tcl & tk ports on FreeBSD 10 amd64

From: Cy Schubert <Cy.Schubert_at_komquats.com>
Date: Tue, 03 Dec 2013 21:32:56 -0800
Hi,

Sorry for cross posting but this concerns both lists.

Over the last month or so I've been upgrading my prod infrastructure from 9 
to 10. It's mostly complete except for a number of issues. One issue, just 
solved today (circumvented is a better word), is exmh crashing 10.0 on 
amd64 while on i386 there are no issues with exmh.

It appears that the tcl and tk ports (all three of them, 8.4, 8.5, and 8.6) 
will panic 10.0 on amd64 (but not i386) when the ports are built with 
threading support.

I haven't had a chance to look at the dump yet but I had a hunch to test 
the ports without threading support enabled, making the panic go away. If I 
don't get to it in time, here is what I haveFatal trap 9: general 
protection fault while in kernel mode
cpuid = 2; apic id = 02
instruction pointer     = 0x20:0xffffffff80957aeb
stack pointer           = 0x28:0xfffffe00f17f9980
frame pointer           = 0x28:0xfffffe00f17f99a0
code segment            = base 0x0, limit 0xfffff, type 0x1b
                        = DPL 0, pres 1, long 1, def32 0, gran 1
processor eflags        = interrupt enabled, resume, IOPL = 0
current process         = 11 (idle: cpu2)
trap number             = 9
timeout stopping cpus
panic: general protection fault
cpuid = 2
KDB: stack backtrace:
db_trace_self_wrapper() at db_trace_self_wrapper+0x2b/frame 
0xfffffe00f17f9510
kdb_backtrace() at kdb_backtrace+0x39/frame 0xfffffe00f17f95c0
panic() at panic+0x153/frame 0xfffffe00f17f9640
trap_fatal() at trap_fatal+0x3a2/frame 0xfffffe00f17f96a0
trap() at trap+0x7bf/frame 0xfffffe00f17f98c0
calltrap() at calltrap+0x8/frame 0xfffffe00f17f98c0
--- trap 0x9, rip = 0xffffffff80957aeb, rsp = 0xfffffe00f17f9980, rbp = 
0xfffffe
00f17f99a0 ---
cpu_idle_hlt() at cpu_idle_hlt+0x2b/frame 0xfffffe00f17f99a0
cpu_idle() at cpu_idle+0x93/frame 0xfffffe00f17f99c0
sched_idletd() at sched_idletd+0x1ee/frame 0xfffffe00f17f9a70
fork_exit() at fork_exit+0x9a/frame 0xfffffe00f17f9ab0
fork_trampoline() at fork_trampoline+0xe/frame 0xfffffe00f17f9ab0
--- trap 0, rip = 0, rsp = 0xfffffe00f17f9b70, rbp = 0 ---
Uptime: 4m42s
Dumping 522 out of 5932 MB:..4%..13%..22%..31%..43%..53%..62%..71%..83%..92%
 so far,

Tcl/tk are tickling some bug somewhere.

Before anyone suggests memory, I've been able to reproduce this on an Intel 
Core i3 machine with 6 GB and an AMD X2 5000+, also with 6 GB, both in 
amd64 mode. Both systems are dual (or multi) boot. The bug does not exhibit 
itself in i386 mode. It also doesn't exhibit itself when tcl/tk are built 
without thread support.

The only application I know of which tickles the bug is mail/exmh2 (I'm the 
maintainer) when using a threaded tcl/tk.

My 11-CURRENT partition on my laptop is still i386 so I haven't been able 
to reproduce it under 11 with amd64.


-- 
Cheers,
Cy Schubert <Cy.Schubert_at_komquats.com>
FreeBSD UNIX:  <cy_at_FreeBSD.org>   Web:  http://www.FreeBSD.org

	The need of the many outweighs the greed of the few.
Received on Wed Dec 04 2013 - 04:33:04 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:40:45 UTC