1
0
Commit Graph

2318 Commits

Author SHA1 Message Date
Megan Wachs
4256d99a9b PLIC: priority/threshold are really WARL (RWSPECIAL). Explain why. 2018-03-05 16:10:05 -08:00
Megan Wachs
41d1a62713 PLIC: Update RegFieldDesc to reflect the fact that source 0 isn't like all the others 2018-03-05 15:29:14 -08:00
Jack Koenig
8c6e745653
Bump chisel and firrtl (#1232)
* Misc changes to better enable autoclonetype
* Bump chisel3 and firrtl and SBT to 1.1.1
2018-03-01 15:19:12 -08:00
Henry Cook
20a8876856
Merge pull request #1190 from freechipsproject/bus-api
BusWrapper API Update
2018-03-01 01:13:50 -08:00
Megan Wachs
cdd2a9227f
Merge pull request #1256 from freechipsproject/json_emit_enums
RegFieldDesc: Emit enumerations into JSON if they exist
2018-02-28 11:32:14 -08:00
Megan Wachs
d13dc8ac2a RegFieldDesc: Emit enumerations if they exist 2018-02-28 09:42:25 -08:00
Andrew Waterman
a48dd575b2
Merge pull request #1254 from freechipsproject/amo-aqrl
Fix mapping of acquire/release AMOs to fence operations
2018-02-27 19:49:40 -06:00
bipult
47d63d6baa
Merge pull request #1251 from freechipsproject/rocket_covers
Added functional covers
2018-02-25 09:01:33 -08:00
Andrew Waterman
eb6e192ec0 Fix mapping of acquire/release AMOs to fence operations
AMO.aq should be implemented as AMO;FENCE, whereas AMO.rl should be
implemented as FENCE;AMO.  These had been swapped.  This error does
not affect cacheable accesses using the blocking D$, nor does it
affect accesses to the data scratchpad, nor does it affect accesses
to strongly ordered I/O regions (which is the default).

Cacheable accesses using the nonblocking D$ and accesses to weakly
ordered I/O regions may manifest memory-ordering violations.  For
these accesses, the workaround is to use AMO.aqrl whenever AMO.aq
or AMO.rl had been used.
2018-02-23 16:39:47 -08:00
Henry Cook
30c0635bb3 subsystem: add some inter-wrapper buffer params 2018-02-23 15:31:18 -08:00
Wesley W. Terpstra
95294bbdcb
PatternPusher: put data at correct address when misaligned (#1249) 2018-02-23 15:20:56 -08:00
Henry Cook
ad823ef43c subsystem: pbus crossing type 2018-02-23 13:52:18 -08:00
Henry Cook
5725e17969 subsystem: even more general coupler methods 2018-02-23 13:52:12 -08:00
Jacob Chang
87eed645d8
Fix JTAG cover description (#1248) 2018-02-23 12:13:31 -08:00
Henry Cook
5b1d72c776 subsystem: expose HasTiles Parameters 2018-02-22 23:46:08 -08:00
Henry Cook
099bbec666 subsystem: more buswrapper coupling methods 2018-02-22 23:45:21 -08:00
Bipul Talukdar
2e548c9ad2 Added functional covers 2018-02-22 23:20:12 -08:00
Andrew Waterman
aad75f2285 Implement misa.C proposal
This proposal hasn't been adopted yet, but anything is better than the
current implementation, where clearing misa.C when the PC is misaligned
is effectively undefined.
2018-02-22 15:12:19 -08:00
Andrew Waterman
c1ee31d133 Fix debug trigger point for stores
In Rocket, debug triggers are supposed to happen before a store
occurs, rather than after.  Previously, we reported the exception
on the store's PC, but the store occurred anyway.  This probably
hasn't been problematic in practice because most stores are
idempotent.
2018-02-22 14:56:57 -08:00
Henry Cook
78883d13e8 subsystem: add TLIdentity.gen and make wrappers more flexible 2018-02-21 18:22:06 -08:00
Henry Cook
eaa908d44f subsystem: more buswrapper methods 2018-02-21 14:52:59 -08:00
Henry Cook
e237f72539 subsystem: XSubsystemModule => XSubsystemModuleImp 2018-02-21 14:52:59 -08:00
Henry Cook
1af02f754e groundtest: fix filename 2018-02-21 14:52:59 -08:00
Henry Cook
030c6f0206 subsystem: bus wrappers now in BaseSubsystem 2018-02-21 14:52:59 -08:00
Henry Cook
b617e26c13 util: augment String and use to name couplers 2018-02-21 14:43:47 -08:00
Henry Cook
a6d3965491 tilelink: bus wrapper scopes called 'couplers' 2018-02-21 14:43:47 -08:00
Henry Cook
57edd7facf subsystem: streamline toTile and fromTile attachment 2018-02-21 14:43:47 -08:00
Wesley W. Terpstra
ef3addee7b diplomacy: put full module + instance path into graphml Description 2018-02-21 14:43:47 -08:00
Henry Cook
62aee56807 diplomacy: base instance names on ValName and module names on className 2018-02-21 14:43:47 -08:00
Henry Cook
3f436a7612 subsystem: new bus attachment api 2018-02-21 14:43:47 -08:00
Henry Cook
8462ea3d5b coreplex => subsystem 2018-02-21 14:42:24 -08:00
Andrew Waterman
8998a97ea1 Preserve WithRV32 behavior: FLEN = 32 2018-02-20 18:28:47 -08:00
Andrew Waterman
1dc1e2c099 support testing RV32D configs 2018-02-20 16:16:39 -08:00
Andrew Waterman
d4fb7ad6a2 DefaultRV32Config should provide fdiv/fsqrt
This is a holdover from before we had built the functional unit.
2018-02-20 16:16:39 -08:00
Andrew Waterman
b487448961 Add FPUParams.fLen option, decoupled from xLen 2018-02-20 16:16:39 -08:00
Andrew Waterman
5e35015651 Minor Rocket fixes to support fLen != xLen 2018-02-20 16:16:39 -08:00
Andrew Waterman
bd29184e11 debug: get beatBytes from pbus, not XLen 2018-02-20 16:16:39 -08:00
Andrew Waterman
62f9b84439 plic: get beatBytes from pbus, not XLen 2018-02-20 16:16:39 -08:00
Andrew Waterman
52e22a1dd8 clint: get beatBytes from pbus, not XLen 2018-02-20 16:16:39 -08:00
Andrew Waterman
1bac2cbdf8 Give Rocket priority over DTIM TL port
The TL port can easily starve the processor, even at only 20% utilization,
because of a bad interaction with the pipeline.  Giving the processor
static priority is OK in practice, since <50% of instructions are loads
and stores in typical workloads.  Even if it executes 100% loads and stores,
it must eventually encounter an I$ miss, taken branch, or exception, so
even malicious code can't permanently starve the TL port.
2018-02-20 11:23:10 -08:00
Megan Wachs
5affd3bec2 RegFieldDesc: fix the output produced for undescribed registers 2018-02-16 10:24:12 -08:00
Megan Wachs
cf7cd03d64
Merge pull request #1239 from freechipsproject/reduce_debug_flags
Reduce Debug Module "flags"
2018-02-16 08:53:41 -08:00
Megan Wachs
d72abb7a12
Debug: revert change to how flags are named 2018-02-15 21:49:32 -08:00
Wesley W. Terpstra
dcfbdabe60 CacheCork: better document edge conditions 2018-02-15 19:14:30 -08:00
Wesley W. Terpstra
ecd069dca4 tilelink: allow FIFO caches
Probably not a smart thing to build, but not illegal!
2018-02-15 19:09:37 -08:00
Wesley W. Terpstra
acecc407a5 HellaCache: we do NOT really support probe below the block size!
If we did, you would somehow have to retain ownership of the
unprobed parts of the block, in case they happened to be dirty.
2018-02-15 19:08:43 -08:00
Megan Wachs
c34b940d9a
ElaborationArtefacts: revert unintentional change 2018-02-15 14:23:54 -08:00
Megan Wachs
e0c3b22d61
RegFieldDesc: same string used to insert/compare 2018-02-15 14:23:27 -08:00
Megan Wachs
b95f68447f RegFieldDesc: Prevent different RegField JSONS from overwriting eachother. 2018-02-15 14:01:47 -08:00
Megan Wachs
64d3731e45 RegFieldDesc: don't put characters into names that need to be sanitized 2018-02-15 13:25:06 -08:00