From c92a7a52243871eb10e0ea2260685def47fb5094 Mon Sep 17 00:00:00 2001 From: David Vernet Date: Wed, 12 Oct 2022 18:20:14 -0500 Subject: bpf: Allow bpf_user_ringbuf_drain() callbacks to return 1 The bpf_user_ringbuf_drain() helper function allows a BPF program to specify a callback that is invoked when draining entries from a BPF_MAP_TYPE_USER_RINGBUF ring buffer map. The API is meant to allow the callback to return 0 if it wants to continue draining samples, and 1 if it's done draining. Unfortunately, bpf_user_ringbuf_drain() landed shortly after commit 1bfe26fb0827 ("bpf: Add verifier support for custom callback return range"), which changed the default behavior of callbacks to only support returning 0. This patch corrects that oversight by allowing bpf_user_ringbuf_drain() callbacks to return 0 or 1. A follow-on patch will update the user_ringbuf selftests to also return 1 from a bpf_user_ringbuf_drain() callback to prevent this from regressing in the future. Fixes: 205715673844 ("bpf: Add bpf_user_ringbuf_drain() helper") Signed-off-by: David Vernet Signed-off-by: Andrii Nakryiko Link: https://lore.kernel.org/bpf/20221012232015.1510043-2-void@manifault.com --- kernel/bpf/verifier.c | 1 + 1 file changed, 1 insertion(+) (limited to 'kernel') diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c index 6f6d2d511c06..9ab7188d8f68 100644 --- a/kernel/bpf/verifier.c +++ b/kernel/bpf/verifier.c @@ -6946,6 +6946,7 @@ static int set_user_ringbuf_callback_state(struct bpf_verifier_env *env, __mark_reg_not_init(env, &callee->regs[BPF_REG_5]); callee->in_callback_fn = true; + callee->callback_ret_range = tnum_range(0, 1); return 0; } -- cgit v1.2.3 From ea68376c8bed5cd156900852aada20c3a0874d17 Mon Sep 17 00:00:00 2001 From: Stanislav Fomichev Date: Fri, 14 Oct 2022 17:24:44 -0700 Subject: bpf: prevent decl_tag from being referenced in func_proto Syzkaller was able to hit the following issue: ------------[ cut here ]------------ WARNING: CPU: 0 PID: 3609 at kernel/bpf/btf.c:1946 btf_type_id_size+0x2d5/0x9d0 kernel/bpf/btf.c:1946 Modules linked in: CPU: 0 PID: 3609 Comm: syz-executor361 Not tainted 6.0.0-syzkaller-02734-g0326074ff465 #0 Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 09/22/2022 RIP: 0010:btf_type_id_size+0x2d5/0x9d0 kernel/bpf/btf.c:1946 Code: ef e8 7f 8e e4 ff 41 83 ff 0b 77 28 f6 44 24 10 18 75 3f e8 6d 91 e4 ff 44 89 fe bf 0e 00 00 00 e8 20 8e e4 ff e8 5b 91 e4 ff <0f> 0b 45 31 f6 e9 98 02 00 00 41 83 ff 12 74 18 e8 46 91 e4 ff 44 RSP: 0018:ffffc90003cefb40 EFLAGS: 00010293 RAX: 0000000000000000 RBX: 0000000000000002 RCX: 0000000000000000 RDX: ffff8880259c0000 RSI: ffffffff81968415 RDI: 0000000000000005 RBP: ffff88801270ca00 R08: 0000000000000005 R09: 000000000000000e R10: 0000000000000011 R11: 0000000000000000 R12: 0000000000000000 R13: 0000000000000011 R14: ffff888026ee6424 R15: 0000000000000011 FS: 000055555641b300(0000) GS:ffff8880b9a00000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 0000000000f2e258 CR3: 000000007110e000 CR4: 00000000003506f0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 Call Trace: btf_func_proto_check kernel/bpf/btf.c:4447 [inline] btf_check_all_types kernel/bpf/btf.c:4723 [inline] btf_parse_type_sec kernel/bpf/btf.c:4752 [inline] btf_parse kernel/bpf/btf.c:5026 [inline] btf_new_fd+0x1926/0x1e70 kernel/bpf/btf.c:6892 bpf_btf_load kernel/bpf/syscall.c:4324 [inline] __sys_bpf+0xb7d/0x4cf0 kernel/bpf/syscall.c:5010 __do_sys_bpf kernel/bpf/syscall.c:5069 [inline] __se_sys_bpf kernel/bpf/syscall.c:5067 [inline] __x64_sys_bpf+0x75/0xb0 kernel/bpf/syscall.c:5067 do_syscall_x64 arch/x86/entry/common.c:50 [inline] do_syscall_64+0x35/0xb0 arch/x86/entry/common.c:80 entry_SYSCALL_64_after_hwframe+0x63/0xcd RIP: 0033:0x7f0fbae41c69 Code: 28 c3 e8 2a 14 00 00 66 2e 0f 1f 84 00 00 00 00 00 48 89 f8 48 89 f7 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 c7 c1 c0 ff ff ff f7 d8 64 89 01 48 RSP: 002b:00007ffc8aeb6228 EFLAGS: 00000246 ORIG_RAX: 0000000000000141 RAX: ffffffffffffffda RBX: 0000000000000000 RCX: 00007f0fbae41c69 RDX: 0000000000000020 RSI: 0000000020000140 RDI: 0000000000000012 RBP: 00007f0fbae05e10 R08: 0000000000000000 R09: 0000000000000000 R10: 00000000ffffffff R11: 0000000000000246 R12: 00007f0fbae05ea0 R13: 0000000000000000 R14: 0000000000000000 R15: 0000000000000000 Looks like it tries to create a func_proto which return type is decl_tag. For the details, see Martin's spot on analysis in [0]. 0: https://lore.kernel.org/bpf/CAKH8qBuQDLva_hHxxBuZzyAcYNO4ejhovz6TQeVSk8HY-2SO6g@mail.gmail.com/T/#mea6524b3fcd6298347432226e81b1e6155efc62c Cc: Yonghong Song Cc: Martin KaFai Lau Fixes: bd16dee66ae4 ("bpf: Add BTF_KIND_DECL_TAG typedef support") Reported-by: syzbot+d8bd751aef7c6b39a344@syzkaller.appspotmail.com Signed-off-by: Stanislav Fomichev Acked-by: Yonghong Song Link: https://lore.kernel.org/r/20221015002444.2680969-2-sdf@google.com Signed-off-by: Martin KaFai Lau --- kernel/bpf/btf.c | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'kernel') diff --git a/kernel/bpf/btf.c b/kernel/bpf/btf.c index eba603cec2c5..35c07afac924 100644 --- a/kernel/bpf/btf.c +++ b/kernel/bpf/btf.c @@ -4436,6 +4436,11 @@ static int btf_func_proto_check(struct btf_verifier_env *env, return -EINVAL; } + if (btf_type_is_resolve_source_only(ret_type)) { + btf_verifier_log_type(env, t, "Invalid return type"); + return -EINVAL; + } + if (btf_type_needs_resolve(ret_type) && !env_type_is_resolved(env, ret_type_id)) { err = btf_resolve(env, ret_type, ret_type_id); -- cgit v1.2.3 From dbe69b29988465b011f198f2797b1c2b6980b50e Mon Sep 17 00:00:00 2001 From: Jiri Olsa Date: Tue, 18 Oct 2022 09:59:34 +0200 Subject: bpf: Fix dispatcher patchable function entry to 5 bytes nop The patchable_function_entry(5) might output 5 single nop instructions (depends on toolchain), which will clash with bpf_arch_text_poke check for 5 bytes nop instruction. Adding early init call for dispatcher that checks and change the patchable entry into expected 5 nop instruction if needed. There's no need to take text_mutex, because we are using it in early init call which is called at pre-smp time. Fixes: ceea991a019c ("bpf: Move bpf_dispatcher function out of ftrace locations") Signed-off-by: Jiri Olsa Acked-by: Peter Zijlstra (Intel) Link: https://lore.kernel.org/r/20221018075934.574415-1-jolsa@kernel.org Signed-off-by: Alexei Starovoitov --- arch/x86/net/bpf_jit_comp.c | 13 +++++++++++++ include/linux/bpf.h | 14 +++++++++++++- kernel/bpf/dispatcher.c | 6 ++++++ 3 files changed, 32 insertions(+), 1 deletion(-) (limited to 'kernel') diff --git a/arch/x86/net/bpf_jit_comp.c b/arch/x86/net/bpf_jit_comp.c index 99620428ad78..00127abd89ee 100644 --- a/arch/x86/net/bpf_jit_comp.c +++ b/arch/x86/net/bpf_jit_comp.c @@ -11,6 +11,7 @@ #include #include #include +#include #include #include #include @@ -388,6 +389,18 @@ out: return ret; } +int __init bpf_arch_init_dispatcher_early(void *ip) +{ + const u8 *nop_insn = x86_nops[5]; + + if (is_endbr(*(u32 *)ip)) + ip += ENDBR_INSN_SIZE; + + if (memcmp(ip, nop_insn, X86_PATCH_SIZE)) + text_poke_early(ip, nop_insn, X86_PATCH_SIZE); + return 0; +} + int bpf_arch_text_poke(void *ip, enum bpf_text_poke_type t, void *old_addr, void *new_addr) { diff --git a/include/linux/bpf.h b/include/linux/bpf.h index 9e7d46d16032..0566705c1d4e 100644 --- a/include/linux/bpf.h +++ b/include/linux/bpf.h @@ -27,6 +27,7 @@ #include #include #include +#include struct bpf_verifier_env; struct bpf_verifier_log; @@ -970,6 +971,8 @@ struct bpf_trampoline *bpf_trampoline_get(u64 key, struct bpf_attach_target_info *tgt_info); void bpf_trampoline_put(struct bpf_trampoline *tr); int arch_prepare_bpf_dispatcher(void *image, void *buf, s64 *funcs, int num_funcs); +int __init bpf_arch_init_dispatcher_early(void *ip); + #define BPF_DISPATCHER_INIT(_name) { \ .mutex = __MUTEX_INITIALIZER(_name.mutex), \ .func = &_name##_func, \ @@ -983,6 +986,13 @@ int arch_prepare_bpf_dispatcher(void *image, void *buf, s64 *funcs, int num_func }, \ } +#define BPF_DISPATCHER_INIT_CALL(_name) \ + static int __init _name##_init(void) \ + { \ + return bpf_arch_init_dispatcher_early(_name##_func); \ + } \ + early_initcall(_name##_init) + #ifdef CONFIG_X86_64 #define BPF_DISPATCHER_ATTRIBUTES __attribute__((patchable_function_entry(5))) #else @@ -1000,7 +1010,9 @@ int arch_prepare_bpf_dispatcher(void *image, void *buf, s64 *funcs, int num_func } \ EXPORT_SYMBOL(bpf_dispatcher_##name##_func); \ struct bpf_dispatcher bpf_dispatcher_##name = \ - BPF_DISPATCHER_INIT(bpf_dispatcher_##name); + BPF_DISPATCHER_INIT(bpf_dispatcher_##name); \ + BPF_DISPATCHER_INIT_CALL(bpf_dispatcher_##name); + #define DECLARE_BPF_DISPATCHER(name) \ unsigned int bpf_dispatcher_##name##_func( \ const void *ctx, \ diff --git a/kernel/bpf/dispatcher.c b/kernel/bpf/dispatcher.c index fa64b80b8bca..04f0a045dcaa 100644 --- a/kernel/bpf/dispatcher.c +++ b/kernel/bpf/dispatcher.c @@ -4,6 +4,7 @@ #include #include #include +#include /* The BPF dispatcher is a multiway branch code generator. The * dispatcher is a mechanism to avoid the performance penalty of an @@ -90,6 +91,11 @@ int __weak arch_prepare_bpf_dispatcher(void *image, void *buf, s64 *funcs, int n return -ENOTSUPP; } +int __weak __init bpf_arch_init_dispatcher_early(void *ip) +{ + return -ENOTSUPP; +} + static int bpf_dispatcher_prepare(struct bpf_dispatcher *d, void *image, void *buf) { s64 ips[BPF_DISPATCHER_MAX] = {}, *ipsp = &ips[0]; -- cgit v1.2.3 From 3d05818707bb2cf133ccdcd29f2d5585b5bc1298 Mon Sep 17 00:00:00 2001 From: Hou Tao Date: Fri, 21 Oct 2022 19:49:12 +0800 Subject: bpf: Wait for busy refill_work when destroying bpf memory allocator A busy irq work is an unfinished irq work and it can be either in the pending state or in the running state. When destroying bpf memory allocator, refill_work may be busy for PREEMPT_RT kernel in which irq work is invoked in a per-CPU RT-kthread. It is also possible for kernel with arch_irq_work_has_interrupt() being false (e.g. 1-cpu arm32 host or mips) and irq work is inovked in timer interrupt. The busy refill_work leads to various issues. The obvious one is that there will be concurrent operations on free_by_rcu and free_list between irq work and memory draining. Another one is call_rcu_in_progress will not be reliable for the checking of pending RCU callback because do_call_rcu() may have not been invoked by irq work yet. The other is there will be use-after-free if irq work is freed before the callback of irq work is invoked as shown below: BUG: kernel NULL pointer dereference, address: 0000000000000000 #PF: supervisor instruction fetch in kernel mode #PF: error_code(0x0010) - not-present page PGD 12ab94067 P4D 12ab94067 PUD 1796b4067 PMD 0 Oops: 0010 [#1] PREEMPT_RT SMP CPU: 5 PID: 64 Comm: irq_work/5 Not tainted 6.0.0-rt11+ #1 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996) RIP: 0010:0x0 Code: Unable to access opcode bytes at 0xffffffffffffffd6. RSP: 0018:ffffadc080293e78 EFLAGS: 00010286 RAX: 0000000000000000 RBX: ffffcdc07fb6a388 RCX: ffffa05000a2e000 RDX: ffffa05000a2e000 RSI: ffffffff96cc9827 RDI: ffffcdc07fb6a388 ...... Call Trace: irq_work_single+0x24/0x60 irq_work_run_list+0x24/0x30 run_irq_workd+0x23/0x30 smpboot_thread_fn+0x203/0x300 kthread+0x126/0x150 ret_from_fork+0x1f/0x30 Considering the ease of concurrency handling, no overhead for irq_work_sync() under non-PREEMPT_RT kernel and has-irq-work-interrupt kernel and the short wait time used for irq_work_sync() under PREEMPT_RT (When running two test_maps on PREEMPT_RT kernel and 72-cpus host, the max wait time is about 8ms and the 99th percentile is 10us), just using irq_work_sync() to wait for busy refill_work to complete before memory draining and memory freeing. Fixes: 7c8199e24fa0 ("bpf: Introduce any context BPF specific memory allocator.") Acked-by: Stanislav Fomichev Signed-off-by: Hou Tao Link: https://lore.kernel.org/r/20221021114913.60508-2-houtao@huaweicloud.com Signed-off-by: Alexei Starovoitov --- kernel/bpf/memalloc.c | 11 +++++++++++ 1 file changed, 11 insertions(+) (limited to 'kernel') diff --git a/kernel/bpf/memalloc.c b/kernel/bpf/memalloc.c index 5f83be1d2018..1cb24323826f 100644 --- a/kernel/bpf/memalloc.c +++ b/kernel/bpf/memalloc.c @@ -493,6 +493,16 @@ void bpf_mem_alloc_destroy(struct bpf_mem_alloc *ma) rcu_in_progress = 0; for_each_possible_cpu(cpu) { c = per_cpu_ptr(ma->cache, cpu); + /* + * refill_work may be unfinished for PREEMPT_RT kernel + * in which irq work is invoked in a per-CPU RT thread. + * It is also possible for kernel with + * arch_irq_work_has_interrupt() being false and irq + * work is invoked in timer interrupt. So waiting for + * the completion of irq work to ease the handling of + * concurrency. + */ + irq_work_sync(&c->refill_work); drain_mem_cache(c); rcu_in_progress += atomic_read(&c->call_rcu_in_progress); } @@ -507,6 +517,7 @@ void bpf_mem_alloc_destroy(struct bpf_mem_alloc *ma) cc = per_cpu_ptr(ma->caches, cpu); for (i = 0; i < NUM_CACHES; i++) { c = &cc->cache[i]; + irq_work_sync(&c->refill_work); drain_mem_cache(c); rcu_in_progress += atomic_read(&c->call_rcu_in_progress); } -- cgit v1.2.3 From fa4447cb73b2bfe7175f1b7ffdc70580fcfcb991 Mon Sep 17 00:00:00 2001 From: Hou Tao Date: Fri, 21 Oct 2022 19:49:13 +0800 Subject: bpf: Use __llist_del_all() whenever possbile during memory draining Except for waiting_for_gp list, there are no concurrent operations on free_by_rcu, free_llist and free_llist_extra lists, so use __llist_del_all() instead of llist_del_all(). waiting_for_gp list can be deleted by RCU callback concurrently, so still use llist_del_all(). Acked-by: Stanislav Fomichev Signed-off-by: Hou Tao Link: https://lore.kernel.org/r/20221021114913.60508-3-houtao@huaweicloud.com Signed-off-by: Alexei Starovoitov --- kernel/bpf/memalloc.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/kernel/bpf/memalloc.c b/kernel/bpf/memalloc.c index 1cb24323826f..4901fa1048cd 100644 --- a/kernel/bpf/memalloc.c +++ b/kernel/bpf/memalloc.c @@ -418,14 +418,17 @@ static void drain_mem_cache(struct bpf_mem_cache *c) /* No progs are using this bpf_mem_cache, but htab_map_free() called * bpf_mem_cache_free() for all remaining elements and they can be in * free_by_rcu or in waiting_for_gp lists, so drain those lists now. + * + * Except for waiting_for_gp list, there are no concurrent operations + * on these lists, so it is safe to use __llist_del_all(). */ llist_for_each_safe(llnode, t, __llist_del_all(&c->free_by_rcu)) free_one(c, llnode); llist_for_each_safe(llnode, t, llist_del_all(&c->waiting_for_gp)) free_one(c, llnode); - llist_for_each_safe(llnode, t, llist_del_all(&c->free_llist)) + llist_for_each_safe(llnode, t, __llist_del_all(&c->free_llist)) free_one(c, llnode); - llist_for_each_safe(llnode, t, llist_del_all(&c->free_llist_extra)) + llist_for_each_safe(llnode, t, __llist_del_all(&c->free_llist_extra)) free_one(c, llnode); } -- cgit v1.2.3