rkward-devel Digest, Vol 86, Issue 12

[Thomas Friedrichsmeier]

A quick wrap up, since we left out the list in the last few mails:

- RKWard currently fails to start on RHEL 8.6, due to a kernel bug.
  Possibly Centos 8 is affected, too.
- Supposedly the issue is fixed in version 4.18.0-392 . This is
  available for Centos, but not for RHEL 8.6, so far.
- If you are affected, and cannot update - for the above , or any other
  reason - you can try a recent snapshot of RKWard git master (such as
  the "daily" Ubuntu builds, or the "nightly" AppImages). Run with the
  commandline option "--quirkmode", in this case.

Regards
Thomas

On Wed, 3 Aug 2022 15:47:01 +0200
Thomas Friedrichsmeier <thomas.friedrichsmeier at kdemail.net> wrote:

> Hi Anders,
> 
> On Wed, 3 Aug 2022 05:48:00 +0000
> Anders Fridberger <anders.fridberger at liu.se> wrote:
> > Hi Thomas
> > I actually enjoy the detective work 😊  
> 
> Glad to hear that.
> 
> > 1)
> > 
> > Thread 6 (Thread 0x7fffc491c700 (LWP 152261)):
> > #0  0x00007fffec48591d in syscall () at /lib64/libc.so.6
> > #1  0x00007fffed38df1c in forkfd_wait4 () at /lib64/libQt5Core.so.5
> > #2  0x00007fffed371f24 in QProcessPrivate::waitForDeadChild() () at
> > /lib64/libQt5Core.so.5 #3  0x00007fffed36bfea in
> > QProcessPrivate::_q_processDied() () at /lib64/libQt5Core.so.5 #4
> > 0x00007fffed371a1e in QProcessPrivate::waitForReadyRead(int) () at
> > /lib64/libQt5Core.so.5 #5  0x000000000062b725 in
> > RKFrontendTransmitter::waitReadLine(QIODevice*, int) () #6
> > 0x000000000062b493 in RKFrontendTransmitter::run() () #7
> > 0x00007fffed229450 in QThreadPrivate::start(void*) () at
> > /lib64/libQt5Core.so.5 #8  0x00007fffea5291cf in start_thread () at
> > /lib64/libpthread.so.0 #9  0x00007fffec485dd3 in clone () at
> > /lib64/libc.so.6  
> 
> Hm, thanks. With those extra keywords, I arrive at:
> 
> https://bugreports.qt.io/browse/QTBUG-100174
> 
> Supposedly that was a kernel bug, fixed in kernel-4.18.0-392.el8
> 
> Earlier you stated, you are using kernel 4.18.0-372.16.1.el8_6.x86_64,
> so yes, this really looks like a candidate for the root cause.
> 
> Any chance for updating?
> 
> (I believe this would actually be the first time 20 years of RKWard,
> that we are directly affected by a kernel-bug.)
> 
> [...]
> 
> > 3)
> > I don't think it's a good idea as a "solution", but I intend to add
> > a "--quirkmode" option to allow to skip the handshake check.
> > Essentially to make sure, that the current problem is the "only"
> > obstacle. Not sure, whether I'll make that in time for the next
> > AppImage build, but I'll keep you posted.  
> 
> Nonetheless, I've added "--quirkmode" as an experimental option, which
> would skip the step of trying to read anything from the child process
> stdout. That step wasn't mentioned as required in the above bug
> report, however, so possibly, that will just delay triggered the bug
> for a few more execution steps.
> 
> Anyway, you may want to try
>   rkward --quirkmode
> with the next AppImage build.
>  
> Regards
> Thomas
> 

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 833 bytes
Desc: OpenPGP digital signature
URL: <http://mail.kde.org/pipermail/rkward-devel/attachments/20220811/9a46d518/attachment.sig>