OK, this can go in as is for now.

In D5838#124795, @skra wrote:

Thanks that you are doing this. I thought that I would do this later, as there was no immediate need to do it in D5730. Thus, thanks again and sorry for trouble.

Thanks that you are doing this. I thought that I would do this later, as there was no immediate need to do it in D5730. Thus, thanks again and sorry for trouble.

I'm happy with it being as-is for now.

thanks!

cool!

Updated diffs to reflect kan@'s comments.

Update diff.

In D5730#123890, @andrew wrote:

This feels a little over engineered.

This feels a little over engineered.

(1) An ipi argument was added to PIC_IPI_SEND method to explicitly specify which ipi should be send. Some controllers use one ISRC for all IPIs (e.g. RPI2 uses a mailbox interrupt), so this simplifies things for them.
(2) Update to current was done.

Change intr_isrc_dispatch() return value from void to int to signal that an interrupt (ISRC) was not handled. Even ISRC with handlers does not guarantee that it happens. So, it's only a PIC responsibility now to act with a stray interrupt - e.g., disable interrupt and does EOI. Note that calling of PIC_DISABLE_INTR() from INTRNG was not enough.

Remove unused ipi_next_num variable.
Include opt_acpi.h to subr_intr.c.
Remove FDT ifdefs from mpic.c as the driver is only for FDT.

Yes, the superpage size should be 2M (PDRSIZE) for mips64. This looks correct. I am not sure it matters for mips32 given there is no superpage support for it yet. The way it is done for mips64 (see https://reviews.freebsd.org/D2535 ) may not work for mips32 given the limited number of available PTE bits. Of course, the mips64 superpage implementation was designed to support multiple superpage sizes (more than just 2M) although only one superpage size is currently implemented. I imagine with mips32 only one superpage size would be supported.

CheriBSD change for reference: https://github.com/CTSRD-CHERI/cheribsd/commit/7ed51f1f4ff2ea4c7ba1cebc101d9dd6e26f3844

In D5312#116404, @imp wrote:

This looks ready to go. If you happen to know why MIPS is different, it would be cool if you could address that in the comment. If not, then no big deal. Differences like that will be a mystery later if we don't document what we know today.

go!

This looks like just code motion, though if there's other subtle changes they are lost on me.

This looks ready to go. If you happen to know why MIPS is different, it would be cool if you could address that in the comment. If not, then no big deal. Differences like that will be a mystery later if we don't document what we know today.

Done in https://svnweb.freebsd.org/base?view=revision&revision=296080

Done with https://svnweb.freebsd.org/base?view=revision&revision=296079

Addressed review comments.
I decided to drop sys/boot/common/self_reloc.c changes, as currently self relocation is not supported by the MIPS ubldr.

The actual MIPS ubldr support will come with D5313.

Thanks for the reviews. I've responded to some of the comments and will take care of them and submit a new version.

I'm unclear how this supports 64-bit mips. Perhaps a quick note about what I'm missing would be in order?

Cleaned up some leftover issues.

I am not sure whether more reviewers are required?

Will resubmit in a different revision.

Ok, I'll test/commit this one first, as i think it's a pre-requisite for the future patches.

Patch updated.

Patch updated.

Patch updated to use sys/sys/intr.h for the common parts between ARM and MIPS.
Also, FDT-related headers are now conditionally included in order to allow MIPS systems that do not support FDT to move to INTRNG should they wish to.

ok, so how about we (read: me) split out the arm intrng definitions into sys/sys/ and then you can update your patch.

cool! ok, committing this now.

Made this revision dependent on D5196 only because the diff was created after the diff for D5196, so they may be better applied in the same order if accepted.

Diff updated.

Renamed sys/mips/mips/intr.c to sys/mips/mips/mips_pic.c per Alexander Kabaev's request (sorry, I must have misunderstood what you meant originally).

I wasn't quite clear, my apologies. I still want to have mips_pic.c, not intr.c in the source. I only wanted intr.c IFF you were to decide to move cpu_establish_hard_intr and friends into a separate file and make them with arbitrary root PICm only then I called for the intr.c file. In other words, rename your intr.c to mips_pic.c for now.