Also explicitly initialize sc->sc_status = HIL_STATUS_BUSY in hil_attach().
Previously hil_intr(9) returned immediately during 'cold', but
all interrupts are enabled at the end of configure() (actually
in MD cpu_configure()) and cold is cleared in confiugre2()
after configure(), so there is a small window when hil interrupts
can be triggered during cold.
It looks there is no problem to process hil_intr() before
hil_attach_deferre() is called via configure2() because
we already check 'sc->sc_status != HIL_STATUS_BUSY' on
processing a kthread.
Note this seems also to appease the similar panic on mame's hp9k370
emulation (though mame's emulation around DMAC looks still incomplete).
Should be pulled up to netbsd-10 and netbsd-9.
this reduces the size of the installed files by over half in most cases,
though the debug set size doesn't really change much (which looks like
close to 1GB of space on amd64 with xdebug installed, similar on arm64,
and about 600MB without xdebug.)
tested by running GDB on a few things, seems just as functional, on amd64,
arm64, and slightly on riscv64.
(first attempt for this feature used "gcc -gz=zlib", but that ends up
making CTF unhappy, but fortunately this works in binutils to create
the .debug files separate to any ctf usage of the main file.)
Otherwise, a pending signal may be delivered on the wrong stack when
we restore the signal mask.
While here:
- Move the botched sp and lr tests earlier.
PR lib/57946
Normally I don't like CHECK(A && B) and favour CHECK(A); CHECK(B)
instead, so you can see which one failed, but in this case the inputs
are displayed anyway so we don't lose anything unless the
floating-point printer is broken.
1. Exercise more edge cases around the largest exponent that can
represent non-integers.
2. - Verify modf can handle all cases modff can.
- Veriy modf can handle some cases modff can't.
- Verify modfl can handle all cases modf or modff can.
- If long double isn't just double, verify modfl can handle some
cases modf and modff can't.
3. Check infinities and NaN.
4. Verify oddness, i.e., modf(-x) = -modf(x).
5. Display more useful diagnostics in case any of this goes wrong.
While here, use CWARNFLAGS.clang instead of an explicit conditional.
In file included from 110_blorp_exec.c:33:
In file included from /home/source/ab/HEAD-llvm/xsrc/external/mit/MesaLib/dist/src/intel/blorp/blorp_genX_exec.h:27:
In file included from /home/source/ab/HEAD-llvm/xsrc/external/mit/MesaLib/dist/src/intel/blorp/blorp_priv.h:30:
/home/source/ab/HEAD-llvm/xsrc/external/mit/MesaLib/dist/src/compiler/nir/nir.h:3840:3: error: redefinition of typedef 'nir_shader' is a C11 feature [-Werror,-Wtypedef-redefinition]
} nir_shader;
^
/home/source/ab/HEAD-llvm/xsrc/external/mit/MesaLib/dist/src/intel/compiler/brw_compiler.h:41:27: note: previous definition is here
typedef struct nir_shader nir_shader;
^
We should arrange to just have LDBL_NBIT unconditionally defined in the
appropriate MD header file, and make LDBL_IMPLICIT_NBIT an alias for
LDBL_NBIT==0. But for now this will do.
Seems likely that there are other parts of libm which would benefit
from being defined unconditionally in terms of ieee_ext_u, with
ieee_ext_u as an alias for ieee_double_u, in this scenario. But I
haven't gone looking yet.
This strtol(3) was introduced in rev 1.12 for PR/57909 after netbsd-10,
but it returns LONG_MAX (0x7FFFFFFF) for 0xFFF00000 on ILP32 hosts and
the wrong loadpoint causes "NOT ENOUGH MEMORY" error by the BOOTROMs
on loading uboot.lif on (at least) my 9000/360 and 9000/425t.
EXT_FRACBITS, the number of bits in the _binary encoding_ that stores
the trailing significand field, is never 113. In IEEE 754 binary128,
it is 112, even though there are 113 bits of precision in the set of
floating-point numbers -- the leading 1 bit is implicit in binary128.
So ld128 platforms like aarch64 and sparc64 were skipping the real
definition and just defining rintl as an alias for rint, which is
wrong.
In contrast, LDBL_MANT_DIG, the number of bits of precision in the set
of floating-point numbers (p, in IEEE 754 parlance), is 113 in IEEE 754
binary128. This is also the constant used in FreeBSD libm anyway. So
let's just use that instead of trying to translate it to our private
EXT_FRACBITS (not defined in FreeBSD) with a fencepos terror. And
delete the buggy rintl=rint alias.
PR lib/58054
Revert the part of rev 1.32 (reapplying "Do away with separate pool_cache
for some kernel objects") that changed the memory allocation for radixtree
nodes from PR_NOWAIT to KM_SLEEP as part of changing from a pool to kmem.
uvm_pageinsert_tree() calls into the radixtree code while holding
the object's vmobjlock, but that same lock is taken by the pagedaemon
in the process of reclaiming pages, and if the pagedaemon happens to
choose the same object to reclaim from that uvm_pageinsert_tree()
is being called on, then these two threads will deadlock.
The previous code already handled memory allocation failures
in uvm_pageinsert_tree() so we can simply change it back to nosleep.
Fixes a hang reported by simonb@, and the fix was also tested by him.
riastradh