Project

General

Profile

Actions

Bug #8878

open

pvscsi panic

Added by Igor Kozhukhov almost 5 years ago. Updated almost 5 years ago.

Status:
New
Priority:
Normal
Assignee:
-
Category:
-
Start date:
2017-11-30
Due date:
% Done:

0%

Estimated time:
Difficulty:
Medium
Tags:
needs-triage
Gerrit CR:
External Bug:

Description

VMware guest with 4GB RAM, 4vCpu
several zvol's with for iscsi

create zpool with mirrors on initiator and produce dd write about 10gb in loop

root@dilos-tar1:/var/crash/myhost# mdb *.4
Loading modules: [ unix genunix specfs dtrace mac cpu.generic uppc apix scsi_vhci zfs sd ip hook neti sockfs arp usba stmf stmf_sbd mm random cpc ptm ufs logindmux nfs ]
> $C
ffffff000a9a4530 pvscsi_transport_command+0x18a(ffffff0254d6b0c0, ffffff8de826b600)
ffffff000a9a4580 pvscsi_start+0x158(ffffff8de826b740, ffffff8de826b738)
ffffff000a9a45d0 scsi_transport+0xbd(ffffff8de826b738)
ffffff000a9a4660 sd_start_cmds+0x164(ffffff025661b040, 0)
ffffff000a9a46b0 sd_core_iostart+0x90(4, ffffff025661b040, ffffff02573ceb80)
ffffff000a9a4740 sd_mapblockaddr_iostart+0x2d3(3, ffffff025661b040, ffffff02573ceb80)
ffffff000a9a4790 sd_xbuf_strategy+0x53(ffffff02573ceb80, ffffff02a3a20b40, ffffff025661b040)
ffffff000a9a4800 xbuf_iostart+0x178(ffffff025b542cc0)
ffffff000a9a4840 ddi_xbuf_qstrategy+0x79(ffffff02573ceb80, ffffff025b542cc0)
ffffff000a9a4880 sdstrategy+0xc0(ffffff02573ceb80)
ffffff000a9a48b0 bdev_strategy+0x59(ffffff02573ceb80)
ffffff000a9a48e0 ldi_strategy+0x4e(ffffff025e6bd4b0, ffffff02573ceb80)
ffffff000a9a4920 vdev_disk_io_start+0x101(ffffff2072497400)
ffffff000a9a4970 zio_vdev_io_start+0xad(ffffff2072497400)
ffffff000a9a49a0 zio_execute+0x76(ffffff2072497400)
ffffff000a9a49d0 zio_nowait+0x2b(ffffff2072497400)
ffffff000a9a4a10 vdev_queue_io_done+0x92(ffffff0850c17030)
ffffff000a9a4a50 zio_vdev_io_done+0xa0(ffffff0850c17030)
ffffff000a9a4a80 zio_execute+0x76(ffffff0850c17030)
ffffff000a9a4b30 taskq_thread+0x2a6(ffffff025e42bec0)
ffffff000a9a4b40 thread_start+8()
> ::status
debugging crash dump vmcore.4 (64-bit) from dilos-tar1
operating system: 5.11 2.0.1.1 (i86pc)
image uuid: 05b4dc5c-2594-672d-f1b4-ffc3d73c2328
panic message: BAD TRAP: type=e (#pf Page fault) rp=ffffff000a9a43f0 addr=ffffff8de826b764
dump content: kernel pages and pages from PID -1

> ffffff0254d6b0c0::whatis
ffffff0254d6b0c0 is allocated from kmem_alloc_768:
            ADDR          BUFADDR        TIMESTAMP           THREAD
                            CACHE          LASTLOG         CONTENTS
ffffff0257aeb668 ffffff0254d6b0c0        972a16968 fffffffffbc39140
                 ffffff024ac2f348 ffffff024d8d7bc0 ffffff02512b4778
                 kmem_cache_alloc_debug+0x3b0
                 kmem_cache_alloc+0xcb
                 kmem_zalloc+0x47
                 ddi_soft_state_zalloc+0x88
                 pvscsi_attach+0x46
                 devi_attach+0x8e
                 attach_node+0x8f
                 i_ndi_config_node+0x75
                 i_ddi_attachchild+0x3e
                 devi_attach_node+0x5e
                 devi_config_one+0x3b2
                 ndi_devi_config_one+0xa8
                 resolve_pathname+0x140
                 e_ddi_hold_devi_by_path+0x17
                 e_devid_cache_to_devt_list+0x24f

> ffffff8de826b600::whatis
ffffff8de826b600 is ffffff8d159e8000+d2883600, freed from the heap vmem arena:
            ADDR TYPE            START              END             SIZE
                                THREAD        TIMESTAMP                 
ffffff02ac54bae0 FREE ffffff8d159e8000 ffffff8de8286000       -762716160

they are input parameters to pvscsi_transport_command()
as we can see - we have second incorrect one - it was freed.

Actions #1

Updated by Igor Kozhukhov almost 5 years ago

  • Subject changed from pvscsci panic to pvscsi panic
Actions #2

Updated by Igor Kozhukhov almost 5 years ago

  • Description updated (diff)
Actions

Also available in: Atom PDF