DomainKey-Signature: a=rsa-sha1; c=nofws; d=sourceware.org; h=list-id
	:list-unsubscribe:list-subscribe:list-archive:list-post
	:list-help:sender:mime-version:reply-to:from:date:message-id
	:subject:to:content-type:content-transfer-encoding; q=dns; s=
	default; b=WmGeqe69caflbCzdNnhSmPo3CQb41Jucw7xN2hJn8mhgVCMmK4Nfu
	xutjeFCadjvSS1XYKUglX0CQ7K1IB+leZuoWIsmi5P+jSKhCvULtaFxyn9vf7G4k
	Giib0ZTLAUrqAW19E0/grgX9apFwXX5uZek2O4WAG4tioHsAa5l108=
Mailing-List: contact cygwin-help AT cygwin DOT com; run by ezmlm
Sender: cygwin-owner AT cygwin DOT com
Mail-Followup-To: cygwin AT cygwin DOT com
MIME-Version: 1.0
Reply-To: fracting AT gmail DOT com
From: Qian Hong <fracting AT gmail DOT com>
Date: Wed, 28 Oct 2015 17:45:19 +0800
Message-ID: <CALd+sZS-EbkUMgmRCyJUK0LCMo6zydzB1rbM9iZFrAcnbfHCRQ@mail.gmail.com>
Subject: Randomly hang when compiling Cygwin on Cygwin on Wine
To: cygwin <cygwin AT cygwin DOT com>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Dear folks,

I try to compile Cygwin on Cygwin on Wine in a loop, and found a
random hang. Usually, it happens more than 1/10.

The hang could happen in different processes, including `sh.exe`,
`gcc.exe`, `as.exe`, etc, but the backtraces are all similar.


Here is one example I get now:

(1) `pstree -p -l` suggest the hanging process might has PID 113:
?(1)─┬─bash(58)───make(93)───sh(89)───make(69)───sh(74)───sh(61)───make(9)───make(44)───sh(129)───sh(145)───make(94)───sh(130)───sh(104)───make(151)───sh(113)
     └─mintty(91)───bash(102)───pstree(141)

(2) `cat /proc/113/cmdline` show below information:

/bin/sh -c gcc -L/newlib-cygwin-all/newlib-cygwin-build/i686-pc-cygwin/winsup/cygwin
-isystem /newlib-cygwin-all/newlib-cygwin/winsup/cygwin/include
-B/newlib-cygwin-all/newlib-cygwin-build/i686-pc-cygwin/newlib/
-isystem /newlib-cygwin-all/newlib-cygwin-build/i686-pc-cygwin/newlib/targ-include
-isystem /newlib-cygwin-all/newlib-cygwin/newlib/libc/include
-I/newlib-cygwin-all/newlib-cygwin/winsup/cygwin/include
-DPACKAGE_NAME=\"newlib\" -DPACKAGE_TARNAME=\"newlib\"
-DPACKAGE_VERSION=\"2.2.0\" -DPACKAGE_STRING=\"newlib\ 2.2.0\"
-DPACKAGE_BUGREPORT=\"\" -DPACKAGE_URL=\"\" -I.
-I/newlib-cygwin-all/newlib-cygwin/newlib/libc/xdr -DHAVE_OPENDIR
-DHAVE_RENAME -DSIGNAL_PROVIDED -D_COMPILING_NEWLIB -DHAVE_BLKSIZE
-DHAVE_FCNTL -DMALLOC_PROVIDED -fno-builtin      -g -O2 -c -o
lib_a-xdr_float.o `test -f 'xdr_float.c' || echo
'/newlib-cygwin-all/newlib-cygwin/newlib/libc/xdr/'`xdr_float.c


(3) I cd to the build target directory and list all files:
/newlib-cygwin-all/newlib-cygwin-build/i686-pc-cygwin/newlib/libc/xdr$ ls
lib_a-dummy.o  lib_a-xdr.o  lib_a-xdr_private.o  Makefile

The above list result suggest that the build target
"lib_a-xdr_float.o" hasn't been generated, which suggested that the
`gcc` process in the above command hasn't run.

(4) Then I check the file descriptors opened by process 113:
$ file /proc/113/fd/*
/proc/113/fd/0: broken symbolic link to /dev/cons0
/proc/113/fd/1: broken symbolic link to /dev/cons0
/proc/113/fd/2: broken symbolic link to /dev/cons0
/proc/113/fd/3: symbolic link to pipe:[296]

(5) I use gdb to attach to the process 113:

[New Thread 64.0x3c]
[New Thread 64.0x4c]
[New Thread 64.0x51]
Reading symbols from /usr/bin/sh.exe...Reading symbols from
/usr/lib/debug//usr/bin/sh.exe.dbg...done.
done.

0xb77a0be6 in ?? ()
(gdb) bt
#0  0xb77a0be6 in __kernel_vsyscall+0xe in [vdso].so
#1  0x7bc82ea5 in server_select+0x315
[/home/fracting/src/wine-patched-auto/dlls/ntdll/server.c:642] in
ntdll
#2  0x7bc8ade9 in wait_objects+0x89
[/home/fracting/src/wine-patched-auto/dlls/ntdll/sync.c:1034] in ntdll
#3  0x7bc8d5e9  in __syscall_NtWaitForMultipleObjects+0x29
[/home/fracting/src/wine-patched-auto/dlls/ntdll/sync.c:1047] in ntdll
#4  0x7b87bbe5 in WaitForMultipleObjectsEx+0x1b5
[/home/fracting/src/wine-patched-auto/dlls/kernel32/sync.c:186] in
kernel32
#5  0x7b87ba21 in WaitForMultipleObjects+0x31
[/home/fracting/src/wine-patched-auto/dlls/kernel32/sync.c:146] in
kernel32
#6  0x61006354 in cygwait(void*, _LARGE_INTEGER*, unsigned int)@12
(object=<optimized out>, timeout=0x7bc4861c, mask=2076474397)
    at /usr/src/debug/cygwin-2.2.1-1/winsup/cygwin/cygwait.cc:76
#7  0x6130a188 in ?? () from /usr/bin/cygwin1.dll
#8  0x00000000 in ?? ()

(gdb) thread 2
[Switching to thread 2 (Thread 64.0x4c)]
#0  0xb77a0be6 in ?? ()
(gdb) bt
#0  0xb77a0be6 in __kernel_vsyscall+0xe in [vdso].so
#1  0x7bc4aadc in __syscall_NtReadFile+0x41c
[/media/workspace/src/wine-patched-auto/dlls/ntdll/file.c:1104] in
ntdll
#2  0x7b83f63d in ReadFile+0x1c6
[/home/fracting/src/wine-patched-auto/dlls/kernel32/file.c:448] in
kernel32
#3  0x610efbae in wait_sig () at
/usr/src/debug/cygwin-2.2.1-1/winsup/cygwin/sigproc.cc:1252
#4  0x611b6c40 in cygthread::main_thread_id () from /usr/bin/cygwin1.dll


(gdb) thread 3
[Switching to thread 3 (Thread 64.0x51)]
#0  0xb77a0be6 in ?? ()
(gdb) bt
#0  0xb77a0be6 in __kernel_vsyscall+0xe in [vdso].so
#1  0x7bc82ea5 in server_select+0x315
[/home/fracting/src/wine-patched-auto/dlls/ntdll/server.c:642] in
ntdll
#2  0x7bc8ade9 in wait_objects+0x89
[/home/fracting/src/wine-patched-auto/dlls/ntdll/sync.c:1034] in ntdll
#3  0x7bc8d5e9 in  __syscall_NtWaitForMultipleObjects+0x29
[/home/fracting/src/wine-patched-auto/dlls/ntdll/sync.c:1047] in ntdll
#4  0x7b87bbe5 in WaitForMultipleObjectsEx+0x1b5
[/home/fracting/src/wine-patched-auto/dlls/kernel32/sync.c:186] in
kernel32
#5  0x7b87b9a3 in WaitForSingleObject+0x2e
[/home/fracting/src/wine-patched-auto/dlls/kernel32/sync.c:126] in
kernel32
#6  0x6100548a in cygthread::stub(void*)@4 (arg=0xacce64) at
/usr/src/debug/cygwin-2.2.1-1/winsup/cygwin/cygthread.cc:114
#7  0x61005400 in cygthread::release(bool) () at
/usr/src/debug/cygwin-2.2.1-1/winsup/cygwin/cygthread.cc:292
#8  0x00acce64 in ?? ()
#9  0x00000051 in ?? ()

Here are the version information I have:

$ uname -a
CYGWIN_NT-5.1  2.2.1(0.289/5/3) 2015-08-20 11:40 i686 Cygwin

$ sh --version
GNU bash, version 4.3.42(4)-release (i686-pc-cygwin)

$ wine --version
wine-1.7.53-859-g6ec39dd (Staging)

This bug also happens with latest git version of Cygwin.


It might be either Wine bug or Cygwin bug, but currently I can't go
further, since:
- I can't reproduce this bug with strace.exe
- I also can't reproduce this bug with Wine's relay debug trace, which
records system API calls.

Could you suggest any hints where should I look deeper?

Thanks!

-- 
Regards,
Qian Hong

-
http://www.winehq.org

--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple