[Bug 1915006] Re: Consider merging 4.0-5 for Hirsute

2021-02-22 Thread Christian Ehrhardt 
** Changed in: glibc (Ubuntu) Status: New => Invalid -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/1915006 Title: Consider merging 4.0-5 for Hirsute To manage notifications about this bug

[Bug 1915006] Re: Consider merging 4.0-5 for Hirsute

2021-02-19 Thread Launchpad Bug Tracker
This bug was fixed in the package chrony - 4.0-5ubuntu2 --- chrony (4.0-5ubuntu2) hirsute; urgency=medium * d/p/lp-1915006-sys_linux-allow-statx-and-fstatat64-in-seccomp-filte.patch: add compatibility for glibc 2.33 (LP: 1915006) chrony (4.0-5ubuntu1) hirsute; urgency=medium

[Bug 1915006] Re: Consider merging 4.0-5 for Hirsute

2021-02-17 Thread Christian Ehrhardt 
Repro in canonistack and locally were good and eventually also in proposed-migration on a retry. It now is ready, but entangled with glibc. -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/1915006

[Bug 1915006] Re: Consider merging 4.0-5 for Hirsute

2021-02-16 Thread Christian Ehrhardt 
This is haunted. While working in former sniff tests and now the fix works. I'm seeing a hang on armhf on another test now. https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac /autopkgtest- hirsute/hirsute/armhf/c/chrony/20210215_234914_e1be6@/log.gz

[Bug 1915006] Re: Consider merging 4.0-5 for Hirsute

2021-02-15 Thread Christian Ehrhardt 
My fix was upstream accepted and successfully tested in arm canonistack instances. Uploading that to Hirsute to not only make 4.0-5* pass but also to un-break it with the new glibc. -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu.

[Bug 1915006] Re: Consider merging 4.0-5 for Hirsute

2021-02-12 Thread Christian Ehrhardt 
I have a fix that works in old glibc environments. With those two I'm proposing that to upstream (and FYI the Debian maintainer). Should pop up any minute at https://listengine.tuxfamily.org/chrony.tuxfamily.org/chrony-dev/2021/02/threads.html Once we have feedback on this (looking at other such

[Bug 1915006] Re: Consider merging 4.0-5 for Hirsute

2021-02-12 Thread Christian Ehrhardt 
But it does not stop there, then I next hit: 0.003749 SYS_327(-100, 0xffdbcc3c, 0xffdbcb50, 0) = -100 <0.000532> 0.000821 --- SIGSYS (Bad system call) --- That is fstatat64 -- You received this bug notification because you

[Bug 1915006] Re: Consider merging 4.0-5 for Hirsute

2021-02-12 Thread Christian Ehrhardt 
chrony 4.0-2ubuntu1 @ glibc 2.32 /usr/sbin/chronyd -d -x -f /etc/chrony/chrony.conf -u root -F -1 2021-02-12T08:56:55Z chronyd version 4.0 starting (+CMDMON +NTP +REFCLOCK +RTC +PRIVDROP +SCFILTER +SIGND +ASYNCDNS +NTS +SECHASH +IPV6 -DEBUG) 2021-02-12T08:56:55Z Wrong owner of /run/chrony (UID !=

[Bug 1915006] Re: Consider merging 4.0-5 for Hirsute

2021-02-12 Thread Christian Ehrhardt 
The bug doesn't have a glibc task to fix it there, but to hold the glibc transition until a fix is available otherwise we break this 21.04. Also adding update-excuse as this is a failing dependent test of glibc. ** Tags added: update-excuse -- You received this bug notification because you are

[Bug 1915006] Re: Consider merging 4.0-5 for Hirsute

2021-02-12 Thread Christian Ehrhardt 
But if we go back libc6 2.33-0ubuntu2 => 2.32-0ubuntu6 and run the same it works just fine. So it is libc6 2.33 that introduces the breakage by adding another syscall that is currently blocked. Looking into past cases we won't change libc for this, but instead want to add this as an allowed

[Bug 1915006] Re: Consider merging 4.0-5 for Hirsute

2021-02-12 Thread Christian Ehrhardt 
Starting it without going to the background confirmed my assumption: ../../chronyd -d -x -f /root/chrony-4.0/test/system/tmp/chronyd.conf -u root -F -1 2021-02-12T08:28:54Z chronyd version 4.0 starting (+CMDMON +NTP +REFCLOCK +RTC +PRIVDROP +SCFILTER +SIGND +ASYNCDNS +NTS +SECHASH +IPV6 -DEBUG)

[Bug 1915006] Re: Consider merging 4.0-5 for Hirsute

2021-02-12 Thread Christian Ehrhardt 
Debugging the testcase: The test checks for filtering, which is compiled in: "+CMDMON +NTP +REFCLOCK +RTC +PRIVDROP +SCFILTER +SIGND +ASYNCDNS +NTS +SECHASH +IPV6 -DEBUG" contains "+SCFILTER" Then it starts chrony, once with -F 1 and once with -F -1 which controls the effect of the SCFilter.

[Bug 1915006] Re: Consider merging 4.0-5 for Hirsute

2021-02-12 Thread Christian Ehrhardt 
Recreating this on canonistack arm64 host with armhf LXD container. Building chrony locally and running the test reproduces the test. It is good that I can re-create this already at build-time. That makes it less complex to switch glibc between 2.32 and 2.33 and doing a rebuild to check. I

[Bug 1915006] Re: Consider merging 4.0-5 for Hirsute

2021-02-12 Thread Vincent Blut
Hi Christian, Could you please send the output of 'strace -o chrony_debug.txt -- chronyd -d -F -1'? -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/1915006 Title: Consider merging 4.0-5 for Hirsute

[Bug 1915006] Re: Consider merging 4.0-5 for Hirsute

2021-02-11 Thread Christian Ehrhardt 
The merge and tests went mostly fine, except for one issue. The test "run_system_tests" executes a subtest called "005-scfilter". This one started to fail before the new upload. [1]: contains a bunch of runs on the former 4.0-2ubuntu1 that failed the same way. With the new code it is the same,

[Bug 1915006] Re: Consider merging 4.0-5 for Hirsute

2021-02-11 Thread Christian Ehrhardt 
MP: https://code.launchpad.net/~paelzer/ubuntu/+source/chrony/+git/chrony/+merge/397790 Debian-MR: https://salsa.debian.org/debian/chrony/-/merge_requests/6 PPA: https://launchpad.net/~ci-train-ppa-service/+archive/ubuntu/4423/+packages Uplaoded to Hirsute ... ** Merge proposal linked: