Hello!
Using default (?) kernel option
CONFIG_DEBUG_KMEMLEAK_EARLY_LOG_SIZE = 400
there's a kernel boot message:
[ 0.339759] kmemleak: Early log buffer exceeded (24635), please
increase DEBUG_KMEMLEAK_EARLY_LOG_SIZE
increasing CONFIG_DEBUG_KMEMLEAK_EARLY_LOG_SIZE to suggested size,
produces the following kernel TPC/OOPS for me, withing 1 minute after
machine boots:
$ zgrep DEBUG_KMEMLEAK /proc/config.gz
CONFIG_HAVE_DEBUG_KMEMLEAK=y
CONFIG_DEBUG_KMEMLEAK=y
CONFIG_DEBUG_KMEMLEAK_EARLY_LOG_SIZE=24636
# CONFIG_DEBUG_KMEMLEAK_TEST is not set
# CONFIG_DEBUG_KMEMLEAK_DEFAULT_OFF is not set
$ systemd-analyze
Startup finished in 7.670s (kernel) + 22.578s (userspace) = 30.249s
graphical.target reached after 22.474s in userspace
[ OK ] Started Samba SMB Daemon.
[ OK ] Reached target Multi-User System.
[ OK ] Reached target Graphical Interface.
Starting Update UTMP about System Runlevel Changes...
[ OK ] Started Update UTMP about System Runlevel Changes.
[ 31.691103] tun: Universal TUN/TAP device driver, 1.6
Debian GNU/Linux buster/sid ttip console
[ 65.561601] log_unaligned: 9641 callbacks suppressed
[ 65.561618] Kernel unaligned access at TPC[65a5d4] kmemleak_scan+0x2d4/0x820
[ 65.561813] Unable to handle kernel NULL pointer dereference in mna handler
[ 65.561819] at virtual address 000000000000001a
[ 65.561890] current->{active_,}mm->context = 0000000000000244
[ 65.561927] current->{active_,}mm->pgd = fff800409f494000
[ 65.561964] \|/ ____ \|/
[ 65.561964] "@'/ .. \`@"
[ 65.561964] /_| \__/ |_\
[ 65.561964] \__U_/
[ 65.562046] kmemleak(197): Oops [#1]
[ 65.562077] CPU: 8 PID: 197 Comm: kmemleak Not tainted 4.16.0-rc2 #633
[ 65.562121] TSTATE: 0000009980001604 TPC: 000000000065a5d4 TNPC:
000000000065a5d8 Y: 00001243 Not tainted
[ 65.562182] TPC: <kmemleak_scan+0x2d4/0x820>
[ 65.562213] g0: 0000000000000000 g1: fffffffffffffffe g2:
0000000000000001 g3: fffffffffffffffe
[ 65.562266] g4: fff80000fb4491c0 g5: fff80040bc040000 g6:
fff80040a1a24000 g7: 0000000000000200
[ 65.562319] o0: 00060000486c0000 o1: 00060000023fe109 o2:
0000000000000000 o3: 0000000000c28c00
[ 65.562375] o4: 0000000000c28c00 o5: 0000000000000000 sp:
fff80040a1a27481 ret_pc: 000000000065a5e8
[ 65.562432] RPC: <kmemleak_scan+0x2e8/0x820>
[ 65.562465] l0: 0000000000b4dac8 l1: 000000000205fca3 l2:
00000000007fffff l3: 0000000001844800
[ 65.562519] l4: 0000000000be1800 l5: 0000000000000000 l6:
0000000000000000 l7: 0000000000000000
[ 65.562574] i0: 00000000486c0000 i1: 0000000001847000 i2:
0000000000c28c00 i3: 0000000000bf9400
[ 65.562628] i4: 0000000000c28dc0 i5: 0000000001018000 i6:
fff80040a1a27541 i7: 000000000065b10c
[ 65.562688] I7: <kmemleak_scan_thread+0x8c/0x100>
[ 65.562724] Call Trace:
[ 65.562750] [000000000065b10c] kmemleak_scan_thread+0x8c/0x100
[ 65.562797] [000000000048bbb4] kthread+0x114/0x140
[ 65.562840] [0000000000406084] ret_from_fork+0x1c/0x2c
[ 65.562878] [0000000000000000] (null)
[ 65.562912] Disabling lock debugging due to kernel taint
[ 65.562936] Caller[000000000065b10c]: kmemleak_scan_thread+0x8c/0x100
[ 65.562961] Caller[000000000048bbb4]: kthread+0x114/0x140
[ 65.562984] Caller[0000000000406084]: ret_from_fork+0x1c/0x2c
[ 65.563006] Caller[0000000000000000]: (null)
[ 65.563026] Instruction DUMP:
[ 65.563030] 84086001
[ 65.563046] 82007fff
[ 65.563060] 87789401
[ 65.563073] <c200e01c>
[ 65.563088] 80a06000
[ 65.563101] 2248000f
[ 65.563114] ba076001
[ 65.563127] 94102000
[ 65.563140] 7ffffe6e
[ 65.563153]
Not sure how this affects system (except system boots almost 10
seconds slower than usually), since it works and no more OOPS are
generated.
PS: any value which is more than suggested (tested kernel boot with
24636, 25000, 40000) produces TPC/OOPS. Any value which is less than
24635 (tested with 400, 12000,18000, 22000, 24000), does not produce
TPC/OOPS.
--
To unsubscribe from this list: send the line "unsubscribe sparclinux" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
