CONFIG_ARM_PMU (Arm Performance Monitor Unit) is a requirement
to use KVM (virtualization) from Linux 5.11+, as the virtualised
guest has virtualized PMU access.
Signed-off-by: Mathew McBride <matt@traverse.com.au>
Currently KASAN is supported but only the generic one. SW-tag and HW-tag
based KASAN have less impact on memory footprint or performance, and are
worth supporting.
Add choice menu for software and hardware Tag-Based KASAN, in addition
to the generic one.
Signed-off-by: Zhen XIN <zhen.xin@nokia-sbell.com>
[Restructure commit message]
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Building it requires gcc >= 10.2 or clang >= 12.
Using sstrip with its -z argument can produce non-working binaries, like
a segfaulting `getrandom`, so don't allow that combination.
Signed-off-by: Andre Heider <a.heider@gmail.com>
sstrip only has one functional arg. Make that a bool option, which can
easily depend on other knobs then.
This is required to be disabled for the mold linker.
Signed-off-by: Andre Heider <a.heider@gmail.com>
In commit b2d1eb717b65 ("generic: 5.15: enable Werror by default for
kernel compile") CONFIG_WERROR=y was enabled and all warnings/errors
reported with GCC 12 were fixed.
Keeping this in sync with past/future GCC versions is going to be uphill
battle, so lets introduce new KERNEL_WERROR config option, enable it by
default only for tested/known working combinations and on buildbots.
References: #12687
Signed-off-by: Petr Štetiar <ynezz@true.cz>
Currently, ipq807x only covers Qualcomm IPQ807x SoC-s.
However, Qualcomm also has IPQ60xx and IPQ50xx SoC-s under the AX WiSoC-s
and they share a lot of stuff with IPQ807x, especially IPQ60xx so to avoid
duplicating kernel patches and everything lets make a common target with
per SoC subtargets.
Start doing that by renaming ipq807x to qualcommax so that dependencies
on ipq807x target can be updated.
Signed-off-by: Robert Marko <robimarko@gmail.com>
armvirt target has been renamed to armsr (Arm SystemReady),
so the config defaults need to be changed as well.
Signed-off-by: Mathew McBride <matt@traverse.com.au>
The nominal partition type for EFI boot partitions is FAT32,
which has a minimum size of 32MiB on a 512-byte-sector block device.
To ensure that the boot partition is created as FAT32 set a size
well above this minimum.
A useful discussion about EFI partition sizes can be found here:
https://superuser.com/questions/1310927/what-is-the-absolute-minimum-size-a-uefi-system-partition-can-be
I have found 128MiB works pretty consistently across both
tools (mkfs.fat) and firmwares (EDKII)
Signed-off-by: Mathew McBride <matt@traverse.com.au>
This adds a separate package for EFI on Arm SystemReady
compatible machines. 32-bit Arm UEFI is supported as well.
It is very similar to x86-64 EFI setup, without the
need for BIOS backward compatibility and slightly
different default modules.
Signed-off-by: Mathew McBride <matt@traverse.com.au>
This set the CONFIG_FRAME_WARN option depending on some target settings.
It will use the default from the upstream kernel and not the hard coded
value of 1024 now.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
BTF mismatch can occur for a separately-built module even when the ABI
is otherwise compatible and nothing else would prevent successfully
loading. Add a new config to control how mismatches are handled. By
default, preserve the current behavior of refusing to load the
module. If MODULE_ALLOW_BTF_MISMATCH is enabled, load the module but
ignore its BTF information.
Signed-off-by: Tianling Shen <cnsztl@immortalwrt.org>
We only use 5.15 kernel. So remove all those unnecessary symbols
referencing 5.10 or 5.15 kernel.
Can be found with:
git grep -E 'LINUX_5_1(0|5)'
Note that we remove the dependency from "sound-soc-chipdip-dac" instead
of removing the complete kernel package. The 5.15 version bump forgot to
delete the "@LINUX_5_10" dependency. The kernel package is still needed
in 5.15 kernel.
Signed-off-by: Nick Hainke <vincent@systemli.org>
Kernel setting CONFIG_IO_URING supports high-performance I/O for file
access and servers, generally for more performant platforms, and adds
~45 KB to kernel sizes. The need for this on less "beefy" devices is
questionable, as is the size cost considering many platforms have kernel
size limits which require tricky repartitioning if outgrown. The size
cost is also large relative to the ~180 KB bump expected between major
OpenWRT kernel releases.
No OpenWrt packages have hard dependencies on this; samba4 and mariadb
can take advantage if available (+KERNEL_IO_URING:liburing) but
otherwise build and work fine.
Since CONFIG_IO_URING is already managed via the KERNEL_IO_URING setting
in Config-kernel.in (default Y), remove it from those target configs
which unconditionally enable it, and update the defaults to enable it
conditionally only on more powerful 64-bit x86 and arm devices. It may
still be manually enabled as needed for high-performance custom builds.
Signed-off-by: Tony Ambardar <itugrok@yahoo.com>
Make it possible to change the kernel configuration option
CONFIG_HARDLOCKUP_DETECTOR from OpenWrt.
Signed-off-by: Hauke Mehrtens <hmehrtens@maxlinear.com>
Without the default value this still causes a missing symbol. Disable by
default as it depends on FUNCTION_ERROR_INJECTION, which is disabled in
the generic config and we don't have a build symbol to enable that.
Fixes: 500c37c56ff6 ("kernel: add missing symbol")
Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>
Sourcing of image/Config.in will not happen
When a target is installed from target/linux/feeds/
Signed-off-by: Prasun Maiti <prasunmaiti87@gmail.com>
Acked-by: Petr Štetiar <ynezz@true.cz>
Set default values for KERNEL_DEBUG_LL and KERNEL_DEBUG_LL_UART_NONE again
as both of these symbols are non visible if KERNEL_EARLY_PRINTK is not
selected and KConfig wont write their value to .config.
This usually is the intended behaviour, but in OpenWrt we are relying on
the KConfig to set these and disable the debug console settings that
multiple targets like mvebu have set in their kernel config.
This was the behaviour before removing all of the "default n" settings
as KConfig by default considers symbols disabled but they are not visible
anymore and thus their value is not set in .config and build system then
later does not override the values from target kernel config.
So, to restore the behaviour to the previous one lets a default value for
KERNEL_DEBUG_LL and KERNEL_DEBUG_LL_UART_NONE.
Fixes: 8bc72ea7be39 ("treewide: strip useless default n Kconfig lines")
Tested-by: Georgi Valkov <gvalkov@gmail.com>
Signed-off-by: Robert Marko <robimarko@gmail.com>
Kconfig docs say:
> The default value deliberately defaults to 'n' in order to avoid
> bloating the build.
Apply this rule everywhere, to avoid more cloning of bad examples
Signed-off-by: Tony Butler <spudz76@gmail.com>
