- User Since
- May 10 2014, 2:21 PM (184 w, 4 d)
Tue, Nov 21
Sun, Nov 19
Are we expecting to run armv6 code, or should we just support armv7?
This will allow userspace to switch between AArch64 and AArch32 via sigreturn. sys_sigreturn has checks on the PSR bits to ensure they are safe to install that this will break.
I split the syncronous exception out into a new handler in my compat32 patch as it allows further sanity checks, e.g. that the kernel thinks the current process should be in 32-bit mode.
Sat, Nov 18
It feels like this should be 2 drivers. One to talk with the ac100, and a second jsut for the rtc bits.
Fri, Nov 17
Thu, Nov 16
- Be stricter when we use the driver
- Update style base on feedback from kib
Tue, Oct 31
Oct 22 2017
Oct 20 2017
I think the architecture allows us to issue the tlb invalidation, then later execute the dsb & isb instructions as long as we don't try to the old mapping, however will need to check with ARM. We should also just invalidate all entries in pmap_invalidate_range if the length is too long, however in this case it may be better to stop invalidating until after we have removed all the pagetable entries.
Oct 19 2017
Oct 13 2017
Oct 12 2017
Should we enable it on ARMv4/ARMv5 (i.e. old ARM systems)? If so you should add it to sys/arm/conf/std.arm.
Oct 10 2017
Oct 9 2017
Oct 8 2017
- Add a public domain comment to fpu.h
- Move _efirt to an aarch64 specific section in modules/Makefile
Oct 6 2017
Oct 5 2017
Oct 3 2017
Oct 2 2017
Mostly style issues.
Oct 1 2017
Sep 28 2017
Sep 24 2017
Move the MI file under sys/dev/efidev
Sep 22 2017
I left the name as it all architectures that currently support EFI RT use a 1:1 map.
- Move the locking macros to a MD header
- Ddd subr_efi.c to the module Makefile
Sep 21 2017
Sep 15 2017
Sep 14 2017
Sep 13 2017
When uploading a diff through the website can you include context. There are commands you can use at https://wiki.freebsd.org/Phabricator#Create_a_Revision_via_Web_Interface
Sep 12 2017
I have a slight concern about a non-current thread being run at the same time as this leading to a potential data race.
Sep 11 2017
On ARM platforms I expect we'll need more bits. Linux already has AT_HWCAP2, it might pay to include it from the start.
Sep 8 2017
Sep 7 2017
Sep 6 2017
Sep 5 2017
Sep 3 2017
Sep 1 2017
Based on the Linux source the A73 value is 0xD09. It might pay to also add it at the same time.
Aug 31 2017
Aug 26 2017
Aug 22 2017
Aug 21 2017
Aug 19 2017
Aug 13 2017
Can you update this with more context.
Aug 12 2017
I think Semihalf was only interested in the kernel debugger when they added this.
Aug 11 2017
Yes. I'm not sure why I made the save KPI look like it does. I think curthread is correct at the point it's called in cpu_switch(), however it would pay to check that.
On arm64 if a thread has used the vfp unit we will store the state on context switch. This means we should only need to call vfp_save_state when we need access to the fpu registers from the current thread. I don't think fill_fpregs has been tested as lldb didn't have vfp support when we did the initial port.
Where is this being called from? I think the bug is calling vfp_save_state on a non-current thread.
Aug 10 2017
Aug 7 2017
There is now access to the platform object from outside of platform.c. You could add a method to platform_pl310_if.m to handle this. Each SoC can then implement this if needed.