src/sys/i386, branch main

pmap: Do not use PMAP_LOCK_INIT with kernel_pmap

2026-04-01T11:15:42Z

The kernel_pmap lock is a bit special: it does not need the DUPOK flag, and it really belongs to a different lock class. If it belongs to the same class as regular pmap locks, then witness may report warnings when performing UMA allocations under a regular pmap lock, if the allocation triggers a pmap_growkernel() call. Replace instances of PMAP_LOCK_INIT(kernel_pmap) with inline mtx_init() calls to silence some witness warnings for harmless behaviour I see with some uncommitted test programs. Reviewed by: alc, kib MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D56185

routing: Retire ROUTE_MPATH compile option

2026-03-27T20:25:33Z

The ROUTE_MPATH compile option was introduced to test the new multipath implementation. Since compiling it has no overhead and it's enabled by default, remove it. Reviewed by: melifaro, markj Relnotes: yes Differential Revision: https://reviews.freebsd.org/D55884

kernel: Be clearer about what ZSTDIO is for

2026-03-21T07:21:45Z

ZSTDIO is not just for core dumps, and it is now required by ZFS, so move it (and GZIO, which is still just for dumps) to a new “Compression support” section in GENERIC, and add them both to MINIMAL. MFC after: 1 week Reviewed by: imp Differential Revision: https://reviews.freebsd.org/D55940

hwpmc: Remove left over k7, xscale and p4 references pmc

2026-03-07T15:17:43Z

Support for these processors was removed a few years ago, but a few references remain that should be removed. Sponsored by: Netflix Reviewed by: imp Pull Request: https://github.com/freebsd/freebsd-src/pull/2039

x86: change signatures of ipi_{bitmap,swi}_handler() to take pointer

2026-03-07T03:58:48Z

to the frame instead of the frame itself. It is some stretch of the amd64 ABI, and is not easily fullfilled when handlers are called from C and not asm. In particular, the struct frame is passed by value but is modified by callees, with the expectation that the caller will see the modifications. Reviewed by: markj Sponsored by: The FreeBSD Foundation MFC after: 1 week Differential revision: https://reviews.freebsd.org/D55679

sys: Retire le(4)

2026-03-02T19:10:15Z

We retired most obsolete 10 and 10/100 Ethernet NIC drivers in 2019 -- see commits following ebcf740a32ae ("FCP-101: remove obsolete 10 and 10/100 Ethernet drivers.). le(4) was retained with with the note "Emulated by QEMU, alternatives don't yet work for mips64." MIPS has since been removed from the tree and emulators and virtual machines offer many other, more suitable devices. Reviewed by: brooks Relnotes: Yes Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D55516

hwpmc: Initial support for AMD IBS

2026-02-27T21:22:16Z

This patch adds support for AMD IBS. It adds a new class of performance counter that cotains two events: ibs-fetch and ibs-op events. Unlike most existing sampled events, IBS events provide a number of values containing extra information regarding the sample. To support this we use the existing callchain event, and introduce a new flag for multipart payloads. The first 8 bytes of the pc_sample contains a header that defines up to four payloads. Sponsored by: Netflix Reviewed by: imp,mhorne Pull Request: https://github.com/freebsd/freebsd-src/pull/2022

i386,amd64: Explicitly set ECX=0 in do_cpuid() to be future-proof

2026-02-20T15:26:20Z

In principle, do_cpuid() should only be used for CPUID leaves without sub-leaves. Even accessing sub-leaf zero (ECX=0), one must use cpuid_count(ax, 0) rather than cpuid(ax). However, one might assume do_cpuid(ax) is equivalent to cpuid_count(ax, 0), but the old do_cpuid() did not initialize ECX before executing the CPUID instruction. If ECX contained a non-zero value, the instruction could return unexpected results, potentially leading to subtle and hard-to-debug issues, especially in ported code. To be future-proof and to help port code, adjust do_cpuid(ax) to be cpuid_count(ax, 0) to explicitly set ECX=0. It's believed that this change does not fix any real bugs in FreeBSD. See also the DragonFly commit: https://github.com/DragonFlyBSD/DragonFlyBSD/commit/0087a1d163488a57787a9a6431dd94070b1988d4 Signed-off-by: Aaron LI Reviewed by: kib Pull Request: https://github.com/freebsd/freebsd-src/pull/2027

Revert "pcb.h: mark struct pcb to be preserved"

2026-02-09T18:55:57Z

kgdb only uses the marked fields from dumppcb for initial frame reconstruction. This reverts commit 8f23665fed2fbaf4481359b4d2fcdd7b9feb40e3.

pcb.h: mark struct pcb to be preserved

2026-02-09T04:17:16Z

There are programs that depend on this structure (e.g. kernel debuggers) that breaks when the ABI changes. Signed-off-by: Minsoo Choo Reviewed by: kib MFC after: 1 week Differential Revision: https://reviews.freebsd.org/D55149