buildroot.busybox.net archive mirror
 help / color / mirror / Atom feed
* [Buildroot] [Bug 15461] New: QtVirtualKeyboard segfaults
@ 2023-03-20  2:43 bugzilla
  2023-03-21 13:23 ` [Buildroot] [Bug 15461] " bugzilla
                   ` (7 more replies)
  0 siblings, 8 replies; 9+ messages in thread
From: bugzilla @ 2023-03-20  2:43 UTC (permalink / raw)
  To: buildroot

https://bugs.busybox.net/show_bug.cgi?id=15461

            Bug ID: 15461
           Summary: QtVirtualKeyboard segfaults
           Product: buildroot
           Version: 2023.02
          Hardware: Other
                OS: Linux
            Status: NEW
          Severity: normal
          Priority: P5
         Component: Other
          Assignee: unassigned@buildroot.uclibc.org
          Reporter: info@resilience-theatre.com
                CC: buildroot@uclibc.org
  Target Milestone: ---

QtVirtualKeybaord program which works on 2022.02.9 segfaults on 2023.02.

My working defconfig for Pinephone (ARM) fails to run QT/QML program which
incldues virtualkeyboard after upgrade to 2023.02.

Steps to reproduce: 

* Clone 2023.02 buildroot
* Arm defconfig with all standard QT packages (5.15.8)
* Setup QT Creator with toolchain and sysroot to BR produced
* Create sample QML application and include QtVirtualKeyboard
* Try to run program

Results on 2023.02:

ASSERT: "targetObjectIndex >= 0" in file
../../include/QtQml/5.15.8/QtQml/private/../../../../../src/qml/qml/qqmlpropertycachecreator_p.h,
line 828

On 2022.02.9 this works without and erros.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
buildroot mailing list
buildroot@buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

^ permalink raw reply	[flat|nested] 9+ messages in thread

* [Buildroot] [Bug 15461] QtVirtualKeyboard segfaults
  2023-03-20  2:43 [Buildroot] [Bug 15461] New: QtVirtualKeyboard segfaults bugzilla
@ 2023-03-21 13:23 ` bugzilla
  2023-03-21 13:42 ` bugzilla
                   ` (6 subsequent siblings)
  7 siblings, 0 replies; 9+ messages in thread
From: bugzilla @ 2023-03-21 13:23 UTC (permalink / raw)
  To: buildroot

https://bugs.busybox.net/show_bug.cgi?id=15461

--- Comment #1 from Thomas Petazzoni <thomas.petazzoni@bootlin.com> ---
Buildroot 2022.02 defaults to using GCC 10.x.
Buildroot 2023.02 defaults to using GCC 11.x, but still supports GCC 10.x.

Could you try staying with Buildroot 2023.02, but use GCC 10.x instead of GCC
11.x ?

My colleague Alexis Lothoré is also chasing down a Qt5 problem with Buildroot
2023.02 (but using the Bootlin toolchain), and reverting to an older toolchain
solved the problem, which seems to point to a GCC issue.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
buildroot mailing list
buildroot@buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

^ permalink raw reply	[flat|nested] 9+ messages in thread

* [Buildroot] [Bug 15461] QtVirtualKeyboard segfaults
  2023-03-20  2:43 [Buildroot] [Bug 15461] New: QtVirtualKeyboard segfaults bugzilla
  2023-03-21 13:23 ` [Buildroot] [Bug 15461] " bugzilla
@ 2023-03-21 13:42 ` bugzilla
  2023-03-22  0:53 ` bugzilla
                   ` (5 subsequent siblings)
  7 siblings, 0 replies; 9+ messages in thread
From: bugzilla @ 2023-03-21 13:42 UTC (permalink / raw)
  To: buildroot

https://bugs.busybox.net/show_bug.cgi?id=15461

--- Comment #2 from Alexis Lothoré <alexis.lothore@bootlin.com> ---
Hi. I indeed encounter the same kind of issue in a Qt Quick project on Armv7
target
* with Buildroot 2023.02 and Bootlin Toolchain 2022.08, Qt project crashes deep
in qqmlcrashpropertycache.h (in an inlined function)
* trying to build without optimizations (to allow debugging), the application
does not crash anymore
* switching back to Bootlin toolchain 2021.11, the application works, even with
optimizations enabled
* I tried to switch to buildroot internal toolchain BUT with components
versions (gcc, glibc, binutils, kernel headers) from bootlin toolchain 2022.08
=> it still crashes, so it is probably not the Bootlin build of the toolchain
which is the culprit

-- 
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
buildroot mailing list
buildroot@buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

^ permalink raw reply	[flat|nested] 9+ messages in thread

* [Buildroot] [Bug 15461] QtVirtualKeyboard segfaults
  2023-03-20  2:43 [Buildroot] [Bug 15461] New: QtVirtualKeyboard segfaults bugzilla
  2023-03-21 13:23 ` [Buildroot] [Bug 15461] " bugzilla
  2023-03-21 13:42 ` bugzilla
@ 2023-03-22  0:53 ` bugzilla
  2023-03-22  2:08 ` bugzilla
                   ` (4 subsequent siblings)
  7 siblings, 0 replies; 9+ messages in thread
From: bugzilla @ 2023-03-22  0:53 UTC (permalink / raw)
  To: buildroot

https://bugs.busybox.net/show_bug.cgi?id=15461

--- Comment #3 from info@resilience-theatre.com ---
Thank you for good comments. I've been trying to play with optimization level,
any comments how and what you do on optimization in details would be
appreciated. 

I am using buildroot internal toolchain, but will try next that compiler
version change. Indeed I remember seeing some similar things years back which
was triggered by GCC version. If anything, I will update results.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
buildroot mailing list
buildroot@buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

^ permalink raw reply	[flat|nested] 9+ messages in thread

* [Buildroot] [Bug 15461] QtVirtualKeyboard segfaults
  2023-03-20  2:43 [Buildroot] [Bug 15461] New: QtVirtualKeyboard segfaults bugzilla
                   ` (2 preceding siblings ...)
  2023-03-22  0:53 ` bugzilla
@ 2023-03-22  2:08 ` bugzilla
  2023-03-22 13:51 ` bugzilla
                   ` (3 subsequent siblings)
  7 siblings, 0 replies; 9+ messages in thread
From: bugzilla @ 2023-03-22  2:08 UTC (permalink / raw)
  To: buildroot

https://bugs.busybox.net/show_bug.cgi?id=15461

info@resilience-theatre.com changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|---                         |WORKSFORME
             Status|NEW                         |RESOLVED

--- Comment #4 from info@resilience-theatre.com ---
I can confirm that changing GCC (buildroot toolchain) to 10.4.0 fixed my
qtvirtualkeyboard error reported in this bug.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
buildroot mailing list
buildroot@buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

^ permalink raw reply	[flat|nested] 9+ messages in thread

* [Buildroot] [Bug 15461] QtVirtualKeyboard segfaults
  2023-03-20  2:43 [Buildroot] [Bug 15461] New: QtVirtualKeyboard segfaults bugzilla
                   ` (3 preceding siblings ...)
  2023-03-22  2:08 ` bugzilla
@ 2023-03-22 13:51 ` bugzilla
  2023-03-22 16:36 ` bugzilla
                   ` (2 subsequent siblings)
  7 siblings, 0 replies; 9+ messages in thread
From: bugzilla @ 2023-03-22 13:51 UTC (permalink / raw)
  To: buildroot

https://bugs.busybox.net/show_bug.cgi?id=15461

--- Comment #5 from Alexis Lothoré <alexis.lothore@bootlin.com> ---
> Thank you for good comments. I've been trying to play with optimization level, any comments how and what you do on optimization in details would be appreciated.

If you are still interested in this option, I simply added the following to
package/qt5/qt5declarative/qt5declarative.mk:

QT5DECLARATIVE_CONF_OPTS += QMAKE_CXXFLAGS=-O0

I also tried with GCC12, with which the crash is still occuring, so the issue
is still in GCC12. I guess next step will be to try to isolate a qt quick
sample able to reproduce the issue to track the exact issue and/or share it
with GNU people

-- 
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
buildroot mailing list
buildroot@buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

^ permalink raw reply	[flat|nested] 9+ messages in thread

* [Buildroot] [Bug 15461] QtVirtualKeyboard segfaults
  2023-03-20  2:43 [Buildroot] [Bug 15461] New: QtVirtualKeyboard segfaults bugzilla
                   ` (4 preceding siblings ...)
  2023-03-22 13:51 ` bugzilla
@ 2023-03-22 16:36 ` bugzilla
  2023-06-21 13:08 ` bugzilla
  2023-06-21 14:46 ` bugzilla
  7 siblings, 0 replies; 9+ messages in thread
From: bugzilla @ 2023-03-22 16:36 UTC (permalink / raw)
  To: buildroot

https://bugs.busybox.net/show_bug.cgi?id=15461

--- Comment #6 from Peter Seiderer <ps.report@gmx.net> ---
With (at least) gcc-12.2.0 on RPi4 (32-bit) changing from

    BR2_OPTIMIZE_S=y

to

    BR2_OPTIMIZE_2=y

seems sufficient...

Or maybe it is worth a try to bump the versions (as the last on is from
2022-05-17, see [1])...

[1]
https://git.buildroot.net/buildroot/commit/package/qt5?id=558d62490b4486f47248c63aa0eceaaa27e3069c

-- 
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
buildroot mailing list
buildroot@buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

^ permalink raw reply	[flat|nested] 9+ messages in thread

* [Buildroot] [Bug 15461] QtVirtualKeyboard segfaults
  2023-03-20  2:43 [Buildroot] [Bug 15461] New: QtVirtualKeyboard segfaults bugzilla
                   ` (5 preceding siblings ...)
  2023-03-22 16:36 ` bugzilla
@ 2023-06-21 13:08 ` bugzilla
  2023-06-21 14:46 ` bugzilla
  7 siblings, 0 replies; 9+ messages in thread
From: bugzilla @ 2023-06-21 13:08 UTC (permalink / raw)
  To: buildroot

https://bugs.busybox.net/show_bug.cgi?id=15461

--- Comment #7 from stijn.vermeir+bugzilla@dioss.com ---
(In reply to Peter Seiderer from comment #6)
I tried a build today with Qt versions bumped like this:
qtbase            29400a683f96867133b28299c0d0bd6bcf40df35
qtdeclarative     540c4e4a5def8c350a49bb68380b787ae62490c6
qtsvg             7e6a3b38b2e898d90cef68a146cd36fda22e4363
qtvirtualkeyboard 72373522141dd3206183eb5fa56ae1c36a6d4c2b

I got these commit ids from here:
https://invent.kde.org/qt/qt/qt5/-/commit/da2e0a59fa38644a95047fe389d955dc71b31fc8

I was using buildroot 2023.02.2 internal toolchain for this with gcc-12.3.0.

The virtual keyboard still segfaults.

I am also trying with gcc-10 now to see if this fixes the segfault here.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
buildroot mailing list
buildroot@buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

^ permalink raw reply	[flat|nested] 9+ messages in thread

* [Buildroot] [Bug 15461] QtVirtualKeyboard segfaults
  2023-03-20  2:43 [Buildroot] [Bug 15461] New: QtVirtualKeyboard segfaults bugzilla
                   ` (6 preceding siblings ...)
  2023-06-21 13:08 ` bugzilla
@ 2023-06-21 14:46 ` bugzilla
  7 siblings, 0 replies; 9+ messages in thread
From: bugzilla @ 2023-06-21 14:46 UTC (permalink / raw)
  To: buildroot

https://bugs.busybox.net/show_bug.cgi?id=15461

--- Comment #8 from stijn.vermeir+bugzilla@dioss.com ---
Indeed here also no segfault anymore after switching to gcc 10!

-- 
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
buildroot mailing list
buildroot@buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

^ permalink raw reply	[flat|nested] 9+ messages in thread

end of thread, other threads:[~2023-06-21 14:47 UTC | newest]

Thread overview: 9+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2023-03-20  2:43 [Buildroot] [Bug 15461] New: QtVirtualKeyboard segfaults bugzilla
2023-03-21 13:23 ` [Buildroot] [Bug 15461] " bugzilla
2023-03-21 13:42 ` bugzilla
2023-03-22  0:53 ` bugzilla
2023-03-22  2:08 ` bugzilla
2023-03-22 13:51 ` bugzilla
2023-03-22 16:36 ` bugzilla
2023-06-21 13:08 ` bugzilla
2023-06-21 14:46 ` bugzilla

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).