Apple M1 Max, OS 13.3.1 (a) (22E772610a)
XQuartz 2.8.5 (xorg-server 21.1.6)
I've been using XQuartz for 10 years and more, and it's normally very
stable. Lately though uptime is measured in days, not months. Why
might that be?
Yesterday it happened again. As usual, the system was utterly idle. I
stopped work around 6 in the evening, and it crashed overnight while no
one was using the computer.
Oddly, the crash message -- the one that duly captures the details and
is deftly ignored by Apple -- indicates a round number of seconds
of uptime. I didn't look this morning, but on July 7 it reported
"uptime" of 21000000 seconds.
Time Since Wake: 3525967 seconds.
Each crash increments "uptime" by 1,000,000 seconds. The first crash
clocked in IIRC at 16000000, then 17000000, etc.
I can't imagine what those numbers mean; it's clearly not what they
say. The same machine says:
$ uptime
16:29 up 287 days, 10:58, 1 user, load averages: 3.17 3.10 2.95
$ echo $(( 21000000 / (24 *3600)))
243
$ echo $(( 3525967 / (24 *3600)))
40
I restarted XQuartz 3 days ago, not before The Flood. To claim it's
been up as long as the machine (almost) is silly, especially when
the number only grows with each restart.
Time Since Wake might be accurate if we subtract the value from one
crash to the next:
$ echo '(3525967 - 3267838) / (24 * 3600)' | bc -l
2.98760416666666666666
i.e., 3 days.
The system is under no great pressure (especially when idle!)
Activity Monitor reports 53 of 64 GB RAM in use, and
Disk Utility about 50% of 1 TB free.
Console currently reports X11 crashes on July 4, 7, and 10. As
shown below, each one looks very much like a bug, using a space for a
pointer.
Is there anything else I can report or do, or does the usual
reboot-and-hope advice apply?
I will give voice to one sneaking suspicion: the Nano library. The
crash report mentions
MALLOC_NANO (reserved) 600018000000-600020000000 [128.0M]
rw-/rwx SM=NUL ...(unallocated)
About 15 years ago IIRC I reported a bug in nano. It failed to
check for NULL returned by a failed malloc(3). Developed on Linux,
where overallocation of memory is the norm, it apparently assumed (at
least there) that 0 wasn't a possibility. (The requested allocation
was also ridiculously large.) macOS being a BSD, 0 is possible and
deadly.
If on scant evidence my hypothesis is right, many environment variables
are available to let malloc(3) give us better information. Trouble is,
XQuartz is normally started by launchd and me, from the Dock. How to
intervene with an environment variable is black-box mystery.
Bug evidence follows. Many thanks to anyone brave enough to go
where angels fear to tread.
--jkl
[snip]
July 4:
Crashed Thread: 1
Exception Type: EXC_BAD_ACCESS (SIGSEGV)
Exception Codes: KERN_INVALID_ADDRESS at 0x0000000000000020
Exception Codes: 0x0000000000000001, 0x0000000000000020
Termination Reason: Namespace SIGNAL, Code 11 Segmentation fault: 11
Terminating Process: exc handler [8562]
VM Region Info: 0x20 is not in any region. Bytes before following
region: 105553518919648
July 7:
Crashed Thread: 1
Exception Type: EXC_BAD_ACCESS (SIGSEGV)
Exception Codes: KERN_INVALID_ADDRESS at 0x0000000000000020
Exception Codes: 0x0000000000000001, 0x0000000000000020
Termination Reason: Namespace SIGNAL, Code 11 Segmentation fault: 11
Terminating Process: exc handler [85493]
VM Region Info: 0x20 is not in any region. Bytes before following
region: 105553518919648
July 10:
Crashed Thread: 1
Exception Type: EXC_BAD_ACCESS (SIGSEGV)
Exception Codes: KERN_INVALID_ADDRESS at 0x0000000000000020
Exception Codes: 0x0000000000000001, 0x0000000000000020
Termination Reason: Namespace SIGNAL, Code 11 Segmentation fault: 11
Terminating Process: exc handler [8382]
VM Region Info: 0x20 is not in any region. Bytes before following
region: 105553518919648
[pins]
_______________________________________________
Do not post admin requests to the list. They will be ignored.
X11-users mailing list ([email protected])
Help/Unsubscribe/Update your Subscription:
https://lists.apple.com/mailman/options/x11-users/archive%40mail-archive.com
This email sent to [email protected]
