On 2 Aug 2023, at 5:14, Doug Moore wrote:
The branch main has been updated by dougm:
URL:
https://cgit.FreeBSD.org/src/commit/?id=1a7fcf6d51eb67ee3e05fdbb806f7e68f9f53c9c
commit 1a7fcf6d51eb67ee3e05fdbb806f7e68f9f53c9c
Author: Doug Moore <do...@freebsd.org>
AuthorDate: 2023-08-02 03:12:00 +0000
Commit: Doug Moore <do...@freebsd.org>
CommitDate: 2023-08-02 03:12:00 +0000
vm_phys_enqueue_contig: handle npages==0
By letting vm_phys_enqueue_contig handle the case when npages ==
0,
the callers can stop checking it, and the compiler can stop
zero-checking with every call to ffs(). Letting
vm_phys_enqueue_contig
call vm_phys_enqueue_contig for part of its work also saves a few
bytes.
The amd64 object code shrinks by 80 bytes.
Reviewed by: kib
Differential Revision: https://reviews.freebsd.org/D41154
I suspect this commit caused my machine to fail to boot:
WARNING: WITNESS option enabled, expect reduced performance.c
data=0x180 data=0x
panic: vm_phys_enq_range: page 0xfffffe0000006180 and npages 4294967040
are misaligned
cpuid = 0onfigured modules...
time = 1rnel/if_bnxt.ko size 0x3efe8 at 0x1f50000
KDB: stack backtrace:ize 0x6f2698 at 0x1f8f000
db_trace_self_wrapper() at db_trace_self_wrapper+0x2b/frame
0xffffffff82c03e20
vpanic() at vpanic+0x149/frame 0xffffffff82c03e70
panic() at panic+0x43/frame 0xffffffff82c03ed00
vm_phys_enq_range() at vm_phys_enq_range+0x11d/frame 0xffffffff82c03ee0
vm_phys_enqueue_contig() at vm_phys_enqueue_contig+0xdd/frame
0xffffffff82c03f20
vm_page_startup() at vm_page_startup+0xf3d/frame 0xffffffff82c03f90
vm_mem_init() at vm_mem_init+0x1a/frame 0xffffffff82c03fa0
mi_startup() at mi_startup+0x1f1/frame 0xffffffff82c03ff0
btext() at btext+0x23
KDB: enter: panic
This is sufficiently early in boot that I don’t get a core dump, so I
hope the backtrace helps you.
I’ll back out the commit locally and confirm.
Best regards,
Kristof
