Unexplained behaviour of RHEL systems with Broadwell/Haswell CPUs

Issue

• System becomes sluggish with no load for a while.
• System crashes with General Protection Fault.
• Kernel crashes with below logs:

BUG: unable to handle kernel NULL pointer dereference at 0000000000000008
IP: [<ffffffffc027a5ef>] xlog_cil_push+0x18f/0x430 [xfs]
PGD 0
Oops: 0000 [#1] SMP
CPU: 5 PID: 32756 Comm: kworker/5:2 Kdump: loaded Tainted: G        W      ------------ T 3.10.0-1062.1.1.el7.x86_64 #1
Hardware name: HP ProLiant DL380 Gen9/ProLiant DL380 Gen9, BIOS P89 01/22/2018
Workqueue: xfs-cil/dm-19 xlog_cil_push_work [xfs]
task: ffff9baa63f8a0e0 ti: ffff9bab33b4c000 task.ti: ffff9bab33b4c000
RIP: 0010:[<ffffffffc027a5ef>]  [<ffffffffc027a5ef>] xlog_cil_push+0x18f/0x430 [xfs]
Call Trace:
 [<ffffffff918ae168>] ? add_timer+0x18/0x20
 [<ffffffff918bb3db>] ? __queue_delayed_work+0x8b/0x1a0
 [<ffffffffc027a8a5>] xlog_cil_push_work+0x15/0x20 [xfs]
 [<ffffffff918bd0ff>] process_one_work+0x17f/0x440
 [<ffffffff918be368>] worker_thread+0x278/0x3c0
 [<ffffffff918be0f0>] ? manage_workers.isra.26+0x2a0/0x2a0
 [<ffffffff918c50d1>] kthread+0xd1/0xe0
 [<ffffffff918c5000>] ? insert_kthread_work+0x40/0x40
 [<ffffffff91f8cd37>] ret_from_fork_nospec_begin+0x21/0x21
 [<ffffffff918c5000>] ? insert_kthread_work+0x40/0x40
Code: 46 08 48 39 85 58 ff ff ff 74 61 45 31 ff eb 28 0f 1f 40 00 48 8b 70 10 49 89 37 4c 8b 78 10 48 c7 40 10 00 00 00 00 49 8b 46 08 <45> 03 6f 08 48 39 85 58 ff ff ff 74 34 48 89 c7 48 89 85 68 ff
RIP  [<ffffffffc027a5ef>] xlog_cil_push+0x18f/0x430 [xfs]
 RSP <ffff9bab33b4fd48>
CR2: 0000000000000008

[32386.065177] ------------[ cut here ]------------
[32386.070257] kernel BUG at fs/jbd2/journal.c:2482!
[32386.075426] invalid opcode: 0000 [#1] SMP 
[32386.226624] CPU: 31 PID: 7287 Comm: jbd2/dm-96-8 Not tainted 3.10.0-514.el7.x86_64 #1
[32386.235232] Hardware name: HP Superdome2 16s x86, BIOS Bundle: 008.004.084 SFW: 043.025.000 08/16/2016
[32386.245464] task: ffff887f7381edd0 ti: ffff887f7f284000 task.ti: ffff887f7f284000
[32386.253691] RIP: 0010:[<ffffffffa071ace2>]  [<ffffffffa071ace2>] jbd2_journal_put_journal_head+0x142/0x146 [jbd2]
[32386.348816] Stack:
[32386.351025]  ffff881911ea9400 ffff885f7ba34000 ffff887f7f287ca0 ffffffffa0713dbb
[32386.359192]  ffffffff811899e0 ffff885f7ba343a0 ffff881ff6b5ca92 ffff88de9820ce38
[32386.367365]  ffff88dea9015000 ffff881fe8d67600 ffff887f7f287e40 ffffffffa07120f6
[32386.375535] Call Trace:
[32386.378257]  [<ffffffffa0713dbb>] __jbd2_journal_remove_checkpoint+0x5b/0x160 [jbd2]
[32386.386782]  [<ffffffff811899e0>] ? free_pages.part.80+0x40/0x50
[32386.393389]  [<ffffffffa07120f6>] jbd2_journal_commit_transaction+0x1106/0x19a0 [jbd2]
[32386.402100]  [<ffffffff81029569>] ? __switch_to+0xd9/0x4c0
[32386.408131]  [<ffffffffa0716e99>] kjournald2+0xc9/0x260 [jbd2]
[32386.414551]  [<ffffffff810b1600>] ? wake_up_atomic_t+0x30/0x30
[32386.420967]  [<ffffffffa0716dd0>] ? commit_timeout+0x10/0x10 [jbd2]
[32386.427858]  [<ffffffff810b052f>] kthread+0xcf/0xe0
[32386.433218]  [<ffffffff810b0460>] ? kthread_create_on_node+0x140/0x140
[32386.440402]  [<ffffffff81696418>] ret_from_fork+0x58/0x90
[32386.446335]  [<ffffffff810b0460>] ? kthread_create_on_node+0x140/0x140
[32386.453509] Code: c7 c6 80 b5 71 a0 48 c7 c7 e8 d8 71 a0 31 c0 e8 2b 47 f6 e0 48 8b 73 20 49 8b 7c 24 20 e8 f7 f8 ff ff e9 64 ff ff ff 0f 0b 0f 0b <0f> 0b 0f 0b 55 48 89 e5 0f 0b 55 48 89 e5 0f 0b 0f 1f 44 00 00 
[32386.474936] RIP  [<ffffffffa071ace2>] jbd2_journal_put_journal_head+0x142/0x146 [jbd2]
[32386.483653]  RSP <ffff887f7f287c50>