[REGRESSION][4.0-rc1] My i386 fails to do CPU hotplug

[Steven Rostedt]

Running my test suite on 4.0-rc1, my i386 run of my tests crashed with
the following:

Initializing CPU#1
invalid opcode: 0000 [#1] SMP
Modules linked in: ip6t_REJECT nf_reject_ipv6 nf_conntrack_ipv6 nf_defrag_ipv6 ip6table_filter ip6_tables ipv6 ppdev parport_pc microcode r8169 parport
CPU: 1 PID: 0 Comm: swapper/1 Not tainted 4.0.0-rc1-test #2
Hardware name: MSI MS-7823/CSM-H87M-G43 (MS-7823), BIOS V1.6 02/22/2014
task: f2ee8000 ti: f2ee6000 task.ti: f2ee6000
EIP: 0060:[<c040ab1f>] EFLAGS: 00210046 CPU: 1
EIP is at xstate_enable+0x2b/0x30
EAX: 00000007 EBX: 00000001 ECX: 00000000 EDX: 00000000
ESI: f2ee8000 EDI: c144c930 EBP: f2ee7f58 ESP: f2ee7f58
 DS: 007b ES: 007b FS: 00d8 GS: 0000 SS: 0068
CR0: 80050033 CR2: 00000000 CR3: 01139000 CR4: 000000b0
Stack:
 f2ee7f60 c040b47b f2ee7f6c c040a4fd f352ef00 f2ee7f94 c040f40d c0fd8c4c
 00000001 320f4000 f2ee7fec 00000000 02100800 00000000 00000000 f2ee7fb4
 c0425f1a 00000000 00000000 00000000 00000000 00000001 00200002 00000000
Call Trace:
 [<c040b47b>] xsave_init+0x25/0x27
 [<c040a4fd>] fpu_init+0xb3/0xbb
 [<c040f40d>] cpu_init+0x20e/0x225
 [<c0425f1a>] start_secondary+0xe/0x162
Code: 64 8b 15 88 d0 43 c1 89 d0 0d 00 00 04 00 55 39 d0 89 e5 74 09 64 a3 88 d0 43 c1 0f 22 e0 8b 15 3c c9 44 c1 31 c9 a1 38 c9 44 c1 <0f> 01 d1 5d c3 55 89 e5 57 56 53 3e 8d 74 26 00 83 cb ff 89 d7
EIP: [<c040ab1f>] xstate_enable+0x2b/0x30 SS:ESP 0068:f2ee7f58
---[ end trace 9afa73ad583edc40 ]---
Kernel panic - not syncing: Attempted to kill the idle task!


Interesting that the fault happened at "<0f> 01 d1" and that matches:

static inline void xsetbv(u32 index, u64 value)
{
	u32 eax = value;
	u32 edx = value >> 32;

	asm volatile(".byte 0x0f,0x01,0xd1" /* xsetbv */
		     : : "a" (eax), "d" (edx), "c" (index));
}


Doing a bisect, it ended on this commit:

commit 1e02ce4cccdcb9688386e5b8d2c9fa4660b45389
Author: Andy Lutomirski <luto@xxxxxxxxxxxxxx>
Date:   Fri Oct 24 15:58:08 2014 -0700

    x86: Store a per-cpu shadow copy of CR4


Reverting it fixed the regression, and my box can happily hotplug again!

-- Steve
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/