Package: libuv1
Version: 1.9.1-1
Severity: normal

Dear Maintainer,

This issue was found during compilation of moar on arm64 where the packaging
process is killed after 150mns of inactivity [1]. moar is built on top on 

Long story [2] short, moar hangs randomly while waiting for a zombie
gcc subprocess.  The SIGCHLD emitted when gcc exits is not used by
moar (actually libuv1)

On real arm64 HW, this is due to weird SigBlk values (from /proc/$pid/status):

SigPnd: 0000000000000000
ShdPnd: 0000000000010000
SigBlk: 929a31e0ee69d200
SigCgt: 0000000180010000

On qemu/arm64, status files also has weird values:

SigPnd: 0000000000000000
ShdPnd: 0000000000010000
SigBlk: eb8a0fa374d9e800
SigIgn: 0000000000000000
SigCgt: fffffffff781feff

Turns out that moar hang no longer happens on qemu/arm64 when
mips-epoll_pwait patch [3] is removed. Note that I could not find a
similar change on upstream libuv.

Unfortunately, I cannot test this on real HW.

Does anyone have an idea on how to fix this issue ?

All the best


-- System Information:
Debian Release: stretch/sid
  APT prefers unstable
  APT policy: (990, 'unstable'), (500, 'testing'), (1, 'experimental')
Architecture: amd64 (x86_64)
Foreign Architectures: i386, armhf

Kernel: Linux 4.7.0-1-amd64 (SMP w/8 CPU cores)
Locale: LANG=en_GB.UTF-8, LC_CTYPE=en_GB.UTF-8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/dash
Init: systemd (via /run/systemd/system)

Versions of packages libuv1 depends on:
ii  libc6  2.24-5

libuv1 recommends no packages.

libuv1 suggests no packages.

-- no debconf information

Pkg-javascript-devel mailing list

Reply via email to