Hi, In archive-wide Lintian runs, this issue occurred in the following binary packages.
apt_2.1.10_amd64.deb dos2unix_7.4.1-1_amd64.deb manpages-zh_1.6.3.4-1_all.deb mplayer_1.3.0-8+b9_amd64.deb wesnoth-1.14-core_1.14.13-1_amd64.deb A sample command is: man --warnings -E UTF-8 -l -Tutf8 -Z mplayer/usr/share/man/zh_CN/man1/mplayer.1.gz but the screen width must be 80 or smaller to reproduce. Jakub Wilk suggested the minimal code below to reproduce the issue. When the troff subprocess is not killed in a timely manner (within six to ten hours), it will consume all of the host machine's I/O buffers and cause a kernel panic. It was observed several times over the past two weeks. Lintian triaged the issue by setting the value to 120 in this commit, but the change should probably be reverted when this bug is fixed: https://salsa.debian.org/lintian/lintian/-/commit/5be6bd66a9f52872615ef32d234b6d616fd5fa49 The credit for tracking down the issue to groff and suggesting a fix belongs to Jakub Wilk. Thanks! Kind regards Felix Lechner * * * $ mkdir -p usr/share/man/ja/man5 $ printf '\343\201\210\343\201\260\343\200\201.lcs.mit.edu_debian_dists_unstable_contrib_binary\\-i386_Release \n' > usr/share/man/ja/man5/test.5 $ MANWIDTH=80 man --warnings -E UTF-8 -l -Tutf8 -Z usr/share/man/ja/man5/test.5 > /dev/null troff: <standard input>:1: warning [p 1, 0.0i]: cannot adjust line [hangs indefinitely]