Stack Allocation Sampling #1
Assignees
Labels
Comments
hgn
pushed a commit
that referenced
this issue
Sep 1, 2022
Commit c89191c ("x86/entry: Convert SWAPGS to swapgs and remove the definition of SWAPGS") missed one use case of SWAPGS in entry_INT80_compat(). Removing of the SWAPGS macro led to asm just using "swapgs", as it is accepting instructions in capital letters, too. This in turn leads to splats in Xen PV guests like: [ 36.145223] general protection fault, maybe for address 0x2d: 0000 [#1] PREEMPT SMP NOPTI [ 36.145794] CPU: 2 PID: 1847 Comm: ld-linux.so.2 Not tainted 5.19.1-1-default #1 \ openSUSE Tumbleweed f3b44bfb672cdb9f235aff53b57724eba8b9411b [ 36.146608] Hardware name: HP ProLiant ML350p Gen8, BIOS P72 11/14/2013 [ 36.148126] RIP: e030:entry_INT80_compat+0x3/0xa3 Fix that by open coding this single instance of the SWAPGS macro. Fixes: c89191c ("x86/entry: Convert SWAPGS to swapgs and remove the definition of SWAPGS") Signed-off-by: Juergen Gross <[email protected]> Signed-off-by: Borislav Petkov <[email protected]> Reviewed-by: Jan Beulich <[email protected]> Cc: <[email protected]> # 5.19 Link: https://lore.kernel.org/r/[email protected]
hgn
pushed a commit
that referenced
this issue
Sep 1, 2022
Syzkaller reported a triggered kernel BUG as follows: ------------[ cut here ]------------ kernel BUG at kernel/bpf/cgroup.c:925! invalid opcode: 0000 [#1] PREEMPT SMP NOPTI CPU: 1 PID: 194 Comm: detach Not tainted 5.19.0-14184-g69dac8e431af #8 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.16.0-0-gd239552ce722-prebuilt.qemu.org 04/01/2014 RIP: 0010:__cgroup_bpf_detach+0x1f2/0x2a0 Code: 00 e8 92 60 30 00 84 c0 75 d8 4c 89 e0 31 f6 85 f6 74 19 42 f6 84 28 48 05 00 00 02 75 0e 48 8b 80 c0 00 00 00 48 85 c0 75 e5 <0f> 0b 48 8b 0c5 RSP: 0018:ffffc9000055bdb0 EFLAGS: 00000246 RAX: 0000000000000000 RBX: ffff888100ec0800 RCX: ffffc900000f1000 RDX: 0000000000000000 RSI: 0000000000000001 RDI: ffff888100ec4578 RBP: 0000000000000000 R08: ffff888100ec0800 R09: 0000000000000040 R10: 0000000000000000 R11: 0000000000000000 R12: ffff888100ec4000 R13: 000000000000000d R14: ffffc90000199000 R15: ffff888100effb00 FS: 00007f68213d2b80(0000) GS:ffff88813bc80000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 000055f74a0e5850 CR3: 0000000102836000 CR4: 00000000000006e0 Call Trace: <TASK> cgroup_bpf_prog_detach+0xcc/0x100 __sys_bpf+0x2273/0x2a00 __x64_sys_bpf+0x17/0x20 do_syscall_64+0x3b/0x90 entry_SYSCALL_64_after_hwframe+0x63/0xcd RIP: 0033:0x7f68214dbcb9 Code: 08 44 89 e0 5b 41 5c c3 66 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 ff8 RSP: 002b:00007ffeb487db68 EFLAGS: 00000246 ORIG_RAX: 0000000000000141 RAX: ffffffffffffffda RBX: 000000000000000b RCX: 00007f68214dbcb9 RDX: 0000000000000090 RSI: 00007ffeb487db70 RDI: 0000000000000009 RBP: 0000000000000003 R08: 0000000000000012 R09: 0000000b00000003 R10: 00007ffeb487db70 R11: 0000000000000246 R12: 00007ffeb487dc20 R13: 0000000000000004 R14: 0000000000000001 R15: 000055f74a1011b0 </TASK> Modules linked in: ---[ end trace 0000000000000000 ]--- Repetition steps: For the following cgroup tree, root | cg1 | cg2 1. attach prog2 to cg2, and then attach prog1 to cg1, both bpf progs attach type is NONE or OVERRIDE. 2. write 1 to /proc/thread-self/fail-nth for failslab. 3. detach prog1 for cg1, and then kernel BUG occur. Failslab injection will cause kmalloc fail and fall back to purge_effective_progs. The problem is that cg2 have attached another prog, so when go through cg2 layer, iteration will add pos to 1, and subsequent operations will be skipped by the following condition, and cg will meet NULL in the end. `if (pos && !(cg->bpf.flags[atype] & BPF_F_ALLOW_MULTI))` The NULL cg means no link or prog match, this is as expected, and it's not a bug. So here just skip the no match situation. Fixes: 4c46091 ("bpf: Fix KASAN use-after-free Read in compute_effective_progs") Signed-off-by: Pu Lehui <[email protected]> Signed-off-by: Daniel Borkmann <[email protected]> Acked-by: Andrii Nakryiko <[email protected]> Link: https://lore.kernel.org/bpf/[email protected]
hgn
pushed a commit
that referenced
this issue
Sep 1, 2022
Hsin-Wei reported a KASAN splat triggered by their BPF runtime fuzzer which is based on a customized syzkaller: BUG: KASAN: slab-out-of-bounds in bpf_int_jit_compile+0x1257/0x13f0 Read of size 8 at addr ffff888004e90b58 by task syz-executor.0/1489 CPU: 1 PID: 1489 Comm: syz-executor.0 Not tainted 5.19.0 #1 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.13.0-1ubuntu1.1 04/01/2014 Call Trace: <TASK> dump_stack_lvl+0x9c/0xc9 print_address_description.constprop.0+0x1f/0x1f0 ? bpf_int_jit_compile+0x1257/0x13f0 kasan_report.cold+0xeb/0x197 ? kvmalloc_node+0x170/0x200 ? bpf_int_jit_compile+0x1257/0x13f0 bpf_int_jit_compile+0x1257/0x13f0 ? arch_prepare_bpf_dispatcher+0xd0/0xd0 ? rcu_read_lock_sched_held+0x43/0x70 bpf_prog_select_runtime+0x3e8/0x640 ? bpf_obj_name_cpy+0x149/0x1b0 bpf_prog_load+0x102f/0x2220 ? __bpf_prog_put.constprop.0+0x220/0x220 ? find_held_lock+0x2c/0x110 ? __might_fault+0xd6/0x180 ? lock_downgrade+0x6e0/0x6e0 ? lock_is_held_type+0xa6/0x120 ? __might_fault+0x147/0x180 __sys_bpf+0x137b/0x6070 ? bpf_perf_link_attach+0x530/0x530 ? new_sync_read+0x600/0x600 ? __fget_files+0x255/0x450 ? lock_downgrade+0x6e0/0x6e0 ? fput+0x30/0x1a0 ? ksys_write+0x1a8/0x260 __x64_sys_bpf+0x7a/0xc0 ? syscall_enter_from_user_mode+0x21/0x70 do_syscall_64+0x3b/0x90 entry_SYSCALL_64_after_hwframe+0x63/0xcd RIP: 0033:0x7f917c4e2c2d The problem here is that a range of tnum_range(0, map->max_entries - 1) has limited ability to represent the concrete tight range with the tnum as the set of resulting states from value + mask can result in a superset of the actual intended range, and as such a tnum_in(range, reg->var_off) check may yield true when it shouldn't, for example tnum_range(0, 2) would result in 00XX -> v = 0000, m = 0011 such that the intended set of {0, 1, 2} is here represented by a less precise superset of {0, 1, 2, 3}. As the register is known const scalar, really just use the concrete reg->var_off.value for the upper index check. Fixes: d2e4c1e ("bpf: Constant map key tracking for prog array pokes") Reported-by: Hsin-Wei Hung <[email protected]> Signed-off-by: Daniel Borkmann <[email protected]> Cc: Shung-Hsi Yu <[email protected]> Acked-by: John Fastabend <[email protected]> Link: https://lore.kernel.org/r/984b37f9fdf7ac36831d2137415a4a915744c1b6.1661462653.git.daniel@iogearbox.net Signed-off-by: Alexei Starovoitov <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Aktuell ist es schwierig im Linux die aktuelle Stackverwendung zu analysieren. Über /proc/[pid]/maps kann zwar festgestellt werden, welche Speicherseiten gemapped wurden, was eine granularität von (maximal) 4096 byte erbringt.
Eine Perf erweiterung wäre es beim samplen nicht nur %RIP Register auszulesen, sondern auch den aktuellen Stackpointer (%ESP bei x96-64 oder %SP bei Arm/RISC-V).
Problem: sampeln beginnt nicht, nachdem das program sofort startete, Nutzer können sich ja zu einen bereits (sehr) lange laufende Applikation "attachen". Wenn nun bei jeden auslesen der %ESP genommen wird - welche Aussage könnte dann getroffen werden. Hier wäre vielmehr die Frage: kommt man an die Stack Start Adresse heran? Dies könnte man auf der Maillingliste erfragen. Bei unwindung muss es ja auch eine Abbruchbedinung geben, wann stoppt man beispielsweise beim unwinding?
Evtl kann man hier schauen: https://github.com/d99kris/stackusage.git - was wird hier als Referenz verwendet? Wie wird hier das Delta berechnet?
The text was updated successfully, but these errors were encountered: