On 8/22/2022 10:51 AM, Paul Moore wrote: > On Mon, Aug 22, 2022 at 1:35 PM Casey Schaufler <casey@xxxxxxxxxxxxxxxx> wrote: >> I see the following BUG reported when I run the SELinux test suite on >> 6.0-rc1 and 6.0-rc2. Is this a known issue? I'm running on Fedora 36 >> under KVM. There's nothing fancy in my config, it's lightly tweaked >> generic with the test suite modifications applied. I've attached the >> config just in case. >> >> >> <4>[ 471.300818] ------------[ cut here ]------------ >> <2>[ 471.300822] kernel BUG at include/linux/mmap_lock.h:156! >> <4>[ 471.300842] invalid opcode: 0000 [#1] PREEMPT SMP PTI >> <4>[ 471.300850] CPU: 6 PID: 4666 Comm: service_provide Not tainted 6.0.0-rc2stack+ #1380 >> <4>[ 471.300856] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.16.0-1.fc36 04/01/2014 >> <4>[ 471.300860] RIP: 0010:find_vma+0x34/0x80 >> <4>[ 471.300869] Code: 48 83 ec 08 48 8b 47 78 48 85 c0 74 18 48 89 f3 e8 f1 85 fe ff 48 85 c0 74 12 48 83 c4 08 5b 5d c3 cc cc cc cc e8 e2 5b a4 00 <0f> 0b 48 8b 55 08 48 85 d2 75 15 eb e3 48 8d 42 e0 48 3b 5a e0 73 >> <4>[ 471.300875] RSP: 0018:ffffb5f041cefa18 EFLAGS: 00010292 >> <4>[ 471.300881] RAX: 000000000000033f RBX: ffff8fa55ce76810 RCX: 0000000000000000 >> <4>[ 471.300886] RDX: 0000000000000001 RSI: ffffffff9e68680f RDI: 00000000ffffffff >> <4>[ 471.300889] RBP: ffff8fa64175f700 R08: 0000000000000000 R09: ffffb5f041cef798 >> <4>[ 471.300893] R10: 0000000000000003 R11: ffffffff9eb44b08 R12: 0000000000000008 >> <4>[ 471.300896] R13: 0000000000000000 R14: ffff8fa64bbc99a8 R15: 00007f61f5cc2000 >> <4>[ 471.300900] FS: 00007fa9a3afd800(0000) GS:ffff8fa6bbb80000(0000) knlGS:0000000000000000 >> <4>[ 471.300905] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 >> <4>[ 471.300909] CR2: 00007fa9a3c2eb50 CR3: 0000000059488004 CR4: 0000000000370ee0 >> <4>[ 471.300918] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 >> <4>[ 471.300922] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 >> <4>[ 471.300926] Call Trace: >> <4>[ 471.300929] <TASK> >> <4>[ 471.300933] binder_alloc_new_buf+0x4a/0x850 >> <4>[ 471.300947] binder_transaction+0x553/0x3120 > Yes, there is a problem in the binder code. OK, thanks. I'll proceed without the binder test. > The good news is that a > fix has been identified and a patch is working its way upstream. I've > tested a prior version of the linked patch below with good effect. > Unfortunately I was still seeing some XFS problems in the test suite > last time I ran it which I haven't yet had time to investigate, I'm > not sure if Ondrej is seeing something similar with the XFS portions > of the test suite. > > https://lore.kernel.org/lkml/CAFqZXNuOvt_p0ZNjh89Vt9sGQ=CE5_GLr+XOT6OkzxcHdEHgKQ@xxxxxxxxxxxxxx/T/#m49642734166118c4666b0d5f84f4e9f96aeda12b >
