[qt/qt/qtbase]: Summary of bulk changes made

Git repository change summary for qt/qt/qtbase
Pushed by mirror-service into branch '6.11'.
Changed from 7d433852cc4803692f8d5d86d54e7495878d0a57 to ecac1eb9c02d53a45bf9f55da2a42c6d045b1c4b
Acknowledgement was received that this change introduces only existing code that has been pushed to another public open source repository.

This change contains the following new commits:

Git commit 97e71d6a58005492b548a8aab6f02a80283e6d7f by Alexei Cazacov on 14/08/2026 at 11:22..
Doc: Housekeeping: Fix missing full stops in \brief's and \summarie's

This commit fixes missing full stops in \brief's and \summarie's in
various QDoc files.

Change-Id: I3e82c225fbeee2d2bcf77772bcdc693718e0ab01
Reviewed-by: Topi Reinio <[email protected]>
(cherry picked from commit a714385255cb6b3f6a1717c919301bb3a360f748)
Reviewed-by: Qt Cherry-pick Bot <[email protected]>
(cherry picked from commit 750465c06063d54a0923be5b64919dcb16b5581d)
Reviewed-by: Alexei Cazacov <[email protected]>
https://invent.kde.org/qt/qt/qtbase/-/commit/97e71d6a58005492b548a8aab6f02a80283e6d7f

Git commit 660e36c95b76e76ab10a6cd4f0d34f3ecd3578d4 by Alexei Cazacov on 14/08/2026 at 11:22..
Doc: Housekeeping: Fix the formatting of \sa commands

Change-Id: I893d3507df85995005f285c3362a939877b8f331
Reviewed-by: Mats Honkamaa <[email protected]>
(cherry picked from commit 0ecd5a271694937debb156880be2ba50fd3b40db)
Reviewed-by: Qt Cherry-pick Bot <[email protected]>
(cherry picked from commit 98dce3ff7f57f3235bb8d66094e0a42b643877de)
Reviewed-by: Alexei Cazacov <[email protected]>
https://invent.kde.org/qt/qt/qtbase/-/commit/660e36c95b76e76ab10a6cd4f0d34f3ecd3578d4

Git commit e91a2d3f2ce0acc9dffbcb7f190f586e850b1140 by Qt Cherry-pick Bot (on behalf of Laszlo Agocs) on 14/08/2026 at 11:22..
linuxfb: Fix crash when no DRM output is discovered

output(0) indexed an empty list when all connectors are disconnected
or every output is off or skipped in the KMS config. Fail the screen
initialization instead, which the caller already handles.

Pick-to: 6.8
Change-Id: I56b3da27fa95b3ad35da2dec0aad6148c402e85b
Reviewed-by: Andy Nichols <[email protected]>
(cherry picked from commit c5347d5e4323760b52c2cb66717627668fd854b0)
Reviewed-by: Qt Cherry-pick Bot <[email protected]>
(cherry picked from commit 08564ab834304d4131fd3043eb0d26d79e9e23c6)
https://invent.kde.org/qt/qt/qtbase/-/commit/e91a2d3f2ce0acc9dffbcb7f190f586e850b1140

Git commit b9112fd7f12a92abdc322af722ff7aa66508118b by Qt Cherry-pick Bot (on behalf of Laszlo Agocs) on 14/08/2026 at 11:22..
linuxfb: Fail instead of showing a blank screen on an unknown format

determineFormat() returns Format_Invalid for an unrecognized bitfield
layout or an 8-bit depth, but initialize() reported success anyway. The
resulting null QImage left the display permanently blank, with only a
QPainter warning to explain it.

Pick-to: 6.8
Change-Id: I9460e97e82b3615f99cd65662176966f6ba01f0a
Reviewed-by: Andy Nichols <[email protected]>
(cherry picked from commit a99028d6c76ae0dbd8450981cc8bdeb57a739f31)
Reviewed-by: Qt Cherry-pick Bot <[email protected]>
(cherry picked from commit 182ea6c651f9ef2cfd58ec393fd98b6fc16b4247)
https://invent.kde.org/qt/qt/qtbase/-/commit/b9112fd7f12a92abdc322af722ff7aa66508118b

Git commit d1173347e6755518f8edbd85170a066eaf0c034e by Qt Cherry-pick Bot (on behalf of Laszlo Agocs) on 14/08/2026 at 11:22..
Fix touch device to screen mapping for symlinked device nodes

The table was keyed on the canonical path of the configured touchDevice
but looked up with the device node string the input handler was given.
Those agree when the device comes from discovery, which always yields
/dev/input/eventN, so the common case worked. They differ when the node
is named explicitly by a stable symlink, for example
QT_QPA_EVDEV_TOUCHSCREEN_PARAMETERS=/dev/input/by-id/..., and then the
lookup silently missed and touches went to the primary screen.

Canonicalize on lookup as well, fall back to a literal match, and warn
instead of storing an empty key when the configured path cannot be
resolved.

Pick-to: 6.8
Change-Id: I59bf26e16e8441a4976222ca0e31a9b7337d7a99
Reviewed-by: Andy Nichols <[email protected]>
(cherry picked from commit 9981b42aa1d3c6edd8ea2cd221cf29b86571e2a9)
Reviewed-by: Qt Cherry-pick Bot <[email protected]>
(cherry picked from commit ce06d9f07360967c8b1a0f5241525c3a9955203d)
https://invent.kde.org/qt/qt/qtbase/-/commit/d1173347e6755518f8edbd85170a066eaf0c034e

Git commit 585aaa8ded27d3aac0294d46b3c99b8ec90d2cfe by Qt Cherry-pick Bot (on behalf of Laszlo Agocs) on 14/08/2026 at 11:22..
kms: Do not put uninitialized stack bytes into a DRM mode blob

parseModeline() set only the fields it parses and the four it zeroed
explicitly, leaving name[] and the rest of the caller's stack
drmModeModeInfo untouched. The struct is passed whole to
drmModeCreatePropertyBlob(), so up to 32 uninitialized bytes ended up in
a property blob that other DRM clients can read back.

Zero the whole struct instead, which also covers any field added later.

Pick-to: 6.8
Change-Id: I69afb522de18fb3a928a3edf359c6595895b2491
Reviewed-by: Andy Nichols <[email protected]>
(cherry picked from commit d769af94fd05112873dce5147fdd68d71b3d107d)
Reviewed-by: Qt Cherry-pick Bot <[email protected]>
(cherry picked from commit c0e54ac92341374d888481a4ed3d4993173920e4)
https://invent.kde.org/qt/qt/qtbase/-/commit/585aaa8ded27d3aac0294d46b3c99b8ec90d2cfe

Git commit 967c4f0a883075d21916a54898888ac1cdfc25c5 by Qt Cherry-pick Bot (on behalf of Laszlo Agocs) on 14/08/2026 at 11:22..
linuxfb: Honor the DRM device from the KMS config

The DRM backend passed a hardcoded /dev/dri/card0 to QKmsDevice, and
since the base class only consults the config when the given path is
empty, the "device" key of QT_QPA_EGLFS_KMS_CONFIG was silently ignored
here, unlike in the eglfs KMS integrations. A board whose DRM device is
not card0 could not be configured.

Passing an empty path is not an option, as QKmsDevice then aborts when
there is no config file, so keep card0 as the fallback.

Pick-to: 6.8
Change-Id: Icaef369d80ec42b6090d64b85f9cc1a83d9afc30
Reviewed-by: Andy Nichols <[email protected]>
(cherry picked from commit 8d94499f785f8a5c9be2e6cb1a92fc1fb40e5eee)
Reviewed-by: Qt Cherry-pick Bot <[email protected]>
(cherry picked from commit c0f2183eef07a69358f567aba361c41bf6f5855c)
https://invent.kde.org/qt/qt/qtbase/-/commit/967c4f0a883075d21916a54898888ac1cdfc25c5

Git commit cb10c2207cc8e238b18838fb4a43b652e1046d7b by Qt Cherry-pick Bot (on behalf of Laszlo Agocs) on 14/08/2026 at 11:22..
vkkhrdisplay: Log through a category instead of unconditional qDebug

Every startup of a shipped application printed the display and mode
enumeration and a full plane capability dump to stdout. Move all of it
behind a qt.qpa.vkkhrdisplay category, as the other embedded platform
plugins do. The qWarning calls report real failures and are left alone.

Pick-to: 6.8
Change-Id: I75cfe393a08c5352f89eed8103f8b7a646da5c93
Reviewed-by: Andy Nichols <[email protected]>
(cherry picked from commit 28cf5581b8beafd1685db89991f539ef711d2047)
Reviewed-by: Qt Cherry-pick Bot <[email protected]>
(cherry picked from commit 0b2103ddebec90004a9fc490f184ef26ed3e3ec4)
https://invent.kde.org/qt/qt/qtbase/-/commit/cb10c2207cc8e238b18838fb4a43b652e1046d7b

Git commit 89a44daa6e4e88571610ec9d664737a001379dc4 by Qt Cherry-pick Bot (on behalf of Laszlo Agocs) on 14/08/2026 at 11:22..
eglfs: Validate cursorsPerRow of the cursor atlas

The value was read with toDouble() and converted to int, which is
undefined for anything outside int range, and only checked with a
Q_ASSERT that is compiled out in release builds. A missing or zero entry
then divided by zero two lines later, and a huge one overflowed
Qt::LastCursor + cursorsPerRow.

Read it with toInt(), which yields 0 for a missing key and for anything
that is not a whole number in range, and reject values outside
1..Qt::LastCursor + 1, more cursors per row than there are cursors being
meaningless. Only custom atlases set through QT_QPA_EGLFS_CURSOR are
affected; the built-in one is unchanged.

Task-number: QTBUG-149070
Pick-to: 6.8
Change-Id: Iace91803d1897a957eca11ad247b3351be3b36d7
Reviewed-by: Andy Nichols <[email protected]>
(cherry picked from commit 5bff0cac99125b602dd02ff1ef4e000a737b3b2a)
Reviewed-by: Qt Cherry-pick Bot <[email protected]>
(cherry picked from commit 24908b665f1e9ef5e6ddfc3bb76ba6763b8008c6)
https://invent.kde.org/qt/qt/qtbase/-/commit/89a44daa6e4e88571610ec9d664737a001379dc4

Git commit b2d68de31cada03674242f08e7062c7d2e479951 by Qt Cherry-pick Bot (on behalf of Laszlo Agocs) on 14/08/2026 at 11:22..
kms: Avoid undefined shifts when tracking CRTC usage

m_crtc_allocator is a 32 bit mask, but the CRTC index shifted into
it was bounded only by the number of CRTCs the device reports, and
the shifts were done on a signed 1, so index 31 was already
undefined. Bound the index where it originates, in
crtcForConnector(), which covers all four users, and make the
shifts unsigned.

No device exposes more than 31 CRTCs today, so this is hygiene
rather than an observable fix.

Pick-to: 6.8
Change-Id: Ia532174388f668687dd1d6dfc39ae5c014230c97
Reviewed-by: Andy Nichols <[email protected]>
(cherry picked from commit 957a533bb6d7c824cf2a9b83f125e2a185a547ef)
Reviewed-by: Qt Cherry-pick Bot <[email protected]>
(cherry picked from commit 5a5300e06d35bc5a12eff26007608b0ff638453a)
https://invent.kde.org/qt/qt/qtbase/-/commit/b2d68de31cada03674242f08e7062c7d2e479951

Git commit a05e8a67a3a40236e12c4b535f341b2b2b38b656 by Qt Cherry-pick Bot (on behalf of Laszlo Agocs) on 14/08/2026 at 11:22..
devicediscovery: Check the udev device before using it

udev_device_new_from_syspath() can fail, and the result was passed
straight on to udev_device_get_devnode() and udev_device_unref().

Pick-to: 6.8
Change-Id: Ibe1504967760552885302b3087ac9829c5dc1260
Reviewed-by: Andy Nichols <[email protected]>
(cherry picked from commit 28862699f08654a8136cdb1b7684f44c82680e95)
Reviewed-by: Qt Cherry-pick Bot <[email protected]>
(cherry picked from commit 5c1085c7ba227ee3df55b80408a118f4d0ba2544)
https://invent.kde.org/qt/qt/qtbase/-/commit/a05e8a67a3a40236e12c4b535f341b2b2b38b656

Git commit ecac1eb9c02d53a45bf9f55da2a42c6d045b1c4b by Qt Cherry-pick Bot (on behalf of Laszlo Agocs) on 14/08/2026 at 11:22..
vkkhrdisplay: Use the first suitable plane, not the last

The loop had no break and kept overwriting the selection, so the
highest numbered suitable plane won, while the comment above it
says to use the plane that was found. Keep the first match instead.

Guarding the assignment rather than breaking out preserves the
capability logging for the remaining planes, which is the only
other thing the loop does.

Pick-to: 6.8
Change-Id: Id27f7cd9af1efed0c28c7eaef31432226fc1ad20
Reviewed-by: Andy Nichols <[email protected]>
(cherry picked from commit 863a42b89565de98918c4c5ca741841f29886573)
Reviewed-by: Qt Cherry-pick Bot <[email protected]>
(cherry picked from commit 43cb234206ff83ecb5b46d753505910d59f42835)
https://invent.kde.org/qt/qt/qtbase/-/commit/ecac1eb9c02d53a45bf9f55da2a42c6d045b1c4b