It should be lost during merge from upstream.
We may introduce a similar hack again, if it is *really* required;
inflate_fast() may be dropped by using slow path unconditionally.
DOS command line arguments are provided as struct psp at 0x0000;
see doscommain.c.
Recent versions of gcc and clang are clever enough to optimize code
block involving NULL dereference into ud2 insn.
Sprinkle -fno-delete-null-pointer-checks to doscommain.c to
prevent this behavior.
Note that dosboot.com for netbsd-9 and later was broken due to
this ``over optimization''. gcc 5.5.0 and clang 4.0.0 in netbsd-8
generate correct codes without this workaround.
XXX
Are there still use cases for dosboot.com? Does anyone want to
boot NetBSD from real-mode DOS in 2023?
This update makes this driver more than just an example and allows for:
o More than one pin to be attached to a gpioirq instance. That is,
the mask parameter can be greater than 0x01 now.
o A /dev/gpioirqN device that allows GPIO pin interrupts to be
transported into userland. This is a device that can be opened for
reading with a simple fixed output indicating the device unit, pin
number and current pin state.
This update was used as part of a physical intrusion detection system
where multiple switches (i.e. window magnetic reed switches and etc.)
are tied to a bunch of GPIO inputs with userland software that reacts
to the pins changing state.
Print the correct input, and print the rounding mode for clarity so
you don't have to cross-reference it by line number.
PR port-mips/57680
XXX pullup-10
Pull request: https://github.com/maleic1618/NetBSD/pull/1:
"""
This PR fixes some bugs and MP-ifies ena(4), including RSS.
On t3.small instance, its forward throughput has grown as followings:
64 bytes 256 bytes
-current 1flow 21.58 110.51
-current 8flow 28.27 123.19
patched 1flow 37.20 142.21
patched 8flow 92.26 366.85
(Unit: Mbps, NET_MPSAFE enabled, Using https://github.com/iij/ipgen)
It also worked well on a aarch64 machine.
Addendum: I forgot to mention that this PR is reviewed ozaki-r@n.o.
"""
Code contributed by KUSABA Takeshi <t-kusaba@iij.ad.jp>
Reviewed also by ozaki-r@n.o
(it can and should be better in the ure(4) driver, but the support
is not yet ported over.)
also add a note about USB 3 speeds. this device is 2.5G, and i've
seen it do just over 1G speeds. other 1G devices also use USB 3
speeds not mentioned here.
bump date.
Except gcc doesn't implement this pragma, so make it conditional.
And clang only supports it on some architectures, so just leave it
out for now with a comment about why.
PR port-mips/57680
XXX pullup-10
At least for addition operations, anyway.
Somewhat redundant with the test t_fe_round added by maya@ but this
gives two minimal pairs to easily diagnose exactly what the rounding
mode is when the wrong one was selected.
PR port-mips/57680
XXX pullup-10
Tested on NWS-3260, which was sent from ryo@'s belongins and repaired
by me, and also tested on my NWS-3470D, including Xorg mono server.
(Note X.org server on NEWS machines requires keymap modifications)
Also add proper initialization on consinit() in NMB-253 case.
Worth to pull up to netbsd-10 and netbsd-9.
Add support in dtrace for SMAP, so that actions like copyinstr() work.
It would be better if dtrace could use the SMAP_* hotpatch macros directly,
but the hotpatching code does not currently operate on kernel modules,
so we'll use some tiny functions in the base kernel for now.
1. Don't touch the obsolete IFF_ALLMULTI.
2. Set ETHER_F_ALLMULTI if we're accepting all multicast addresses.
3. If any multicast entry range is not a single address, accept all
multicast addresses.
- Can't touch if_flags without IFNET_LOCK.
- Can't take IFNET_LOCK in SIOCADDMULTI/SIOCDELMULTI path.
Instead, cache IFF_PROMISC and IFF_ALLMULTI on if_init under a lock we
can take in this path.
XXX Is IFF_ALLMULTI relevant any more? Hasn't it been moved to
ethercom flags?
XXX Should not take sc_lock around if_init/stop -- IFNET_LOCK is
enough. Should narrow scope of sc_lock to be just tick/mii/multi
stuff.
Right now, Substring_Words terminates each word with a '\0', but that's
an implementation detail that is not required by the interface, so don't
rely on it.
rin