based on v-spec 0.7.1, support
sections: 14/15.3 ~ 15.4
element size: 32
Signed-off-by: Bruce Hoult <bruce@hoult.org>
Signed-off-by: Chih-Min Chao <chihmin.chao@sifive.com>
Signed-off-by: Dave Wen <dave.wen@sifive.com>
based on v-spec 0.7.1, support
section: 7
element size: 8/16/32/64
Signed-off-by: Bruce Hoult <bruce@hoult.org>
Signed-off-by: Chih-Min Chao <chihmin.chao@sifive.com>
Signed-off-by: Dave Wen <dave.wen@sifive.com>
Signed-off-by: Zakk Chen <zakk.chen@sifive.com>
based on v-spec 0.7.1, support
sections: 12/13/15.1 ~ 15.2/16/17
element size: 8/16/32/64
support ediv: 1
Signed-off-by: Bruce Hoult <bruce@hoult.org>
Signed-off-by: Chih-Min Chao <chihmin.chao@sifive.com>
Signed-off-by: Dave Wen <dave.wen@sifive.com>
1. add integer signed add/subu and unsigend add/sub saturation function
2. merge these with mulhi helper
Signed-off-by: Chih-Min Chao <chihmin.chao@sifive.com>
the default vector parameters are defined in configuration time but can
be changed throught command-line option
Signed-off-by: Dave Wen <dave.wen@sifive.com>
1. configure option "--with-varch"
the option defines the default u-arch implementatiton-decided parameter
VLEN: vector register length in bit
SLEN: striping distance in bit
ELEN: max element size in bit
ex: --with-vector=v128:e32:s128
2. add __int128_t type checking
3. add --varch command option and help message
ex: --varch=v512:e64:s512
Signed-off-by: Dave Wen <dave.wen@sifive.com>
It is a preparatory commit.
vector extension has supported segment load/store which have
the same prefix and a serial number as suffix.
ex:
base :
vlsegb.v
variation:
vlseg2.v, vlseg3b ... vlseg8b
dynamic string can reduce the typing effort
Signed-off-by: Chih-Min Chao <chihmin.chao@sifive.com>
It is preparatory commit for vector extension.
v-ext has hundresds of new instructions and mixing them with scalar instructions
messes up code.
Separate each extension into different list to make thing clean
Signed-off-by: Chih-Min Chao <chihmin.chao@sifive.com>
* Clean up debug module options.
1. Instead of passing each one a few levels deep, create
debug_module_config_t which contains them all.
2. Rename all those command line options so they start with --dm for
debug module.
3. Add --dm-no-halt-groups to disable halt group support.
* Update changelog.
This is used to make sure that OpenOCD can work on targets that don't
support abstract access to CSR registers. It replaces a simpler hack,
which caused #266.
* Implement hasel/hawindow support.
This should allow simultaneous resume and halt to work.
* Fix anyrunning/anyhalted bits.
* Add --without-hasel argument for testing.
* Make halt/resume times more equal.
Switching threads after every instruction executed in debug mode leads
to a lot of extra instructions being executed on the "other" thread when
both are really supposed to halt/resume near-simultaneously. Fixed that
by adding wfi to debug_rom.S, and implementing it to switch to the other
hart as well as check for JTAG input.
When resuming, write the hart ID to the debug ROM so that the DM knows
which hart actually resumed. (Before simultaneous resume it just assumed
the current one.)
Also got rid of resume symbol in debug_rom.S since it had no purpose.
* Preserve Debug ROM entry points.
* Make sure minstret is correct when wfi happens.
Several people have raised the concern that dynamically linking the
built-in components of Spike causes more headaches than it's worth.
IIRC, the only reason we did this is to better support the
`--extension=libfoo.so` feature.
The spec says that e.g. MEI takes priority over SEI. We got this right in
the common case that SEI is delegated to S-mode, but we reversed it in the
undelegated case.
The destination privilege was correct, so this wasn't much of a problem,
but it is technically noncompliant.
Resolves#288
* Update debug_defines from latest spec.
* Implement halt groups.
This lets the debugger halt multiple harts near simultaneously.
* Revert encoding, which I updated accidentally.
PMP checks should unconditionally fail if the PMP matches part of, but
not all of, an access. We got this right, but went too far: we checked
whether _any_ PMP matches in this manner. In fact, only the first PMP
that maches any of the bytes should be checked in this manner.
Chih-Min Chao
Tim Newsome
Andrew Waterman
Scott Beamer
Andrew Waterman
nmeum
Luís Marques
Dinesh Thirumurthy
Dinesh Thirumurthy
naufal