Support non-static mappings, allocating space from kernel_map.
Still TODO: platform-specific hook for a vmem arena that supplies
the VAs for a given bus.
Add bus_space_mmap() to the simple implmentation, and provide a default
implementation that returns failure; if a platform wants to support this,
it will need to provide a _bus_space_mmap() implementation that matches
the address ranges that are to be supported for this operation.
Update to 2026b, previous was 2026a
Release 2026b - 2026-04-22 23:06:43 -0700
Changes to code
zic no longer mishandles a last transition to a new time type.
zic no longer overflows a buffer when generating a TZ string like
"PST-167:59:58PDT-167:59:59,M11.5.6/-167:59:59,M12.5.6/-167:59:59",
which can occur with adversarial input. (Thanks to Naveed Khan.)
zic no longer generates a longer TZif file than necessary when
an earlier time zone abbreviation is a suffix of a later one.
As a nice side effect, zic no longer overflows a buffer when given
a long series of abbreviations, each a suffix of the next.
(Buffer overflow reported by Arthur Chan.)
zic no longer overflows an int when processing input like `Zone
[4 lines not shown]
For machines where the real time clock is off or not existing:
- offer a config menu to set date and time
- at build time burn the unix time of the build into the binary (or the
reproducable build timestamp)
- when the current clock claims we have a date older than two days before
the embedded build time, show an error message and go to the date and
time setup step automatically
arm32/bus_dma: fix interger overflow on 32 bit kernels
On Raspberry Pi 4B with BCM2711, dr_busbase is 0xc0000000 and
dr_len is 0x40000000, so "dr->dr_busbase + dr->dr_len" causes
wraparound.
Fixes panic on Raspberry Pi 4B during attaching vcmbox0:
> panic: _bus_dma_busaddr_to_paddr: curaddr 0xc5b6c000not in range
Re-factor all of the badaddr() variants into a single badaddr_read()
that takes an access size and can return a value. Add a badaddr_write()
that similarly takes a size.
when '-i[rR]' is given, prompt the user for examination of directories as well as for removal
Previously, interactively and recursively removing a directory hierarchy
did not iteratively prompt the user for confirmation for both descending
into as well as removing directories.
This uses the patch suggested by rvp at SDF.ORG on tech-userlevel back in
February to more closely match the behavior of macOS and Linux as well
as user expectations.
Diable a few options unlikely to be all that useful on a NeXT:
- COMPAT_NOMID, COMPAT_SUNOS
- file-system MFS (TMPFS is preferred)
- dse, se (SCSI Ethernet)
- sl (SLIP), ppp / pppoe
Shrinks the kernel by ~167KB.
