luis R. Rodriguez <mcgrof@gmail.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
| * WDEV: ath5k, fix lock imbalanceJiri Slaby2008-02-20
| | | | | | | | | | | | | | | | | | | | | | | | Omitted lock causes sparse warning drivers/net/wireless/ath5k/base.c:1682:1: warning: context imbalance in 'ath5k_tasklet_rx' - different lock contexts for basic block Add the lock to the guilty fail path. Signed-off-by: Jiri Slaby <jirislaby@gmail.com> Acked-by: Nick Kossifidis <mickflemm@gmail.com> Cc: Luis R. Rodriguez <mcgrof@gmail.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
| * p54usb: add USB ID for Linksys WUSB54G ver 2John W. Linville2008-02-20
| | | | | | | | | | | | | | | | | | Based on report from Cavan Carroll <cavancarroll@hotmail.com>: http://bugzilla.kernel.org/show_bug.cgi?id=9863 Cc: Cavan Carroll <cavancarroll@hotmail.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
| * p54usb: add USB ID for Phillips CPWUA054Ivo Couckuyt2008-02-20
| | | | | | | | | | | | | | | | | | Retarget of an old patch against prism54usb in linux-wireless archive: http://marc.info/?l=linux-wireless&m=117449935810254&w=2 Cc: Ivo Couckuyt <iv0co@yahoo.co.uk> Signed-off-by: John W. Linville <linville@tuxdriver.com>
| * [NETNS]: Namespace leak in pneigh_lookup.Denis V. Lunev2008-02-20
| | | | | | | | | | | | | | release_net is missed on the error path in pneigh_lookup. Signed-off-by: Denis V. Lunev <den@openvz.org> Signed-off-by: David S. Miller <davem@davemloft.net>
| * [SCTP]: Pick up an orphaned sctp_sockets_allocated counter.Pavel Emelyanov2008-02-20
| | | | | | | | | | | | | | | | | | | | | | | | This counter is currently write-only. Drawing an analogy with the similar tcp counter, I think that this one should be pointed by the sockets_allocated members of sctp_prot and sctpv6_prot. Signed-off-by: Pavel Emelyanov <xemul@openvz.org> Acked-by: Vlad Yasevich <vladislav.yasevich@hp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
| * veth: fix dev refcount raceDaniel Lezcano2008-02-20
| | | | | | | | | | | | | | | | | | | | | | | | | | | | When deleting the veth driver, veth_close calls netif_carrier_off for the two extremities of the network device. netif_carrier_off on the peer device will fire an event and hold a reference on the peer device. Just after, the peer is unregistered taking the rtnl_lock while the linkwatch_event is scheduled. If __linkwatch_run_queue does not occurs before the unregistering, unregister_netdevice will wait for the dev refcount to reach zero holding the rtnl_lock and linkwatch_event will wait for the rtnl_lock and hold the dev refcount. Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com> Signed-off-by: David S. Miller <davem@davemloft.net>
| * [NIU]: More BMAC alt MAC address fixes.Matheos Worku2008-02-20
| | | | | | | | | | | | | | | | | | | | | | From: Matheos Worku <Matheos.Worku@Sun.COM> 1) niu_enable_alt_mac() needs to be adjusted so that the mask is computed properly for the BMAC case. 2) BMAC has 6 alt MAC addresses available, not 7. Signed-off-by: David S. Miller <davem@davemloft.net>
| * [NETFILTER]: Use __u32 in struct nf_inet_addrPatrick McHardy2008-02-19
| | | | | | | | | | | | | | | | As reported by David Woodhouse <dwmw2@infradead.org>, using u_int32_t in struct nf_inet_addr breaks the busybox build. Fix by using __u32. Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
| * [NETFILTER]: xt_iprange: fix subtraction-based comparisonJan Engelhardt2008-02-19
| | | | | | | | | | | | | | | | | | The host address parts need to be converted to host-endian first before arithmetic makes any sense on them. Signed-off-by: Jan Engelhardt <jengelh@computergmbh.de> Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
| * [NETFILTER]: xt_hashlimit: remove unneeded struct memberJan Engelhardt2008-02-19
| | | | | | | | | | | | | | | | | | | | | | By allocating ->hinfo, we already have the needed indirection to cope with the per-cpu xtables struct match_entry. [Patrick: do this now before the revision 1 struct is used by userspace] Signed-off-by: Jan Engelhardt <jengelh@computergmbh.de> Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
| * [NETFILTER]: Make sure xt_policy.h is unifdef'ed.Robert P. J. Day2008-02-19
| | | | | | | | | | | | | | | | | | Since the header file xt_policy.h tests __KERNEL__, it should be unifdef'ed before exporting to userspace. Signed-off-by: Robert P. J. Day <rpjday@crashcourse.ca> Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
| * [NETFILTER]: Fix incorrect use of skb_make_writableJoonwoo Park2008-02-19
| | | | | | | | | | | | | | | | | | http://bugzilla.kernel.org/show_bug.cgi?id=9920 The function skb_make_writable returns true or false. Signed-off-by: Joonwoo Park <joonwpark81@gmail.com> Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
| * [NETFILTER]: xt_u32: drop the actually unused variable from u32_match_itPavel Emelyanov2008-02-19
| | | | | | | | | | | | | | | | | | | | The int ret variable is used only to trigger the BUG_ON() after the skb_copy_bits() call, so check the call failure directly and drop the variable. Signed-off-by: Pavel Emelyanov <xemul@openvz.org> Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
| * [NETFILTER]: {ip,ip6,nfnetlink}_queue: fix SKB_LINEAR_ASSERT when mangling ↵Patrick McHardy2008-02-19
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | packet data As reported by Tomas Simonaitis <tomas.simonaitis@gmail.com>, inserting new data in skbs queued over {ip,ip6,nfnetlink}_queue triggers a SKB_LINEAR_ASSERT in skb_put(). Going back through the git history, it seems this bug is present since at least 2.6.12-rc2, probably even since the removal of skb_linearize() for netfilter. Linearize non-linear skbs through skb_copy_expand() when enlarging them. Tested by Thomas, fixes bugzilla #9933. Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
| * ipv4/fib_hash.c: fix NULL dereferenceAdrian Bunk2008-02-19
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Unless I miss a guaranteed relation between between "f" and "new_fa->fa_info" this patch is required for fixing a NULL dereference introduced by commit a6501e080c318f8d4467679d17807f42b3a33cd5 ("[IPV4] FIB_HASH: Reduce memory needs and speedup lookups") and spotted by the Coverity checker. Eric Dumazet says: Hum, you are right, kmem_cache_free() doesnt allow a NULL object, like kfree() does. Signed-off-by: Adrian Bunk <bunk@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
| * net/9p/trans_virtio.c: kmalloc() enough memoryAdrian Bunk2008-02-19
| | | | | | | | | | | | | | | | The Coverity checker spotted that less memory than required was allocated. Signed-off-by: Adrian Bunk <bunk@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
| * [RTNL]: Add missing link netlink attribute policy definitionsThomas Graf2008-02-19
| | | | | | | | | | | | | | | | | | | | | | | | IFLA_LINK is no longer a write-only attribute on the kernel side and must thus be validated. Same goes for the newly introduced IFLA_LINKINFO. Fixes undefined behaviour if either of the attributes are not well formed. Signed-off-by: Thomas Graf <tgraf@suug.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
| * [NET]: Messed multicast lists after dev_mc_sync/unsyncJorge Boncompte [DTI2]2008-02-19
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Commit a0a400d79e3dd7843e7e81baa3ef2957bdc292d0 ("[NET]: dev_mcast: add multicast list synchronization helpers") from you introduced a new field "da_synced" to struct dev_addr_list that is not properly initialized to 0. So when any of the current users (8021q, macvlan, mac80211) calls dev_mc_sync/unsync they mess the address list for both devices. The attached patch fixed it for me and avoid future problems. Signed-off-by: Jorge Boncompte [DTI2] <jorge@dti2.net> Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
* | Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc-2.6Linus Torvalds2008-02-24
|\ \ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | * git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc-2.6: [SPARC64]: make IOMMU code respect the segment boundary limits [SPARC64]: Fix cpu trampoline et al. mismatch warnings. [SPARC64]: More sparse warning fixes in process.c [SPARC64]: Fix sparse warning wrt. fault_in_user_windows. [SPARC64]: Kill show_regs32(). [SPARC64]: Fix sparse warnings wrt. __show_regs(). [SPARC64]: Kill show_stackframe{,32}(). [SPARC64]: Fix sparse warnings wrt. machine_alt_power_off().
| * | [SPARC64]: make IOMMU code respect the segment boundary limitsFUJITA Tomonori2008-02-21
| | | | | | | | | | | | | | | Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp> Signed-off-by: David S. Miller <davem@davemloft.net>
| * | [SPARC64]: Fix cpu trampoline et al. mismatch warnings.Sam Ravnborg2008-02-21
| | | | | | | | | | | | | | | Signed-off-by: Sam Ravnborg <sam@ravnborg.org> Signed-off-by: David S. Miller <davem@davemloft.net>
| * | [SPARC64]: More sparse warning fixes in process.cDavid S. Miller2008-02-20
| | | | | | | | | | | | | | | | | | | | | | | | arch/sparc64/kernel/process.c:504:17: warning: symbol 'sparc_do_fork' was not declared. Should it be static? arch/sparc64/kernel/process.c:655:5: warning: symbol 'dump_fpu' was not declared. Should it be static? arch/sparc64/kernel/process.c:708:16: warning: symbol 'sparc_execve' was not declared. Should it be static? Signed-off-by: David S. Miller <davem@davemloft.net>
| * | [SPARC64]: Fix sparse warning wrt. fault_in_user_windows.David S. Miller2008-02-20
| | | | | | | | | | | | | | | | | | arch/sparc64/kernel/process.c:467:6: warning: symbol 'fault_in_user_windows' was not declared. Should it be static? Signed-off-by: David S. Miller <davem@davemloft.net>
| * | [SPARC64]: Kill show_regs32().David S. Miller2008-02-20
| | | | | | | | | | | | | | | | | | Unused, noticed via sparse. Signed-off-by: David S. Miller <davem@davemloft.net>
| * | [SPARC64]: Fix sparse warnings wrt. __show_regs().David S. Miller2008-02-19
| | | | | | | | | | | | | | | | | | arch/sparc64/kernel/process.c:219:6: warning: symbol '__show_regs' was not declared. Should it be static? Signed-off-by: David S. Miller <davem@davemloft.net>
| * | [SPARC64]: Kill show_stackframe{,32}().David S. Miller2008-02-19
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Noticed via sparse: arch/sparc64/kernel/process.c:215:6: warning: symbol 'show_stackframe' was not declared. Should it be static? arch/sparc64/kernel/process.c:243:6: warning: symbol 'show_stackframe32' was not declared. Should it be static? It is totally unused. Signed-off-by: David S. Miller <davem@davemloft.net>
| * | [SPARC64]: Fix sparse warnings wrt. machine_alt_power_off().David S. Miller2008-02-19
| | | | | | | | | | | | | | | | | | arch/sparc64/kernel/process.c:123:6: warning: symbol 'machine_alt_power_off' was not declared. Should it be static? Signed-off-by: David S. Miller <davem@davemloft.net>
* | | Fix u132-hcd.c compile errorMirco Tischler2008-02-23
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This fixes the following compile error caused by commit 3a2d5b700132f35401f1d9e22fe3c2cab02c2549 ("PM: Introduce PM_EVENT_HIBERNATE callback state") CC [M] drivers/usb/host/u132-hcd.o drivers/usb/host/u132-hcd.c: In function ‘u132_suspend’: drivers/usb/host/u132-hcd.c:3224: error: expected expression before ‘int’ drivers/usb/host/u132-hcd.c:3225: error: ‘ports’ undeclared (first use in this function) ... Signed-off-by: Mirco Tischler <mt-ml@gmx.de> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
* | | fix vmsas.c file permissionsOliver Pinter2008-02-23
| | | | | | | | | | | | | | | Signed-off-by: Oliver Pinter <oliver.pntr@gmail.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
* | | Add memory barrier semantics to wake_up() & coLinus Torvalds2008-02-23
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Oleg Nesterov and others have pointed out that on some architectures, the traditional sequence of set_current_state(TASK_INTERRUPTIBLE); if (CONDITION) return; schedule(); is racy wrt another CPU doing CONDITION = 1; wake_up_process(p); because while set_current_state() has a memory barrier separating setting of the TASK_INTERRUPTIBLE state from reading of the CONDITION variable, there is no such memory barrier on the wakeup side. Now, wake_up_process() does actually take a spinlock before it reads and sets the task state on the waking side, and on x86 (and many other architectures) that spinlock is in fact equivalent to a memory barrier, but that is not generally guaranteed. The write that sets CONDITION could move into the critical region protected by the runqueue spinlock. However, adding a smp_wmb() to before the spinlock should now order the writing of CONDITION wrt the lock itself, which in turn is ordered wrt the accesses within the spinlock (which includes the reading of the old state). This should thus close the race (which probably has never been seen in practice, but since smp_wmb() is a no-op on x86, it's not like this will make anything worse either on the most common architecture where the spinlock already gave the required protection). Acked-by: Oleg Nesterov <oleg@tv-sign.ru> Acked-by: Dmitry Adamushko <dmitry.adamushko@gmail.com> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Nick Piggin <nickpiggin@yahoo.com.au> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
generated by cgit v1.2.2 (git 2.25.0) at 2025-02-10 02:54:01 -0500