bug#38912: 27.0.60; PDumper meets segmentation fault when evil is loaded

[Daniel Colascione]

> [Please use "Reply All" to reply, so that the bug address is kept on
> the CC list.]
>
>> From: NiwTinray <niwtrx@icloud.com>
>> Date: Sun, 5 Jan 2020 13:25:07 +0800
>>
>> > I cannot reproduce this from "emacs -Q" because 'use-package' is not a
>> > known function.  Can you show a recipe starting from "emacs -Q",
>> > please?
>>
>> Here. I've attached a minimal script file that helps reproduce this bug.
>>
>> (require 'package)
>> (package-initialize)
>> (add-to-list 'package-archives
>>              '("melpa-stable" . "https://stable.melpa.org/packages/";) t)
>> (unless (package-installed-p 'evil)
>>   (package-refresh-contents)
>>   (package-install 'evil))
>> (require 'evil)
>> (dump-emacs-portable "/tmp/test.pdmp")
>>
>> The script downloads the package "evil" from Melpa stable, load the evil
>> package
>> and dumps an image to /tmp/test.pdmp.
>>
>> > Also, does this happen if you add -Q to the Emacs invocation after
>> > dumping?  If not, there's more detail missing in your report: the
>> > customizations in your init files.
>>
>>
>> Sure. Please download this file, and run the command:
>>
>> emacs --batch -Q --script evil.el
>>
>> To see the bug happen, load the test.pdmp file:
>>
>> emacs -Q --dump-file /tmp/test.pdmp
>>
>> You should see a segmentation fault:
>>
>> [1]    23369 segmentation fault (core dumped)  emacs -Q --dump-file
>> /tmp/test.pdmp
>>
>> I run debugger inside src/.gdbinit using the command:
>>
>> gdb -x .gdbinit --args ./emacs --dump-file /tmp/test.pdmp
>>
>> And logged backtrace. See my second attachment:
>>
>> (base) omnisky :: ~/emacs/src ‹emacs-27*› » gdb -x .gdbinit --args
>> ./emacs --dump-file /tmp/test.pdmp
>> GNU gdb (Ubuntu 7.11.1-0ubuntu1~16.5) 7.11.1
>> Copyright (C) 2016 Free Software Foundation, Inc.
>> License GPLv3+: GNU GPL version 3 or later
>> <http://gnu.org/licenses/gpl.html>
>> This is free software: you are free to change and redistribute it.
>> There is NO WARRANTY, to the extent permitted by law.  Type "show
>> copying"
>> and "show warranty" for details.
>> This GDB was configured as "x86_64-linux-gnu".
>> Type "show configuration" for configuration details.
>> For bug reporting instructions, please see:
>> <http://www.gnu.org/software/gdb/bugs/>.
>> Find the GDB manual and other documentation resources online at:
>> <http://www.gnu.org/software/gdb/documentation/>.
>> For help, type "help".
>> Type "apropos word" to search for commands related to "word"...
>> Reading symbols from ./emacs...done.
>> warning: File "/home/ntr/emacs/src/.gdbinit" auto-loading has been
>> declined by your `auto-load safe-path' set to
>> "$debugdir:$datadir/auto-load".
>> To enable execution of this file add
>>      add-auto-load-safe-path /home/ntr/emacs/src/.gdbinit
>> line to your configuration file "/home/ntr/.gdbinit".
>> To completely disable this security protection add
>>      set auto-load safe-path /
>> line to your configuration file "/home/ntr/.gdbinit".
>> For more information about this security protection see the
>> "Auto-loading safe path" section in the GDB manual.  E.g., run from the
>> shell:
>>      info "(gdb)Auto-loading safe path"
>> SIGINT is used by the debugger.
>> Are you sure you want to change it? (y or n) [answered Y; input not from
>> terminal]
>> Environment variable "DISPLAY" not defined.
>> TERM = xterm-24bits
>> Breakpoint 1 at 0x411df0: file emacs.c, line 370.
>> Breakpoint 2 at 0x4bfe60: file xterm.c, line 10130.
>> (gdb) r
>> Starting program: /home/ntr/emacs/src/emacs --dump-file /tmp/test.pdmp
>> /home/ntr/emacs/src/emacs:
>> /raid_sdc/home/ntr/anaconda3/lib/libtiff.so.5: no version information
>> available (required by /home/ntr/emacs/src/emacs)
>> [Thread debugging using libthread_db enabled]
>> Using host libthread_db library
>> "/lib/x86_64-linux-gnu/libthread_db.so.1".
>>
>> Program received signal SIGSEGV, Segmentation fault.
>> 0x00000000004f12d0 in Fcurrent_active_maps (olp=olp@entry=XIL(0x30),
>> position=position@entry=XIL(0)) at keymap.c:1541
>> 1541       && NILP (KVAR (current_kboard,
>> Voverriding_terminal_local_map))
>> (gdb) xbacktrace
>> "key-binding" (0xffffd5c8)
>> "turn-on-undo-tree-mode" (0xffffd758)
>> "global-undo-tree-mode-enable-in-buffers" (0xffffd948)
>> "run-hooks" (0xffffd9e8)
>> "run-mode-hooks" (0xffffdbc0)
>> "minibuffer-inactive-mode" (0xffffdd40)
>> (gdb)
>
> In my debug build of Emacs 27.0.60 I get an assertion violation while
> dumping, which probably is already a sign of trouble.
>
> Daniel, any ideas?

I haven't had a chance over the past few days to repro this problem, but I
hope to do so this weeekend. The messages about the assertion failure
*during* dumping do seem likely unrelated. The easiest way to debug this
particular crash is with rr. Run the original test as "rr record emacs
[args]", then run "rr replay". The latter will dump you in a GDB prompt.
Type "cont" and run the replay of Emacs until you get to the SIGSEGV. Run
"watch -l [expression]" to break out of execution whenever that value
changes, then run "reverse-cont" to run the replay *backwards* until you
get to the code that changed the variable with the bad value.