1
0
Commit Graph

231 Commits

Author SHA1 Message Date
Wesley W. Terpstra
b240505a15 rocketchip: move memory channel Xbar from coreplex to rocketchip
We want to keep the banks split in the outer SoC if there is an L3.
Furthermore, each channel might go to different memory subsystems,
like DDR/HMC/Zero, from rocketchip.
2017-02-03 17:19:21 -08:00
Wesley W. Terpstra
93b2fa197e Artefact output (#545)
* build: stop using empty .prm file

* generator: general-purpose mechanism for creating elaboration artefacts
2017-02-02 19:24:55 -08:00
Wesley W. Terpstra
280af9684b BankedL2Config: use the same LazyModule for all L2 banks
This makes it much easier for banked coherence managers to support
cross-bank functionality, like a common control port, for example.
2017-01-30 14:02:59 -08:00
Wesley W. Terpstra
24ee7f45f5 rocketchip: pass variable l1tol2 connections into coreplex 2017-01-29 11:18:36 -08:00
Wesley W. Terpstra
03f2fe02ac coreplex: support rational crossing to L2 (#534) 2017-01-27 17:09:43 -08:00
Wesley W. Terpstra
3fc55298ef coreplex: provide coherence managers with geometry information 2017-01-23 15:50:39 -08:00
Wesley W. Terpstra
38c9ddffcc BankedL2: move TLFilter BEFORE coherence manager
This lets smart caches exclude the sets that are filtered.
2017-01-21 13:23:07 -08:00
Wesley W. Terpstra
dcadd5a006 coreplex: move TLBuffers for L2 and socBus 2017-01-20 22:23:36 -08:00
Wesley W. Terpstra
3a5e5a65f8 coreplex: support multiple memory channels via diplomatic trickery 2017-01-19 19:07:14 -08:00
Wesley W. Terpstra
258abc5629 coreplex: re-enable stateless L2 config 2017-01-19 19:07:14 -08:00
Wesley W. Terpstra
bf7823f1c8 tilelink2: split suportsAcquire into T and B variants 2017-01-19 19:07:13 -08:00
Henry Cook
e0411c6cde [coreplex] bugfix: re-enable multicore configs via WithNCores 2017-01-19 17:48:04 -08:00
Henry Cook
307f938b88 [rocket] bugfix: fixes #517 2017-01-19 17:48:04 -08:00
Henry Cook
9a6634cd40 Add TLBuffers on the L1 backends and blind exit points (#513)
* [coreplex] add TLBuffers on the exit points from the Tile and Coreplex
* [config] WithBootROMFile
2017-01-17 11:57:23 -08:00
Henry Cook
74b6a8d02b Refactor Tile to use cake pattern (#502)
* [rocket] Refactor Tile into cake pattern with traits
* [rocket] cacheDataBits &etc in HasCoreParameters
* [rocket] pass TLEdgeOut implicitly rather than relying on val edge in HasCoreParameters
* [rocket] frontend and icache now diplomatic
* [rocket] file name capitalization
* [rocket] re-add hook for inserting externally-defined Cores
* [rocket] add FPUCoreIO
* [groundtest] move TL1 Config instances to where they are used
* [unittest] remove legacy unit tests
* [groundtest] remove legacy device tests
2017-01-16 18:24:08 -08:00
Wesley W. Terpstra
52bb6cd9d9 Configs: use a uniform syntax without Match exceptions (#507)
* Configs: use a uniform syntax without Match exceptions

The old style of specifying Configs used total functions.  The only way to
indicate that a key was not matched was to throw an exception.  Not only was
this a performance concern, but it also caused confusing error messages
whenever you had a match failure from a lookup within a lookup.  The
exception could get handled by an outer-lookup that then reported the wrong
key as missing.
2017-01-13 14:41:19 -08:00
Wesley W. Terpstra
020fbe8be9 diplomacy: make config.Parameters available in bundle connect()
This makes it posisble to use Parameters to control Monitors.
However, we need to make all LazyModules carry Parameters.
2016-12-07 12:24:01 -08:00
Henry Cook
f3d0692619 Make a directory for the config package (#464)
* [config] make dir structure mirror packages

* [config] expunge max_int
2016-12-05 10:42:16 -08:00
Wesley W. Terpstra
b7963eca4e copyright: ran scripts/modify-copyright 2016-11-27 22:15:43 -08:00
Wesley W. Terpstra
a17753983a coreplex: allow legacy devices to override the config string (#458) 2016-11-25 19:38:24 -08:00
Wesley W. Terpstra
2b80386a9e rocketchip: TileInterrupts needs a TLCacheEdge (#456) 2016-11-25 17:02:29 -08:00
Wesley W. Terpstra
0baa1c9a45 coreplex: CacheBlockOffsetBits was wrong!
This bug is ancient. I don't understand how it never mattered before.
Anyway, in processors with a custom CacheBlockBytes, this value is wrong!
The symptom is that TL1 components end up missing high address bits.
This causes, for example, a system to jump to 0 instead of RAM.

I don't understand how this very serious bug did not cause problems before.
2016-11-24 18:32:44 -08:00
Wesley W. Terpstra
9f1c668c4f config: when modifying Parameters, subordinate lookups use top 2016-11-23 20:44:45 -08:00
Henry Cook
dae6772624 factor out common cache subcomponents into uncore.util 2016-11-23 12:09:35 -08:00
Wesley W. Terpstra
13190a5de0 rocketchip: re-add AXI4 interface 2016-11-22 17:27:58 -08:00
Wesley W. Terpstra
c230580157 coreplex: rename RocketPlex => RocketTiles 2016-11-22 17:27:58 -08:00
Wesley W. Terpstra
bbabcf67ff coreplex: width adapter should happen as part of coherence manager
In the future we will want the L2 to be wider on the backside so that
we can take advantage of fat DDR controllers (256bits/beat).
2016-11-22 17:27:58 -08:00
Wesley W. Terpstra
3d644b943c coreplex: configString is a property of the RISCVPlatform 2016-11-21 21:13:26 -08:00
Wesley W. Terpstra
e8be365b5d rocketchip: remove GlobalAddrMap completely 2016-11-21 21:13:26 -08:00
Wesley W. Terpstra
d1328a6b6f rocketchip: remove most uses of GlobalAddrMap 2016-11-18 19:38:02 -08:00
Wesley W. Terpstra
001d9821bd Merge remote-tracking branch 'origin/master' into tl2-tile 2016-11-18 18:19:41 -08:00
Wesley W. Terpstra
be8121eeaf coreplex: fix clock crossing 2016-11-18 17:15:57 -08:00
Wesley W. Terpstra
0082d713af coreplex: disable Stateless config until we implement adapter 2016-11-18 16:23:16 -08:00
Wesley W. Terpstra
a6188efc41 rocketchip: break infinite Config loops 2016-11-18 16:18:33 -08:00
Wesley W. Terpstra
37a3c22639 rocketchip: move from using cde to config 2016-11-18 16:18:33 -08:00
Wesley W. Terpstra
30425d1665 rocketchip: eliminate all Knobs 2016-11-18 14:31:42 -08:00
Richard Xia
bab504cc3f Add various granular and composable configs. 2016-11-18 11:30:07 -08:00
Wesley W. Terpstra
179c93db42 tilelink2 broadcast: make it controlled via Config 2016-11-17 17:26:49 -08:00
Wesley W. Terpstra
c82b371354 rocketchip: remove obsolete TL1 config 2016-11-17 14:24:45 -08:00
Wesley W. Terpstra
dfc3a0dafb tilelink2: do not depend on obsolete TL1 configuration 2016-11-17 14:07:53 -08:00
Henry Cook
24e3216fcf coreplex: allow zero interrupt sink/sources 2016-11-16 16:50:36 -08:00
Wesley W. Terpstra
06a7b95d0d tilelink2 broadcast: support bufferless Config 2016-11-16 12:25:11 -08:00
Wesley W. Terpstra
10e459fedb rocket: change connection between rocketchip and coreplex
* rtc and dtm are now crossed half-and-half on the two sides
* groundtest no longer uses riscv platform traits
2016-11-15 18:27:52 -08:00
Henry Cook
71315d5cf5 WIP scala compile and firrtl elaborate; monitor error 2016-11-11 13:07:45 -08:00
Wesley W. Terpstra
32fd11935c rocketchip: use TL2 and AXI4 for memory subsytem 2016-11-04 13:36:47 -07:00
Wesley W. Terpstra
4a2cf6431b coreplex: make 'mem' port an Option until we can use a Seq 2016-11-04 13:35:36 -07:00
Wesley W. Terpstra
8f757a9135 coreplex: rename BankedL2 trait to BankedL2CoherenceManagers 2016-11-04 13:35:36 -07:00
Wesley W. Terpstra
d03046d11c coreplex: fix BankedL2 line width 2016-11-04 13:35:36 -07:00
Wesley W. Terpstra
da3cc3b299 coreplex: TileLink2 l1tol2 memory channels 2016-11-03 22:18:28 -07:00
Wesley W. Terpstra
f83d1d0aaf coreplex: rename trait CoreplexRISCVPlatform
This makes it clear we are talking about the devices one expects in the
platform, not the ISA.
2016-10-31 11:42:47 -07:00
Wesley W. Terpstra
4a0b29850c coreplex: reattach clint interrupt 2016-10-31 11:42:47 -07:00
Wesley W. Terpstra
aabd17d935 rocketchip: must create bundles within Module scope
1. Bundles be created after base class Module constructor runs
2. Bundles must be created before Module(...) runs

Solution: pass a bundle constructor to the cake base class

Require the constructor to take a parameter so people don't use it by
accident; they should get a type error.

Consistently name all the cake arguments with an _io, _coreplex, _outer,
so that they don't shadow the base class variables you should be using.
2016-10-31 11:42:47 -07:00
Wesley W. Terpstra
d52615c39e coreplex: one IntNode per tile 2016-10-31 11:42:47 -07:00
Wesley W. Terpstra
e97844f71e coreplex: make it possible to override the ConfigString 2016-10-31 11:42:47 -07:00
Wesley W. Terpstra
688e1bffdf rocketchip: pull rtcTick out of the coreplex 2016-10-31 11:42:47 -07:00
Wesley W. Terpstra
d51b0b5c02 rocketchip: use self-type 2016-10-31 11:42:47 -07:00
Wesley W. Terpstra
841a31479a coreplex: fix TinyConfig 2016-10-31 11:42:47 -07:00
Wesley W. Terpstra
ba529c3716 rocketchip: use TileLink2 interrupts 2016-10-31 11:42:47 -07:00
Wesley W. Terpstra
6505431eac coreplex: use self-type constraints 2016-10-31 11:42:47 -07:00
Wesley W. Terpstra
ac886026e6 rocketchip: reduce number of type parameters 2016-10-31 11:42:47 -07:00
Wesley W. Terpstra
401fd378b4 rocketchip: include devices from cbus in ConfigString 2016-10-31 11:42:13 -07:00
Wesley W. Terpstra
a73aa351ca rocketchip: fix all clock crossings 2016-10-31 11:42:13 -07:00
Wesley W. Terpstra
825c253a72 rocketchip: move TL2 and cake pattern into Coreplex 2016-10-31 11:42:13 -07:00
Wesley W. Terpstra
dddb50a942 BuildTiles: convert to LazyTile 2016-10-31 11:42:13 -07:00
Wesley W. Terpstra
b99662796d PLIC: converted to TL2 2016-10-31 11:42:13 -07:00
Wesley W. Terpstra
bddfa4d69b Debug: make address configurable 2016-10-31 11:42:13 -07:00
Wesley W. Terpstra
c3dacca39a rocketchip: remove pbus; TL2 has swallowed it completely 2016-10-31 11:42:08 -07:00
Wesley W. Terpstra
0ae45d0f24 rocketchip: bundle (=> B) need not be delayed; Module is constructed later 2016-10-31 11:41:18 -07:00
Megan Wachs
af924d8c51 DebugModule: Instantiate TL2 DebugModule in BaseCoreplex 2016-10-31 11:41:18 -07:00
Andrew Waterman
53360f4c2c Disable U-mode by default unless S-mode is present 2016-10-08 21:29:40 -07:00
Andrew Waterman
eddf1679f5 Use <> instead of := for bi-directional connections 2016-10-04 22:29:39 -07:00
Andrew Waterman
968851f7e3 Default to configurable priorities
up-to-7 levels is kind of arbitrary, but I'm unwilling to introduce
a new Parameter at the moment.
2016-10-04 22:29:39 -07:00
Wesley W. Terpstra
f05298d9bc tilelink2: move general-purpose code out of tilelink2 package 2016-10-03 16:22:28 -07:00
Andrew Waterman
2bdf8c2be7 Merge branch 'master' into move-to-util 2016-09-29 14:42:11 -07:00
Howard Mao
ab3219cf6e don't use Scala to Chisel implicit conversions outside of rocket 2016-09-29 14:35:42 -07:00
Howard Mao
9910c69c67 Move a bunch more things into util package
A lot of utility code was just being imported willy-nilly from one
package to another. This moves the common code into util to make things
more sensible. The code moved were

 * The AsyncQueue and AsyncDecoupledCrossing from junctions.
 * All of the code in rocket's util.scala
 * The BlackBox asynchronous reset registers from uncore.tilelink2
 * The implicit definitions from uncore.util
2016-09-29 14:23:42 -07:00
Andrew Waterman
e928b741ce Default mtvec=0, not None
Setting it to None was a mistake.  It makes it far harder to
diagnose boot bugs, as you end up fetching from random addreses
after trapping.
2016-09-29 13:52:41 -07:00
Howard Mao
c45cc76cef Get rid of remaining MemIO code
The only thing we were still using it for was for the MIFDataBits
and MIFTagBits parameters. We replace these with EdgeDataBits and
EdgeIDBits.
2016-09-27 16:28:17 -07:00
Howard Mao
7d6fb950b6 Give TileLink IDs more sensible names
* Outermost -> MCtoEdge
 * MMIO_Outermost -> MMIOtoEdge

Then the corresponding parameters objects are

 * L1toL2 -> innerParams
 * L2toMC -> outerMemParams
 * L2toMMIO -> outerMMIOParams
 * MCtoEdge -> edgeMemParams
 * MMIOtoEdge -> edgeMMIOParams
2016-09-27 12:48:01 -07:00
Howard Mao
8a55521b01 move memory width adapter from coreplex to periphery 2016-09-27 12:48:01 -07:00
Howard Mao
201e247f73 Factor coreplex IO connection into separate trait (#350)
This would allow, for instance, putting the coreplex on a separate clock
domain and crossing the IOs over through asynchronous queues.

The ExampleMultiClockTop* classes are removed since they no longer fit
into the class hierarchy.
2016-09-27 11:55:32 -07:00
Howard Mao
ea9f0a868f TileLink utility objects should not take implicit parameters
We have a handful of TileLink-related helper objects
(wrappers, unwrappers, width adapters, and enqueuers). Previously, using
them could be error-prone, because you had to make sure the implicit
parameters they took in had the same TLId as the TileLinkIO bundles
passed in as inputs. This is rather silly, we should just use the
parameters in the bundle.
2016-09-26 17:28:21 -07:00
Howard Mao
803739a95c Make sure coreplex mmio's TLId is correct (thanks to zizztux) 2016-09-26 17:28:21 -07:00
Henry Cook
1e54820f8c Merge remote-tracking branch 'origin/master' into unittest-config 2016-09-22 16:03:51 -07:00
Henry Cook
47c5d1a992 [WIP] Move RocketTestSuite generation into RocketchipGenerator 2016-09-22 14:31:45 -07:00
Yunsup Lee
1b1ef3be07 simplify base Coreplex bundle 2016-09-21 18:29:28 -07:00
Yunsup Lee
d2df6397cd rename trc (tile reset clock) bundles to tcr (tile clock reset) 2016-09-21 18:29:28 -07:00
Yunsup Lee
5bb575ef74 rename internal/external MMIO network to cbus/pbus respectively 2016-09-21 18:29:28 -07:00
Henry Cook
fd5e00fed9 [coreplex] rename Testing.scala -> RocketTestSuite.scala 2016-09-21 17:35:39 -07:00
Yunsup Lee
7afd630d3e add multiclock support to Coreplex 2016-09-21 16:55:26 -07:00
Henry Cook
ed91e9a89b Merge remote-tracking branch 'origin' into testharness-refactor 2016-09-20 13:03:21 -07:00
Andrew Waterman
d0572d6aab Allow reset vector to be set dynamically
A chip's power-up sequence, or awake-from-sleep sequence, may wish to
set the reset PC based upon dynamic properties, e.g., the settings of
external pins.  Support this by passing the reset vector to the Coreplex.
ExampleTop simply hard-wires the reset vector, as was the case before.

Additionally, allow MTVEC to *not* be reset.  In most cases, including
riscv-tests, pk, and bbl, overriding MTVEC is one of the first things
that the boot sequence does.  So the reset value is superfluous.
2016-09-19 17:18:03 -07:00
Richard Xia
63f13ae7ce Merge remote-tracking branch 'origin/master' into rxia-testharness-refactor 2016-09-16 17:10:52 -07:00
Andrew Waterman
86b70c8c59 Rename PRCI to CoreplexLocalInterrupter
That's all it's doing (there wasn't much PRC).
2016-09-16 14:26:34 -07:00
Wesley W. Terpstra
f05222a072 testconfigs: disable atomics until AtomicAbsorber finished 2016-09-15 21:28:56 -07:00
Wesley W. Terpstra
669e3b0d96 Regression: fix-up address lookup 2016-09-15 21:28:56 -07:00
Wesley W. Terpstra
fb24e847fd rocketchip: globals are for sissies 2016-09-15 21:28:56 -07:00
Wesley W. Terpstra
644f8fe974 rocketchip: switch to TL2 mmio + port PRCI 2016-09-15 21:28:56 -07:00
Jack Koenig
f2fe437fa4 Use CDEMatchError for improved performance (#304) 2016-09-15 19:47:18 -07:00
Henry Cook
be9ddae77f make groundtest and unitest peers of rocketchip, with their own packages, harnesses and configs 2016-09-15 13:04:01 -07:00
Henry Cook
c6f252a913 Remove Option from success flag in coreplex; just use a sane default. 2016-09-15 12:19:22 -07:00
Howard Mao
49863944c4 merge ClientTileLinkEnqueuer and ClientUncachedTileLinkEnqueuer objects into TileLinkEnqueuer 2016-09-14 21:36:27 -07:00
Howard Mao
f5db83a72f NTiles should not be a Knob 2016-09-14 21:16:54 -07:00
Yunsup Lee
e404bea2ee Merge branch 'master' into move-bootrom 2016-09-14 18:58:48 -07:00
Yunsup Lee
97809b183f refactor unittest framework
as a result, there's another SUITE that needs to run
2016-09-14 18:10:21 -07:00
Yunsup Lee
710f1ec020 Move BootROM from Coreplex to Periphery 2016-09-14 16:09:59 -07:00
Andrew Waterman
565444c40e Make UnitTestCoreplex cope with an external MMIO network 2016-09-14 12:19:21 -07:00
Andrew Waterman
c3ddff809b Move PRCI from Coreplex to always-on block, where it belongs 2016-09-14 11:01:05 -07:00
Andrew Waterman
5566bf1b13 Don't route PLIC interrupts through PRCI
The PLIC is local to the Coreplex, and PRCI should not be.
2016-09-14 11:01:05 -07:00
Henry Cook
7dd4492abb First cut at refactoring unittests into a top-level utility. Individual tests co-located with their DUT. No functional changes. 2016-09-13 20:30:14 -07:00
Andrew Waterman
61cbe6164d Add option to execute JAL from decode stage
This is particularly helpful for designs that don't have a BTB, but
it becomes the critical path for designs with RVC.  Caveat emptor.
2016-09-13 02:32:00 -07:00
Andrew Waterman
266a2f24bd Disable Mul early out by default if XLen == 32
With a default unroll of 8, it doesn't help performance, but costs area.
2016-09-12 16:50:08 -07:00
Howard Mao
f3cdeb08c6 pass nMemChannels to coreplex through CoreplexConfig 2016-09-12 12:40:10 -07:00
Yunsup Lee
bb3f514e8d now able to add periphery devices through traits
Unfortunately, I had to touch a lot of code, which weren't quite possible to split up into multiple commits.
This commit gets rid of the "extra" infrastructure to add periphery devices into Top.
2016-09-10 23:39:29 -07:00
Colin Schmidt
254f49093c only use companion objects for types 2016-09-07 12:32:34 -07:00
Andrew Waterman
70cfd7ce13 Make DefaultRV32Config be RV32IMAFCS, not RV32IMC
The latter is more the domain of TinyConfig.
2016-09-07 01:58:25 -07:00
Andrew Waterman
9fea4c83da Add RV32F support 2016-09-07 00:05:39 -07:00
Howard Mao
f34843f1b9 fix assignment of incoherent vector 2016-09-04 10:12:16 -07:00
Yunsup Lee
a4c1942958 flatten Coreplex module hierarchy 2016-09-02 17:45:08 -07:00
Andrew Waterman
63679bb019 Add support for L1 data scratchpads instead of caches
They fit in the same part of the address space as DRAM would be, and
are coherent (because they are not cacheable).

They are currently limited to single cores without DRAM.  We intend
to lift both restrictions, probably when we add support for
heterogeneous tiles.
2016-09-02 16:22:07 -07:00
Andrew Waterman
fb50f7c9dd Set default TileLink width to XLen 2016-09-02 15:27:54 -07:00
Andrew Waterman
6872000f5e Merge pull request #239 from ucb-bar/move_rtc
Move RTC
2016-09-02 15:17:49 -07:00
Megan Wachs
8163a6b597 Make it easier to override the 'placeholder' Real-Time-Clock, to allow more real-world applications 2016-09-02 11:11:40 -07:00
Andrew Waterman
c05ba1e864 Add TileId parameter, generalizing GroundTestId
This usually shouldn't be used in Tiles that are meant to be P&R'd once
and multiply instantiated, as their RTL would no longer be homogeneous.
However, it is useful for conditionalizing RTL generation for
heterogeneous tiles.
2016-09-02 00:10:50 -07:00
Andrew Waterman
f4524e4c91 Add PML for Boolean.option; use it 2016-08-31 13:43:04 -07:00
Andrew Waterman
8dbee2b133 Don't conditionalize running bmarks on UseVM 2016-08-29 13:43:29 -07:00
Andrew Waterman
f91552a650 Add performance counter support 2016-08-29 12:31:52 -07:00
Andrew Waterman
9ca82dd397 reset default MulDiv config to moderately fast default
Closes #228.

In commit 3f8c60bbd6 I inadvertently
changed the configuration while refactoring it.
2016-08-29 12:31:52 -07:00
Henry Cook
93c801f598 Streamline the Generator App and associated utilities. Remove deprecated call to chiselMain and useless Chisel2 args. Update arguments to sbt run. (#227) 2016-08-25 17:26:28 -07:00
mwachs5
22ffe36258 Add a queue for timing QoR between L2->MMIO network (#217) 2016-08-19 22:51:49 -07:00
Howard Mao
7b20609d4d reorganize moving non-submodule packages into src/main/scala 2016-08-19 13:45:23 -07:00