lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-ID: <20190919102643.GA9400@dingwall.me.uk>
Date:   Thu, 19 Sep 2019 10:26:43 +0000
From:   James Dingwall <james@...gwall.me.uk>
To:     linux-kernel@...r.kernel.org
Cc:     Kees Cook <keescook@...omium.org>,
        Anton Vorontsov <anton@...msg.org>,
        Colin Cross <ccross@...roid.com>,
        Tony Luck <tony.luck@...el.com>,
        Boris Ostrovsky <boris.ostrovsky@...cle.com>,
        Juergen Gross <jgross@...e.com>
Subject: pstore does not work under xen

Hi,

I have been investigating a regression in our environment where pstore 
(efi-pstore specifically but I suspect this would affect all 
implementations) no longer works after upgrading from a 4.4 to 5.0 
kernel when running under xen.  (This is an Ubuntu kernel but I don't 
think there are patches which affect this area.)

In kernel/panic.c the flow of panic() is roughly:

dump_stack();

atomic_notifier_call_chain(&panic_notifier_list, 0, buf);

kmsg_dump(KMSG_DUMP_PANIC);


pstore registers a kdump callback which would normally be invoked by the 
call to kmsg_dump() however in Xen there is a panic_notifier registered 
which never returns preventing the pstore record being generated.  The 
implementation of xen_panic_event() has changed since v4.4 but I don't 
understand how this may have changed the behaviour.

Adding a couple of printks in arch/x86/xen/enlighten.c:

@@ -277,8 +278,10 @@ void xen_emergency_restart(void)
 static int
 xen_panic_event(struct notifier_block *this, unsigned long event, void *ptr)
 {
+       printk(KERN_WARNING "enter xen_panic_event()\n");
        if (!kexec_crash_loaded())
                xen_reboot(SHUTDOWN_crash);
+       printk(KERN_WARNING "exit xen_panic_event()\n");
        return NOTIFY_DONE;
 }

Only the first is printed when triggering a crash (echo c > 
/proc/sysrq-trigger)

[ 1185.458761] sysrq: SysRq : Trigger a crash
[ 1185.476937] Kernel panic - not syncing: sysrq triggered crash
[ 1185.495747] CPU: 1 PID: 19241 Comm: bash Tainted: P           OE     5.0.0-27-generic #4
[ 1185.513387] Hardware name: HP ProLiant EC200a/ProLiant EC200a, BIOS U26 05/21/2018
[ 1185.530705] Call Trace:
[ 1185.548683]  dump_stack+0x63/0x85
[ 1185.566553]  panic+0xfe/0x2b4
[ 1185.583634]  sysrq_handle_crash+0x15/0x20
[ 1185.600594]  __handle_sysrq+0x9f/0x170
[ 1185.617613]  write_sysrq_trigger+0x34/0x40
[ 1185.634271]  proc_reg_write+0x3e/0x60
[ 1185.651407]  __vfs_write+0x1b/0x40
[ 1185.668140]  vfs_write+0xb1/0x1a0
[ 1185.685087]  ksys_write+0x5c/0xe0
[ 1185.701190]  __x64_sys_write+0x1a/0x20
[ 1185.717719]  do_syscall_64+0x5a/0x120
[ 1185.733839]  entry_SYSCALL_64_after_hwframe+0x44/0xa9
[ 1185.749739] RIP: 0033:0x7fe42e4f5154
[ 1185.764726] Code: 89 02 48 c7 c0 ff ff ff ff c3 66 2e 0f 1f 84 00 00 00 00 00 66 90 48 8d 05 b1 07 2e 00 8b 00 85 c0 75 13 b8 01 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 54 f3 c3 66 90 41 54 55 49 89 d4 53 48 89 f5
[ 1185.797937] RSP: 002b:00007fff092d0358 EFLAGS: 00000246 ORIG_RAX: 0000000000000001
[ 1185.814701] RAX: ffffffffffffffda RBX: 0000000000000002 RCX: 00007fe42e4f5154
[ 1185.831241] RDX: 0000000000000002 RSI: 000055c1c224bf10 RDI: 0000000000000001
[ 1185.848405] RBP: 000055c1c224bf10 R08: 000000000000000a R09: 0000000000000001
[ 1185.869816] R10: 000000000000000a R11: 0000000000000246 R12: 00007fe42e7d1760
[ 1185.888686] R13: 0000000000000002 R14: 00007fe42e7cd2a0 R15: 00007fe42e7cc760
[ 1185.905369] Kernel Offset: disabled
[ 1185.920295] enter xen_panic_event()
(XEN) Hardware Dom0 crashed: rebooting machine in 5 seconds.
(d1) Checking store ...xenstored: Checking store complete.xenstored: Checking store ...xenstored: Checking store complete.xenstored: Checking store ...xenstored: Checking store complete.xenstored: Checking store ...

Sorry for the long Cc list, I'm not sure who's court this falls in.

Thanks,
James

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ