user/sven/linux.git/include/linux, branch v4.19.17

block: use rcu_work instead of call_rcu to avoid sleep in softirq

2019-01-22T20:40:35Z

commit 94a2c3a32b62e868dc1e3d854326745a7f1b8c7a upstream. We recently got a stack by syzkaller like this: BUG: sleeping function called from invalid context at mm/slab.h:361 in_atomic(): 1, irqs_disabled(): 0, pid: 6644, name: blkid INFO: lockdep is turned off. CPU: 1 PID: 6644 Comm: blkid Not tainted 4.4.163-514.55.6.9.x86_64+ #76 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.10.2-1ubuntu1 04/01/2014 0000000000000000 5ba6a6b879e50c00 ffff8801f6b07b10 ffffffff81cb2194 0000000041b58ab3 ffffffff833c7745 ffffffff81cb2080 5ba6a6b879e50c00 0000000000000000 0000000000000001 0000000000000004 0000000000000000 Call Trace: [] __dump_stack lib/dump_stack.c:15 [inline] [] dump_stack+0x114/0x1a0 lib/dump_stack.c:51 [] ___might_sleep+0x291/0x490 kernel/sched/core.c:7675 [] __might_sleep+0xb3/0x270 kernel/sched/core.c:7637 [] slab_pre_alloc_hook mm/slab.h:361 [inline] [] slab_alloc_node mm/slub.c:2610 [inline] [] slab_alloc mm/slub.c:2692 [inline] [] kmem_cache_alloc_trace+0x2c3/0x5c0 mm/slub.c:2709 [] kmalloc include/linux/slab.h:479 [inline] [] kzalloc include/linux/slab.h:623 [inline] [] kobject_uevent_env+0x2c7/0x1150 lib/kobject_uevent.c:227 [] kobject_uevent+0x1f/0x30 lib/kobject_uevent.c:374 [] kobject_cleanup lib/kobject.c:633 [inline] [] kobject_release+0x229/0x440 lib/kobject.c:675 [] kref_sub include/linux/kref.h:73 [inline] [] kref_put include/linux/kref.h:98 [inline] [] kobject_put+0x72/0xd0 lib/kobject.c:692 [] put_device+0x25/0x30 drivers/base/core.c:1237 [] delete_partition_rcu_cb+0x1d4/0x2f0 block/partition-generic.c:232 [] __rcu_reclaim kernel/rcu/rcu.h:118 [inline] [] rcu_do_batch kernel/rcu/tree.c:2705 [inline] [] invoke_rcu_callbacks kernel/rcu/tree.c:2973 [inline] [] __rcu_process_callbacks kernel/rcu/tree.c:2940 [inline] [] rcu_process_callbacks+0x59c/0x1c70 kernel/rcu/tree.c:2957 [] __do_softirq+0x299/0xe20 kernel/softirq.c:273 [] invoke_softirq kernel/softirq.c:350 [inline] [] irq_exit+0x216/0x2c0 kernel/softirq.c:391 [] exiting_irq arch/x86/include/asm/apic.h:652 [inline] [] smp_apic_timer_interrupt+0x8b/0xc0 arch/x86/kernel/apic/apic.c:926 [] apic_timer_interrupt+0xa5/0xb0 arch/x86/entry/entry_64.S:746 [] ? audit_kill_trees+0x180/0x180 [] fd_install+0x57/0x80 fs/file.c:626 [] do_sys_open+0x45e/0x550 fs/open.c:1043 [] SYSC_open fs/open.c:1055 [inline] [] SyS_open+0x32/0x40 fs/open.c:1050 [] entry_SYSCALL_64_fastpath+0x1e/0x9a In softirq context, we call rcu callback function delete_partition_rcu_cb(), which may allocate memory by kzalloc with GFP_KERNEL flag. If the allocation cannot be satisfied, it may sleep. However, That is not allowed in softirq contex. Although we found this problem on linux 4.4, the latest kernel version seems to have this problem as well. And it is very similar to the previous one: https://lkml.org/lkml/2018/7/9/391 Fix it by using RCU workqueue, which allows sleep. Reviewed-by: Paul E. McKenney Signed-off-by: Yufen Yu Signed-off-by: Jens Axboe Signed-off-by: Greg Kroah-Hartman

MIPS: BCM47XX: Setup struct device for the SoC

2019-01-22T20:40:33Z

commit 321c46b91550adc03054125fa7a1639390608e1a upstream. So far we never had any device registered for the SoC. This resulted in some small issues that we kept ignoring like: 1) Not working GPIOLIB_IRQCHIP (gpiochip_irqchip_add_key() failing) 2) Lack of proper tree in the /sys/devices/ 3) mips_dma_alloc_coherent() silently handling empty coherent_dma_mask Kernel 4.19 came with a lot of DMA changes and caused a regression on bcm47xx. Starting with the commit f8c55dc6e828 ("MIPS: use generic dma noncoherent ops for simple noncoherent platforms") DMA coherent allocations just fail. Example: [ 1.114914] bgmac_bcma bcma0:2: Allocation of TX ring 0x200 failed [ 1.121215] bgmac_bcma bcma0:2: Unable to alloc memory for DMA [ 1.127626] bgmac_bcma: probe of bcma0:2 failed with error -12 [ 1.133838] bgmac_bcma: Broadcom 47xx GBit MAC driver loaded The bgmac driver also triggers a WARNING: [ 0.959486] ------------[ cut here ]------------ [ 0.964387] WARNING: CPU: 0 PID: 1 at ./include/linux/dma-mapping.h:516 bgmac_enet_probe+0x1b4/0x5c4 [ 0.973751] Modules linked in: [ 0.976913] CPU: 0 PID: 1 Comm: swapper Not tainted 4.19.9 #0 [ 0.982750] Stack : 804a0000 804597c4 00000000 00000000 80458fd8 8381bc2c 838282d4 80481a47 [ 0.991367] 8042e3ec 00000001 804d38f0 00000204 83980000 00000065 8381bbe0 6f55b24f [ 0.999975] 00000000 00000000 80520000 00002018 00000000 00000075 00000007 00000000 [ 1.008583] 00000000 80480000 000ee811 00000000 00000000 00000000 80432c00 80248db8 [ 1.017196] 00000009 00000204 83980000 803ad7b0 00000000 801feeec 00000000 804d0000 [ 1.025804] ... [ 1.028325] Call Trace: [ 1.030875] [<8000aef8>] show_stack+0x58/0x100 [ 1.035513] [<8001f8b4>] __warn+0xe4/0x118 [ 1.039708] [<8001f9a4>] warn_slowpath_null+0x48/0x64 [ 1.044935] [<80248db8>] bgmac_enet_probe+0x1b4/0x5c4 [ 1.050101] [<802498e0>] bgmac_probe+0x558/0x590 [ 1.054906] [<80252fd0>] bcma_device_probe+0x38/0x70 [ 1.060017] [<8020e1e8>] really_probe+0x170/0x2e8 [ 1.064891] [<8020e714>] __driver_attach+0xa4/0xec [ 1.069784] [<8020c1e0>] bus_for_each_dev+0x58/0xb0 [ 1.074833] [<8020d590>] bus_add_driver+0xf8/0x218 [ 1.079731] [<8020ef24>] driver_register+0xcc/0x11c [ 1.084804] [<804b54cc>] bgmac_init+0x1c/0x44 [ 1.089258] [<8000121c>] do_one_initcall+0x7c/0x1a0 [ 1.094343] [<804a1d34>] kernel_init_freeable+0x150/0x218 [ 1.099886] [<803a082c>] kernel_init+0x10/0x104 [ 1.104583] [<80005878>] ret_from_kernel_thread+0x14/0x1c [ 1.110107] ---[ end trace f441c0d873d1fb5b ]--- This patch setups a "struct device" (and passes it to the bcma) which allows fixing all the mentioned problems. It'll also require a tiny bcma patch which will follow through the wireless tree & its maintainer. Fixes: f8c55dc6e828 ("MIPS: use generic dma noncoherent ops for simple noncoherent platforms") Signed-off-by: Rafał Miłecki Signed-off-by: Paul Burton Acked-by: Hauke Mehrtens Cc: Christoph Hellwig Cc: Linus Walleij Cc: linux-wireless@vger.kernel.org Cc: Ralf Baechle Cc: James Hogan Cc: linux-mips@linux-mips.org Cc: linux-kernel@vger.kernel.org Cc: stable@vger.kernel.org # v4.19+ Signed-off-by: Greg Kroah-Hartman

sunrpc: use-after-free in svc_process_common()

2019-01-16T21:04:37Z

commit d4b09acf924b84bae77cad090a9d108e70b43643 upstream. if node have NFSv41+ mounts inside several net namespaces it can lead to use-after-free in svc_process_common() svc_process_common() /* Setup reply header */ rqstp->rq_xprt->xpt_ops->xpo_prep_reply_hdr(rqstp); <<< HERE svc_process_common() can use incorrect rqstp->rq_xprt, its caller function bc_svc_process() takes it from serv->sv_bc_xprt. The problem is that serv is global structure but sv_bc_xprt is assigned per-netnamespace. According to Trond, the whole "let's set up rqstp->rq_xprt for the back channel" is nothing but a giant hack in order to work around the fact that svc_process_common() uses it to find the xpt_ops, and perform a couple of (meaningless for the back channel) tests of xpt_flags. All we really need in svc_process_common() is to be able to run rqstp->rq_xprt->xpt_ops->xpo_prep_reply_hdr() Bruce J Fields points that this xpo_prep_reply_hdr() call is an awfully roundabout way just to do "svc_putnl(resv, 0);" in the tcp case. This patch does not initialiuze rqstp->rq_xprt in bc_svc_process(), now it calls svc_process_common() with rqstp->rq_xprt = NULL. To adjust reply header svc_process_common() just check rqstp->rq_prot and calls svc_tcp_prep_reply_hdr() for tcp case. To handle rqstp->rq_xprt = NULL case in functions called from svc_process_common() patch intruduces net namespace pointer svc_rqst->rq_bc_net and adjust SVC_NET() definition. Some other function was also adopted to properly handle described case. Signed-off-by: Vasily Averin Cc: stable@vger.kernel.org Fixes: 23c20ecd4475 ("NFS: callback up - users counting cleanup") Signed-off-by: J. Bruce Fields v2: added lost extern svc_tcp_prep_reply_hdr() Signed-off-by: Vasily Averin Signed-off-by: Greg Kroah-Hartman

x86, modpost: Replace last remnants of RETPOLINE with CONFIG_RETPOLINE

2019-01-16T21:04:29Z

commit e4f358916d528d479c3c12bd2fd03f2d5a576380 upstream. Commit 4cd24de3a098 ("x86/retpoline: Make CONFIG_RETPOLINE depend on compiler support") replaced the RETPOLINE define with CONFIG_RETPOLINE checks. Remove the remaining pieces. [ bp: Massage commit message. ] Fixes: 4cd24de3a098 ("x86/retpoline: Make CONFIG_RETPOLINE depend on compiler support") Signed-off-by: WANG Chao Signed-off-by: Borislav Petkov Reviewed-by: Zhenzhong Duan Reviewed-by: Masahiro Yamada Cc: "H. Peter Anvin" Cc: Andi Kleen Cc: Andrew Morton Cc: Andy Lutomirski Cc: Arnd Bergmann Cc: Daniel Borkmann Cc: David Woodhouse Cc: Geert Uytterhoeven Cc: Jessica Yu Cc: Jiri Kosina Cc: Kees Cook Cc: Konrad Rzeszutek Wilk Cc: Luc Van Oostenryck Cc: Michal Marek Cc: Miguel Ojeda Cc: Peter Zijlstra Cc: Tim Chen Cc: Vasily Gorbik Cc: linux-kbuild@vger.kernel.org Cc: srinivas.eeda@oracle.com Cc: stable Cc: x86-ml Link: https://lkml.kernel.org/r/20181210163725.95977-1-chao.wang@ucloud.cn Signed-off-by: Greg Kroah-Hartman

mm, hmm: use devm semantics for hmm_devmem_{add, remove}

2019-01-13T08:51:04Z

commit 58ef15b765af0d2cbe6799ec564f1dc485010ab8 upstream. devm semantics arrange for resources to be torn down when device-driver-probe fails or when device-driver-release completes. Similar to devm_memremap_pages() there is no need to support an explicit remove operation when the users properly adhere to devm semantics. Note that devm_kzalloc() automatically handles allocating node-local memory. Link: http://lkml.kernel.org/r/154275559545.76910.9186690723515469051.stgit@dwillia2-desk3.amr.corp.intel.com Signed-off-by: Dan Williams Reviewed-by: Christoph Hellwig Reviewed-by: Jérôme Glisse Cc: "Jérôme Glisse" Cc: Logan Gunthorpe Cc: Balbir Singh Cc: Michal Hocko Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds Signed-off-by: Greg Kroah-Hartman

mm, devm_memremap_pages: fix shutdown handling

2019-01-13T08:51:04Z

commit a95c90f1e2c253b280385ecf3d4ebfe476926b28 upstream. The last step before devm_memremap_pages() returns success is to allocate a release action, devm_memremap_pages_release(), to tear the entire setup down. However, the result from devm_add_action() is not checked. Checking the error from devm_add_action() is not enough. The api currently relies on the fact that the percpu_ref it is using is killed by the time the devm_memremap_pages_release() is run. Rather than continue this awkward situation, offload the responsibility of killing the percpu_ref to devm_memremap_pages_release() directly. This allows devm_memremap_pages() to do the right thing relative to init failures and shutdown. Without this change we could fail to register the teardown of devm_memremap_pages(). The likelihood of hitting this failure is tiny as small memory allocations almost always succeed. However, the impact of the failure is large given any future reconfiguration, or disable/enable, of an nvdimm namespace will fail forever as subsequent calls to devm_memremap_pages() will fail to setup the pgmap_radix since there will be stale entries for the physical address range. An argument could be made to require that the ->kill() operation be set in the @pgmap arg rather than passed in separately. However, it helps code readability, tracking the lifetime of a given instance, to be able to grep the kill routine directly at the devm_memremap_pages() call site. Link: http://lkml.kernel.org/r/154275558526.76910.7535251937849268605.stgit@dwillia2-desk3.amr.corp.intel.com Signed-off-by: Dan Williams Fixes: e8d513483300 ("memremap: change devm_memremap_pages interface...") Reviewed-by: "Jérôme Glisse" Reported-by: Logan Gunthorpe Reviewed-by: Logan Gunthorpe Reviewed-by: Christoph Hellwig Cc: Balbir Singh Cc: Michal Hocko Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds Signed-off-by: Greg Kroah-Hartman

netfilter: nf_tables: fix suspicious RCU usage in nft_chain_stats_replace()

2019-01-13T08:50:57Z

[ Upstream commit 4c05ec47384ab3627b62814e8f886e90cc38ce15 ] basechain->stats is rcu protected data which is updated from nft_chain_stats_replace(). This function is executed from the commit phase which holds the pernet nf_tables commit mutex - not the global nfnetlink subsystem mutex. Test commands to reproduce the problem are: %iptables-nft -I INPUT %iptables-nft -Z %iptables-nft -Z This patch uses RCU calls to handle basechain->stats updates to fix a splat that looks like: [89279.358755] ============================= [89279.363656] WARNING: suspicious RCU usage [89279.368458] 4.20.0-rc2+ #44 Tainted: G W L [89279.374661] ----------------------------- [89279.379542] net/netfilter/nf_tables_api.c:1404 suspicious rcu_dereference_protected() usage! [...] [89279.406556] 1 lock held by iptables-nft/5225: [89279.411728] #0: 00000000bf45a000 (&net->nft.commit_mutex){+.+.}, at: nf_tables_valid_genid+0x1f/0x70 [nf_tables] [89279.424022] stack backtrace: [89279.429236] CPU: 0 PID: 5225 Comm: iptables-nft Tainted: G W L 4.20.0-rc2+ #44 [89279.430135] Call Trace: [89279.430135] dump_stack+0xc9/0x16b [89279.430135] ? show_regs_print_info+0x5/0x5 [89279.430135] ? lockdep_rcu_suspicious+0x117/0x160 [89279.430135] nft_chain_commit_update+0x4ea/0x640 [nf_tables] [89279.430135] ? sched_clock_local+0xd4/0x140 [89279.430135] ? check_flags.part.35+0x440/0x440 [89279.430135] ? __rhashtable_remove_fast.constprop.67+0xec0/0xec0 [nf_tables] [89279.430135] ? sched_clock_cpu+0x126/0x170 [89279.430135] ? find_held_lock+0x39/0x1c0 [89279.430135] ? hlock_class+0x140/0x140 [89279.430135] ? is_bpf_text_address+0x5/0xf0 [89279.430135] ? check_flags.part.35+0x440/0x440 [89279.430135] ? __lock_is_held+0xb4/0x140 [89279.430135] nf_tables_commit+0x2555/0x39c0 [nf_tables] Fixes: f102d66b335a4 ("netfilter: nf_tables: use dedicated mutex to guard transactions") Signed-off-by: Taehee Yoo Signed-off-by: Pablo Neira Ayuso Signed-off-by: Sasha Levin

platform-msi: Free descriptors in platform_msi_domain_free()

2019-01-09T16:38:42Z

commit 81b1e6e6a8590a19257e37a1633bec098d499c57 upstream. Since the addition of platform MSI support, there were two helpers supposed to allocate/free IRQs for a device: platform_msi_domain_alloc_irqs() platform_msi_domain_free_irqs() In these helpers, IRQ descriptors are allocated in the "alloc" routine while they are freed in the "free" one. Later, two other helpers have been added to handle IRQ domains on top of MSI domains: platform_msi_domain_alloc() platform_msi_domain_free() Seen from the outside, the logic is pretty close with the former helpers and people used it with the same logic as before: a platform_msi_domain_alloc() call should be balanced with a platform_msi_domain_free() call. While this is probably what was intended to do, the platform_msi_domain_free() does not remove/free the IRQ descriptor(s) created/inserted in platform_msi_domain_alloc(). One effect of such situation is that removing a module that requested an IRQ will let one orphaned IRQ descriptor (with an allocated MSI entry) in the device descriptors list. Next time the module will be inserted back, one will observe that the allocation will happen twice in the MSI domain, one time for the remaining descriptor, one time for the new one. It also has the side effect to quickly overshoot the maximum number of allocated MSI and then prevent any module requesting an interrupt in the same domain to be inserted anymore. This situation has been met with loops of insertion/removal of the mvpp2.ko module (requesting 15 MSIs each time). Fixes: 552c494a7666 ("platform-msi: Allow creation of a MSI-based stacked irq domain") Cc: stable@vger.kernel.org Signed-off-by: Miquel Raynal Signed-off-by: Marc Zyngier Signed-off-by: Greg Kroah-Hartman

ptr_ring: wrap back ->producer in __ptr_ring_swap_queue()

2019-01-09T16:38:33Z

[ Upstream commit aff6db454599d62191aabc208930e891748e4322 ] __ptr_ring_swap_queue() tries to move pointers from the old ring to the new one, but it forgets to check if ->producer is beyond the new size at the end of the operation. This leads to an out-of-bound access in __ptr_ring_produce() as reported by syzbot. Reported-by: syzbot+8993c0fa96d57c399735@syzkaller.appspotmail.com Fixes: 5d49de532002 ("ptr_ring: resize support") Cc: "Michael S. Tsirkin" Cc: John Fastabend Cc: Jason Wang Signed-off-by: Cong Wang Acked-by: Michael S. Tsirkin Signed-off-by: David S. Miller Signed-off-by: Greg Kroah-Hartman

mm: don't miss the last page because of round-off error

2018-12-29T12:37:59Z

commit 68600f623d69da428c6163275f97ca126e1a8ec5 upstream. I've noticed, that dying memory cgroups are often pinned in memory by a single pagecache page. Even under moderate memory pressure they sometimes stayed in such state for a long time. That looked strange. My investigation showed that the problem is caused by applying the LRU pressure balancing math: scan = div64_u64(scan * fraction[lru], denominator), where denominator = fraction[anon] + fraction[file] + 1. Because fraction[lru] is always less than denominator, if the initial scan size is 1, the result is always 0. This means the last page is not scanned and has no chances to be reclaimed. Fix this by rounding up the result of the division. In practice this change significantly improves the speed of dying cgroups reclaim. [guro@fb.com: prevent double calculation of DIV64_U64_ROUND_UP() arguments] Link: http://lkml.kernel.org/r/20180829213311.GA13501@castle Link: http://lkml.kernel.org/r/20180827162621.30187-3-guro@fb.com Signed-off-by: Roman Gushchin Reviewed-by: Andrew Morton Cc: Johannes Weiner Cc: Michal Hocko Cc: Tejun Heo Cc: Rik van Riel Cc: Konstantin Khlebnikov Cc: Matthew Wilcox Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds Signed-off-by: Greg Kroah-Hartman