[LU-17724] sanity-sec test_69: crash the setting sptlrpc.gss.rsi_upcall=prog

• Type: Bug
• Resolution: Fixed
• Priority: Minor
• Fix Version/s: Lustre 2.16.0
• Affects Version/s: Lustre 2.16.0
• Severity: 3
• Assignee: Sebastien Buisson
• Reporter: Maloo

Description

This issue was created by maloo for S Buisson <[e-mail]>

This issue relates to the following test suite run: https://testing.whamcloud.com/test_sets/d5790b0a-ed8c-4e21-a5c3-6395565db7b7

test_69 failed with the following error:

trevis-102vm1 crashed during sanity-sec test_69

Test session details:
clients: https://build.whamcloud.com/job/lustre-reviews/103999 - 4.18.0-513.18.1.el8_9.x86_64
servers: https://build.whamcloud.com/job/lustre-reviews/103999 - 4.18.0-513.18.1.el8_lustre.x86_64

Stack trace is:

[ 5405.670870] Lustre: DEBUG MARKER: /usr/sbin/lctl set_param sptlrpc.gss.rsi_upcall=prog
[ 5405.802972] BUG: unable to handle kernel paging request at 00007ffe5b8cc009
[ 5405.805753] Oops: 0001 [#1] SMP PTI
[ 5405.806391] CPU: 0 PID: 477214 Comm: lctl 4.18.0-513.18.1.el8_lustre.x86_64 #1
[ 5405.809262] RIP: 0010:string_nocheck+0x12/0x70
[ 5405.825615] Call Trace:
[ 5405.830279]  string+0x44/0x60
[ 5405.830811]  vsnprintf+0x340/0x520
[ 5405.831437]  libcfs_debug_msg+0x889/0xbc0 [libcfs]
[ 5405.832428]  ? upcall_cache_set_upcall+0x37e/0x3f0 [obdclass]
[ 5405.834146]  rsi_upcall_seq_write+0x2e5/0x390 [ptlrpc_gss]
[ 5405.835291]  proc_reg_write+0x39/0x60
[ 5405.836017]  vfs_write+0xa5/0x1b0
[ 5405.836653]  ksys_write+0x4f/0xb0
[ 5405.837245]  do_syscall_64+0x5b/0x1b0

VVVVVVV DO NOT REMOVE LINES BELOW, Added by Maloo for auto-association VVVVVVV
sanity-sec test_69 - trevis-102vm1 crashed during sanity-sec test_69

Issue Links

• is related to: LU-17718 Coverity: 424705 (String overflow) under upcall_cache.c

Comments

[Peter Jones]

Merged for 2.16

[Gerrit Updater]

"Oleg Drokin <[e-mail]>" merged in patch https://review.whamcloud.com/c/fs/lustre-release/+/54730/
Subject: LU-17724 gss: fix bad use of user buffer in rsi upcall
Project: fs/lustre-release
Branch: master
Current Patch Set:
Commit: fe8c195f7a5ef3e653b6eaff8863c4c94e97e28c

[Gerrit Updater]

"Sebastien Buisson <[e-mail]>" uploaded a new patch: https://review.whamcloud.com/c/fs/lustre-release/+/54730
Subject: LU-17724 gss: fix bad use of user buffer in rsi upcall
Project: fs/lustre-release
Branch: master
Current Patch Set: 1
Commit: 99e562b5def84286f65c9ba7419ff7fa9ae1913a

[Sebastien Buisson]

This crash is caused by the use of a user pointer when printing a debug message after upcall_cache_set_upcall returns an error:

	rc = upcall_cache_set_upcall(rsicache, kbuf, count, true);
	if (rc) {
		CERROR("%s: incorrect rsi upcall %.*s. Valid value for sptlrpc.gss.rsi_upcall is an executable pathname: rc = %d\n",
		       rsicache->uc_name, (int)count, buffer, rc);
		GOTO(out, rc);
	}