On 2025-03-05 14:48, Robert Marko wrote:
On Wed, 5 Mar 2025 at 14:43, Martin Schiller <[email protected]> wrote:
CONFIG_MITIGATE_SPECTRE_BRANCH_HISTORY was disabled with commit
6b007d2512ad (treewide: disable spectre mitigation on unaffected Arm64
targets) for all targets in target/linux/mediatek/filogic folder, but
the Arm Cortex-A73 in the Filogic 880 are affected [0].
So should we enable CONFIG_MITIGATE_SPECTRE_BRANCH_HISTORY for this
subtarget again?
If MT7988 is affected then yes.
Another option we maybe should activate is CONFIG_UNMAP_KERNEL_AT_EL0.
According to [0], the cortex a53 and a73 are not affected, but
mediatek/filogic and mediatek/mt7622 are the only arm64 subtargets where
this feature is disabled.
On the other hand, the target.mk in this subtarget folder states it is
CPU_TYPE:=cortex-a53, which is not correct for all the supported
systems
in this subtarget.
Do we have to split this subtarget into filogic830, filogic880, etc.?
This is intentionally done, as A73 core is backwards compatible with
A53 instructions
since it only adds on top of it, so to not introduce yet another
package target A53 ones
are reused.
Regards,
Robert
Regards,
Martin
[0]
https://developer.arm.com/Arm%20Security%20Center/Speculative%20Processor%20Vulnerability
_______________________________________________
openwrt-devel mailing list
[email protected]
https://lists.openwrt.org/mailman/listinfo/openwrt-devel
_______________________________________________
openwrt-devel mailing list
[email protected]
https://lists.openwrt.org/mailman/listinfo/openwrt-devel
