I observe that the kernel crashes with OOM, the kmalloc-192 cache is
ridiculously large:
> [ 132.409828] kmalloc-192 15750374KB 15750374KB
The stack looks like this, although I'm unsure if this was actually the thread
that allocated so much memory:
[ 132.508968] Kernel panic - not syncing: System is deadlocked on memory
[ 132.515495] CPU: 1 PID: 4444 Comm: iscsi_trx Not tainted
5.3.18-lp152.60-default #1 openSUSE Leap 15.2
[ 132.524799] Hardware name: FUJITSU PRIMERGY TX100 S3/D3009-A1, BIOS V4.6.4.1
R1.10.0 for D3009-A1x 03/26/2013
[ 132.534712] Call Trace:
[ 132.537157] dump_stack+0x66/0x8b
[ 132.540469] panic+0xfe/0x2d7
[ 132.543434] out_of_memory+0x501/0x510
[ 132.547180] __alloc_pages_slowpath+0x9c9/0xd00
[ 132.551711] __alloc_pages_nodemask+0x2c4/0x300
[ 132.556241] new_slab+0x27a/0xaa0
[ 132.559552] ___slab_alloc+0x3d9/0x5b0
[ 132.563301] ? bio_alloc_bioset+0x1e4/0x240
[ 132.567485] ? bio_alloc_bioset+0x1e4/0x240
[ 132.571669] ? __slab_alloc+0x1c/0x30
[ 132.575329] __slab_alloc+0x1c/0x30
[ 132.578816] __kmalloc+0x285/0x2c0
[ 132.582218] bio_alloc_bioset+0x1e4/0x240
[ 132.586227] pscsi_execute_cmd+0x374/0x6e0 [target_core_pscsi]
[ 132.592066] __target_execute_cmd+0x25/0xb0 [target_core_mod]
[ 132.597816] transport_generic_new_cmd+0x15f/0x310 [target_core_mod]
[ 132.604172] transport_handle_cdb_direct+0x34/0x80 [target_core_mod]
[ 132.610528] iscsit_sequence_cmd+0xce/0x1a0 [iscsi_target_mod]
[ 132.616365] iscsit_process_scsi_cmd+0x20/0xb0 [iscsi_target_mod]
[ 132.622461] iscsit_get_rx_pdu+0x684/0x10b0 [iscsi_target_mod]
[ 132.628294] ? wait_for_completion_interruptible+0xbd/0x190
[ 132.633868] ? iscsi_target_tx_thread+0x1f0/0x1f0 [iscsi_target_mod]
[ 132.640224] ? iscsi_target_rx_thread+0xb8/0xf0 [iscsi_target_mod]
[ 132.646406] iscsi_target_rx_thread+0xb8/0xf0 [iscsi_target_mod]
[ 132.652410] kthread+0x10d/0x130
[ 132.655635] ? kthread_park+0xa0/0xa0
[ 132.659295] ret_from_fork+0x1f/0x40
[ 132.662874] Kernel Offset: 0x34c00000 from 0xffffffff81000000 (relocation
range: 0xffffffff80000000-0xffffffffbfffffff)