Re: [PATCH v4 00/45] x86: Kernel IBT

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Sat, Mar 12, 2022 at 7:44 AM Peter Zijlstra <peterz@xxxxxxxxxxxxx> wrote:
>
> On Fri, Mar 11, 2022 at 09:09:38AM -0800, Alexei Starovoitov wrote:
> > On Fri, Mar 11, 2022 at 2:40 AM Peter Zijlstra <peterz@xxxxxxxxxxxxx> wrote:
> > >
> > > On Thu, Mar 10, 2022 at 05:29:11PM +0100, Peter Zijlstra wrote:
> > >
> > > > This seems to cure most of the rest. I'm still seeing one failure:
> > > >
> > > > libbpf: prog 'connect_v4_prog': BPF program load failed: Invalid argument
> > > > libbpf: failed to load program 'connect_v4_prog'
> > > > libbpf: failed to load object './connect4_prog.o'
> > > > test_fexit_bpf2bpf_common:FAIL:tgt_prog_load unexpected error: -22 (errno 22)
> > > > #48/4 fexit_bpf2bpf/func_replace_verify:FAIL
> > >
> > >
> > > Hmm, with those two patches on I get:
> > >
> > > root@tigerlake:/usr/src/linux-2.6/tgl-build# ./test_progs -t fexit
> > > #46 fentry_fexit:OK
> > > #48 fexit_bpf2bpf:OK
> > > #49 fexit_sleep:OK
> > > #50 fexit_stress:OK
> > > #51 fexit_test:OK
> > > Summary: 5/9 PASSED, 0 SKIPPED, 0 FAILED
> > >
> > > On the tigerlake, I suppose I'm doing something wrong on the other
> > > machine because there it's even failing on the pre-ibt kernel image.
> > >
> > > I'll go write up changelogs and stick these on.
> >
> > What is the latest branch I can use to test it?
>
>   git://git.kernel.org/pub/scm/linux/kernel/git/peterz/queue.git x86/ibt
>
> that also include bpf-next. Thanks!

Looks better.
During the build with gcc 8.5 I see:

arch/x86/crypto/crc32c-intel.o: warning: objtool: file already has
.ibt_endbr_seal, skipping
arch/x86/crypto/crc32c-intel.o: warning: objtool: file already has
.orc_unwind section, skipping
  LD [M]  crypto/async_tx/async_xor.ko
  LD [M]  crypto/authenc.ko
make[3]: *** [../scripts/Makefile.modfinal:61:
arch/x86/crypto/crc32c-intel.ko] Error 255
make[3]: *** Waiting for unfinished jobs....

but make clean cures it.
I suspect it's some missing makefile dependency.

and:
vmlinux.o: warning: objtool: ksys_unshare()+0x626: unreachable instruction
which stays even after make clean.

The rcu "false positive" is still there that causes
sporadic hangs during the boot.

The test_progs shows:
Summary: 228/1122 PASSED, 4 SKIPPED, 6 FAILED
(when I remove one test)

That test is actually crashing the kernel:
./test_progs -t mod_race
[   39.202593] bpf_testmod: loading out-of-tree module taints kernel.
[   39.303142] general protection fault, probably for non-canonical
address 0xdffffc0000000000: 0000 [#1] PREEMPT SMP KASAN PTI
[   39.304610] KASAN: null-ptr-deref in range
[0x0000000000000000-0x0000000000000007]
[   39.305514] CPU: 9 PID: 1599 Comm: test_progs Tainted: G
O      5.17.0-rc7-02525-g5dd5efb53cf1 #4
[   39.306675] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996),
BIOS rel-1.12.0-59-gc9ba5276e321-prebuilt.qemu.org 04/01/2014
[   39.308036] RIP: 0010:do_init_module+0x9/0x6f0
[   39.308583] Code: fe ff ff e8 59 13 46 00 e9 7f fe ff ff e8 4f 13
46 00 e9 49 fe ff ff 66 2e 0f 1f 84 00 00 00 00 00 e8 cb 8d eb 1e 48
b8 00 00 <00> 00 00 fc ff df 41 57 49 89 ff 48 c7 c7 20 f6 5c 84 48 89
fa 41
[   39.310815] RSP: 0018:ffff88810f7e7aa0 EFLAGS: 00010282
[   39.311450] RAX: dffffc0000000000 RBX: 0000000000000009 RCX: ffffffff81283b16
[   39.312253] RDX: 0000000000000000 RSI: 0000000000000008 RDI: ffffffffa0224c00
[   39.313031] RBP: ffff88810f7e7ac8 R08: 0000000000000000 R09: fffffbfff0b4d557
[   39.313813] R10: ffffffff85a6aab7 R11: fffffbfff0b4d556 R12: ffff88811171f518
[   39.314591] R13: dffffc0000000000 R14: ffffffffa0224c00 R15: ffff88810f7e7e50
[   39.315374] FS:  00007f8e1b981700(0000) GS:ffff8881f6a80000(0000)
knlGS:0000000000000000
[   39.316293] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[   39.316984] CR2: 00007fdf39350ff0 CR3: 000000011952e006 CR4: 00000000003706e0
[   39.317860] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[   39.318680] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
[   39.319467] Call Trace:
[   39.319744]  <TASK>
[   39.319982]  bpf_trampoline_6442471603_0+0x32/0x1000
[   39.320537]  do_init_module+0x5/0x6f0
[   39.320945]  load_module+0x77c0/0x9c00
[   39.321376]  ? module_frob_arch_sections+0x20/0x20
[   39.321892]  ? ima_post_read_file+0x161/0x180
[   39.322392]  ? ima_read_file+0x140/0x140
[   39.322827]  ? security_kernel_post_read_file+0x55/0xb0
[   39.323406]  ? __x64_sys_fsconfig+0x630/0x630
[   39.323889]  ? fput_many+0x1e/0x120
[   39.324285]  ? __do_sys_finit_module+0xf3/0x150
[   39.324822]  __do_sys_finit_module+0xf3/0x150
[   39.325311]  ? __ia32_sys_init_module+0xb0/0xb0
[   39.325826]  ? rcu_read_lock_held_common+0xe/0xa0
[   39.326349]  ? rcu_read_lock_sched_held+0x5a/0xc0
[   39.326869]  ? rcu_read_lock_bh_held+0xa0/0xa0
[   39.327362]  ? file_open_root+0x1f0/0x1f0
[   39.327812]  ? syscall_trace_enter.isra.17+0x184/0x250
[   39.328411]  do_syscall_64+0x38/0x80
[   39.328812]  entry_SYSCALL_64_after_hwframe+0x44/0xae

The test was designed to check whether the kernel bug is fixed.
If not it would crash the kernel.

Kumar,
you've added that test.
Could you please take a look at why it is crashing in Peter's tree?



[Index of Archives]     [Linux Samsung SoC]     [Linux Rockchip SoC]     [Linux Actions SoC]     [Linux for Synopsys ARC Processors]     [Linux NFS]     [Linux NILFS]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]


  Powered by Linux