1
0
Commit Graph

2133 Commits

Author SHA1 Message Date
Henry Cook
a887baa615 rocket: base trait for reporting ecc errors 2017-09-21 14:58:47 -07:00
Andrew Waterman
88c782cc70 Report D$ uncorrectable errors on C channel 2017-09-20 17:15:11 -07:00
Andrew Waterman
6bc20942b5 Don't cache TL error responses; report access exceptions 2017-09-20 17:01:08 -07:00
Andrew Waterman
9b828a2640 Only look at error signal on last beat 2017-09-20 15:15:21 -07:00
Andrew Waterman
026fa14bf8 Rename trace.addr -> iaddr 2017-09-20 14:32:41 -07:00
Andrew Waterman
5b2f458214 Merge branch 'master' into ma-fetch 2017-09-20 12:18:03 -07:00
Andrew Waterman
f1a506476b Merge pull request #994 from freechipsproject/beu
Add L1 bus-error unit
2017-09-20 12:17:08 -07:00
Andrew Waterman
f5bd639863 Don't write badaddr on misaligned fetch exceptions
It's optional, and we were doing it wrong before, so just don't do it.
2017-09-20 10:52:41 -07:00
Andrew Waterman
db57e943f3 Report TL errors into D$ 2017-09-20 00:05:07 -07:00
Andrew Waterman
aaad73f019 Add an intra-tile xbar 2017-09-20 00:05:07 -07:00
Andrew Waterman
afad25fceb Integrate L1 BusErrorUnit 2017-09-20 00:05:07 -07:00
Andrew Waterman
dbf599f6a1 Support SynchronizerShiftReg(sync = 0)
This makes it easier to parameterize code where the synchronizer
might not always be needed.
2017-09-20 00:05:07 -07:00
Andrew Waterman
79dab487fc Implement bus error unit 2017-09-20 00:05:07 -07:00
Andrew Waterman
ed18acaae0 Report D$ errors 2017-09-20 00:05:07 -07:00
Andrew Waterman
034ea722f4 Report I$ errors 2017-09-20 00:05:07 -07:00
Andrew Waterman
9a175b0fb1 Statically report error correction/detection capability from ECC codes 2017-09-20 00:05:07 -07:00
Andrew Waterman
4d6d6ff641 Add instruction-trace port 2017-09-19 22:59:57 -07:00
Andrew Waterman
acea94bcef Merge pull request #1001 from freechipsproject/address-decoder
Address decoder "improvements"
2017-09-19 22:38:53 -07:00
Jacob Chang
b4fc5104d4 Add cover property API that can be refined through Config PropertyLibrary (#998) 2017-09-19 19:26:54 -07:00
Henry Cook
57e8fe0a6b Merge pull request #1000 from freechipsproject/name-seqmems
try to give seqmems clearer names for use with external tools
2017-09-19 17:59:00 -07:00
Andrew Waterman
87b92cb206 Scan AddressDecoder bits left to right
This heuristic is brittle but fixes deduplication in RocketTile.
2017-09-19 17:47:24 -07:00
Andrew Waterman
72bd89a2af Add another AddressDecoder debug message 2017-09-19 17:47:17 -07:00
Andrew Waterman
fb2ad11347 Improve AddressDecoder optimization function
This function is better 27% of the time but worse 6% of the time.
2017-09-19 17:47:12 -07:00
Henry Cook
8db5bbbae0 try to give seqmems clearer names 2017-09-19 13:41:11 -07:00
Megan Wachs
826fc8ba61 Merge remote-tracking branch 'origin/master' into test_mode_reset 2017-09-18 09:50:27 -07:00
Andrew Waterman
d93d7b9fa4 Only merge stores that aren't yet pending
This fixes a deadlock (and possibly memory corruption, though that is
unconfirmed).  The following sequence manifests it, assuming t0
is 32-byte aligned:

    sw t0, 0(t0)
    sw t0, 16(t0)
    lw t1, 4(t0)
    lw t2, 4(t0)
2017-09-17 15:01:07 -07:00
Megan Wachs
c85333f826 Merge remote-tracking branch 'origin/test_mode_reset' into test_mode_reset 2017-09-17 13:51:46 -07:00
Megan Wachs
215e072e5c test_mode_reset: fix typos 2017-09-17 13:51:40 -07:00
Henry Cook
9b75dd7e5b Merge branch 'master' into test_mode_reset 2017-09-15 17:26:11 -07:00
Megan Wachs
641a8e7eab test_mode_reset: Correct some gender issues. Tie off signals in the test harness 2017-09-15 16:36:35 -07:00
Megan Wachs
6cda4504ac test_mode_reset: use a cleaner interface with bundles and options instead of individual signals 2017-09-15 12:30:39 -07:00
Megan Wachs
ffc514d1bc test_mode_reset: Add missing file 2017-09-14 13:17:37 -07:00
Megan Wachs
a0396b63e8 test_mode_reset: fix one bulk-connect gender issue 2017-09-14 13:16:13 -07:00
Megan Wachs
44edc5fdc3 test_mode_reset: Use simpler apply() method 2017-09-14 13:16:13 -07:00
Megan Wachs
82c00cb656 reset_catch: Allow Test Mode Overrides 2017-09-14 13:16:13 -07:00
Henry Cook
e50d14415e tilelink: more verbose requires 2017-09-13 11:25:42 -07:00
Henry Cook
56dae946b6 coreplex: MemoryBusParams.beatBytes also based on XLen 2017-09-13 11:25:42 -07:00
Henry Cook
b86f4b9bb7 config: use Field defaults over Config defaults
Also rename some keys that had the same class name as their value's class name.
2017-09-13 11:25:42 -07:00
Henry Cook
a7540d35b7 ports: use BigInts instead of Longs and the new x"..." context 2017-09-13 11:25:42 -07:00
Henry Cook
37c5af1c0d diplomacy: add x"..." string context
Enables hex address literals containing underscores.
Converts them to BigInts.
2017-09-13 11:25:42 -07:00
Henry Cook
063ca0ed4a Merge pull request #983 from freechipsproject/kill-paddrbits
Remove global fields PAddrBits and ResetVectorBits
2017-09-11 12:51:10 -07:00
Andrew Waterman
1f606d924f Don't perform in-place correction if there was a recent store (#988)
Since the correction updates the entire word, the WAW hazard detection
logic is not sufficient to prevent overwriting a recent store.  So,
re-read the word after all pending stores have drained.
2017-09-08 16:26:54 -07:00
Henry Cook
9c0bfbd500 tile: remove global Field ResetVectorBits
Reset vector width is determined by systemBus.busView.
Also move some defs from HasCoreParameters to HasTileParameters.
2017-09-08 14:50:59 -07:00
Henry Cook
3133c321b7 scratchpad: remove dependency on HasCoreParameters 2017-09-08 13:55:40 -07:00
Henry Cook
e46aeb7342 tile: remove PAddrBits in favor of SharedMemoryTLEdge 2017-09-08 13:53:36 -07:00
Wesley W. Terpstra
e7de7f3e82 Merge pull request #985 from freechipsproject/flop-interrupts
Add Parameters to diplomatic edges
2017-09-08 13:16:11 -07:00
Andrew Waterman
53dfc5e9be Remove overzealous assertion (#987)
This assertion made sure the D$ controller was able to write the tag RAM
when a cache line was refilled.  However, it is benign if it fails to do
so: the metadata is invalid at this point, so the miss will simply happen
a second time.

This happens when resolving a tag ECC error during hit-under-miss.
2017-09-07 18:17:56 -07:00
Wesley W. Terpstra
e723a3f42b MemoryBus: fanout the A for performance 2017-09-07 16:03:35 -07:00
Wesley W. Terpstra
6879f5bfb1 tilelink: Xbar now allows for fanout control 2017-09-07 16:03:35 -07:00
Wesley W. Terpstra
e831acba9c adapters: support bulk connections 2017-09-07 16:03:35 -07:00
Wesley W. Terpstra
06a244f9f9 diplomacy: rename {Left,Right}Star to refer to {Source,Sink}Cardinality 2017-09-07 16:03:35 -07:00
Wesley W. Terpstra
bef593c21a diplomacy: edges now capture their Parameters 2017-09-07 16:03:35 -07:00
Wesley W. Terpstra
80ed27683e diplomacy: protect against API leakage 2017-09-07 16:03:35 -07:00
Wesley W. Terpstra
1365c5f90c diplomacy: implement DisableMonitors scope 2017-09-07 16:03:35 -07:00
Wesley W. Terpstra
a450357744 tilelink: Monitor construction method is unconditional
Whether or not a Monitor should be placed is decided by diplomacy.
2017-09-07 16:03:35 -07:00
Wesley W. Terpstra
7a8364ef08 diplomacy: leverage new Parameters defaults 2017-09-07 16:03:35 -07:00
Wesley W. Terpstra
655a08f12e config: support default values for Field[T] keys 2017-09-07 16:03:35 -07:00
Wesley W. Terpstra
09d8d476c5 config: require Parameters keys to be Field[T]
This has been good practice for ages. Enforce it.
2017-09-07 16:03:35 -07:00
Wesley W. Terpstra
42f1ae27fc Xbar: use the IdentityModule to encourage wider fanout 2017-09-07 16:03:35 -07:00
Wesley W. Terpstra
5626cdd18f util: add the IdentityModule, useful to dedup wires 2017-09-07 16:03:35 -07:00
Wesley W. Terpstra
1a87ed1193 coreplex: add externalSlaveBuffers configuration option 2017-09-07 16:03:35 -07:00
Wesley W. Terpstra
fd8a51a910 coreplex: rename externalBuffers to externalMasterBuffers 2017-09-07 16:03:35 -07:00
Wesley W. Terpstra
4911a7d44f tilelink Bus: toAsyncSlaves now supports BufferChains 2017-09-07 16:03:35 -07:00
Wesley W. Terpstra
040f7e1d49 tilelink: add Bus.toSyncSlaves for easy BufferChain attachment 2017-09-07 16:03:35 -07:00
Wesley W. Terpstra
d5c6494f59 tilelink: Bus.toRationalSlaves can have a BufferChain 2017-09-07 16:03:35 -07:00
Wesley W. Terpstra
80965e8230 tilelink Buffer: use new :=? adapter API 2017-09-07 16:03:35 -07:00
Wesley W. Terpstra
1b705f62f6 diplomacy: support :=? for unknown star inference 2017-09-07 16:03:35 -07:00
Wesley W. Terpstra
6bfea86dbf config: support p.lift(key) to optionally return a value 2017-09-07 16:03:34 -07:00
Wesley W. Terpstra
2d93262f71 RationalCrossing: use ShiftQueues
These are faster and small don't cost much more.
2017-09-07 16:03:34 -07:00
Wesley W. Terpstra
50d5d8c1fd ShiftQueue: added a helper object 2017-09-07 16:03:34 -07:00
Wesley W. Terpstra
3e3024c256 ShiftQueue: fix bug in !flow case 2017-09-07 16:03:34 -07:00
Wesley W. Terpstra
ed70b243bd plic: support a configurable number of interrupt register stages 2017-09-07 16:03:34 -07:00
Wesley W. Terpstra
9b55063de6 clint: support a configurable number of interrupt register stages 2017-09-07 16:03:34 -07:00
Megan Wachs
126d56b254 synchronizers: I learn how foldRight works 2017-09-07 10:48:27 -07:00
Megan Wachs
1da6cb85ab shiftReg: Make it so that register '0' is always closest to the q output, regardless of the type of shift register created. 2017-09-07 09:57:50 -07:00
Megan Wachs
dcafb5fea3 Merge remote-tracking branch 'origin/master' into async_reg 2017-09-06 11:07:19 -07:00
Megan Wachs
3c4b472f66 shift regs: remove some unnecessary primitives, and add some that actually are necessary 2017-09-06 10:37:59 -07:00
Jim Lawson
f1b7666d21 Jtagresettobool - add explicit toBool cast now required on reset. (#984)
Add explicit toBool cast on reset, for chisel3 compatability
2017-09-06 09:49:47 -07:00
Megan Wachs
777f052f95 regs: Add named/initial value ShiftRegister primitives so they are all in one place 2017-09-05 17:32:53 -07:00
Wesley W. Terpstra
b1cacc56ad SystemBus: restore correct order of FIFOFixer and Buffer 2017-09-05 16:41:39 -07:00
Wesley W. Terpstra
b74a419bfb FrontBus: FIFOFixer should not have a buffer between it and Xbar 2017-09-05 16:27:57 -07:00
Megan Wachs
e9e46db600 sync reg: Rename the file to reflect the more generic shift registers also in the file. 2017-09-05 15:54:25 -07:00
Megan Wachs
5df23c5514 Synchronizers: remove some newlines and unncessary gen's 2017-09-05 15:17:21 -07:00
Wesley W. Terpstra
e65f49b89a FrontBus: attach to splitter for cross-chip visibility 2017-09-05 15:03:41 -07:00
Wesley W. Terpstra
5886025b1a sbus => pbus: 2 buffers should already be enough
There is a buffer on the sbus backside.
There is a buffer on the pbus frontside.

Between them is only an AtomicAutomata.
That should be enough for most designs.
2017-09-05 15:03:38 -07:00
Henry Cook
a902e15987 pbus: clarify that we are adding buffers when attaching to sbus 2017-09-05 15:03:38 -07:00
Henry Cook
8fc4d78c84 frontbus: provide fifofixer on the side of the front bus where masters connect 2017-09-05 15:03:38 -07:00
Megan Wachs
667d966410 TLBuffer: Create a wrapper module for TLBufferChain, to allow for more stable naming 2017-09-05 15:03:38 -07:00
Megan Wachs
94f06dc85c pbus: turn down overkill buffering between PBus and SBus 2017-09-05 15:03:38 -07:00
Megan Wachs
c353f68dc0 buses: name dummy buffers too 2017-09-05 15:03:38 -07:00
Henry Cook
3bde9506c6 coreplex: allow buffer chains on certain bus ports 2017-09-05 15:03:36 -07:00
Megan Wachs
57d0360c35 frontbus: Name the connection. 2017-08-30 18:07:34 -07:00
Megan Wachs
c99afe4c66 buses: Name all the things. 2017-08-30 17:31:42 -07:00
Henry Cook
32cb358c81 coreplex: include optional tile name for downstream name stabilization 2017-08-30 15:48:55 -07:00
Megan Wachs
183fefb2b9 Front/SystemBus: allow naming the intermediate TLNodes that get sprinkled in 2017-08-30 15:27:56 -07:00
Wesley W. Terpstra
d5b62dffda SystemBus: add stupidly many (4 more) buffers from sbus=>pbus
This should probably be reverted.
2017-08-30 14:22:49 -07:00
Henry Styles
f7330028cc Add optional frontbus for peripherals mastering into SBus. Switch FF and Buffer order on non-tile masters into SBus. Buffer non-L2 side of splitter 2017-08-30 14:22:49 -07:00
Wesley W. Terpstra
173f185b17 Merge pull request #976 from freechipsproject/system-buffer
SystemBus: add output buffering
2017-08-30 23:22:13 +02:00
Wesley W. Terpstra
656609d610 SystemBus: split FIFOFixers along bus boundaries
If you have a system with a lot of periphery slaves, you wan to FIFO fix
them on the periphery bus rather than paying the circuit cost at the sbus.
2017-08-30 13:28:11 -07:00
Megan Wachs
a3bc5f2e33 synchronizers: Add a generic shift register and then extend from it, since an asynchronously resettable shift register is also a useful primitive 2017-08-30 12:59:16 -07:00
Megan Wachs
8139014c9e syncrhonizers: Remove unused sync from superclass 2017-08-30 12:33:03 -07:00
Megan Wachs
9dd6c4c32d synchronizers: New chisel ways of cloning type and use simpler lambda function 2017-08-30 12:11:14 -07:00
Megan Wachs
bd32f0c122 synchronizers: properly pass parameters up to the superclass 2017-08-30 11:58:25 -07:00
Megan Wachs
483e63da19 synchronizers: Correctly pass the width through 2017-08-30 11:50:25 -07:00
Megan Wachs
a62ce0afe6 TLBuffer: Add a nodedebugstring for quick browsing of the properties of the buffer. 2017-08-29 10:36:46 -07:00
Megan Wachs
c473538e36 Merge remote-tracking branch 'origin/master' into async_reg 2017-08-28 17:19:03 -07:00
Megan Wachs
451334ac73 Add 1-deep synchronizer register for output of AsyncQueue 2017-08-28 17:18:54 -07:00
Wesley W. Terpstra
bf19440db5 SystemBus: use a full buffer on slaves 2017-08-26 02:47:04 -07:00
Megan Wachs
85c39b2f97 syncregs: Not sure the use case for SynchronizerShiftRegInit, so remove it YAGNI 2017-08-24 17:47:04 -07:00
Megan Wachs
4e773f4738 syncregs: Use synchronizer primivites for LevelSyncCrossing 2017-08-24 17:42:31 -07:00
Megan Wachs
130b24355f syncregs: Use synchronizer primitives for IntXing 2017-08-24 17:39:07 -07:00
Megan Wachs
8b462d1595 syncregs: Use common primitives for AsyncQueue grey code synchronizers 2017-08-24 17:34:07 -07:00
Megan Wachs
3461cb47cc syncregs: Make Reset catcher use the synchronizer primitive 2017-08-24 17:26:38 -07:00
Megan Wachs
c78ee9f0e4 syncreg: Refactor common code 2017-08-24 17:18:04 -07:00
Megan Wachs
d83a6dc6af syncregs: Add utilities for Synchronizing Shift Registers 2017-08-24 16:55:17 -07:00
Megan Wachs
7f683eeb24 async_regs: Make modules have predictable names 2017-08-24 15:33:53 -07:00
Megan Wachs
0f75ebee92 async_reg: Rename the file to match scalastyle 2017-08-24 15:31:29 -07:00
Megan Wachs
103b6bc6d3 systemBus: allowing naming the TLBuffers which get inserted 2017-08-24 14:49:12 -07:00
Wesley W. Terpstra
17134125e1 SystemBus: remove misnamed functions (#972)
These functions were actually for cross connecting chips.
2017-08-24 23:35:01 +02:00
Andrew Waterman
82df766f4a Merge pull request #963 from freechipsproject/interrupt-order
Respect ISA requirements on interrupt priority order
2017-08-18 00:10:19 -07:00
Andrew Waterman
8087a205cc Remove redundant check in interrupt priority encoding
chooseInterrupts already sorts M interrupts above S interrupts.
2017-08-17 22:23:42 -07:00
Andrew Waterman
cbe7c51b50 Respect ISA requirements on interrupt priority order
a62e76cb16
2017-08-17 21:27:08 -07:00
Shreesha Srinath
b1719cfee0 Fixing requirements for PAddrBits (#961)
Previously, the requirement for PAddrBits only checked to be equal or greater than the bundle bits. Changing it to check for these to match exactly as for cases when the PAddrBits greater than address bits we could run into scenarios which cause possible address wrap around issues.
2017-08-17 11:53:59 -07:00
Megan Wachs
1db4b3be9a Merge pull request #957 from freechipsproject/param_jtag_vpi
jtag_vpi: Use Parameterized Black Box
2017-08-14 18:37:30 -07:00
Megan Wachs
8783d51c97 jtag_vpi: Use Parameterized Black Box to allow TestHarnesses to override the clock speed 2017-08-14 17:25:47 -07:00
Wesley W. Terpstra
710a782145 HeterogenousBag: empty bags were being combined! (#956)
This lead to strange firrtl errors when you had two empty
HeterogeneousBags in the same Bundle.
2017-08-14 15:48:42 -07:00
Andrew Waterman
e945f6e265 Merge pull request #955 from freechipsproject/fix-acquire-before-release
Fix acquire before release
2017-08-13 18:29:58 -07:00
Megan Wachs
88332bd885 max-core-cycles: Add a +max-core-cycles PlusArg 2017-08-13 15:47:14 -07:00
Andrew Waterman
3cbc5262ec Don't permit new acquires until the release queue is drained
If the queue is not empty before a dirty miss, C could block D.
I haven't seen this in the wild, but it could happen because of
dirty probe responses backed up in the queue.
2017-08-13 13:18:45 -07:00
Andrew Waterman
0190724492 Actually use the C-channel acquire-before-release queue
oops...
2017-08-13 13:03:35 -07:00
Andrew Waterman
7387f2a93a Don't block D-channel when handling a probe
This is an acquire-before-release regression.
2017-08-12 16:13:24 -07:00
Andrew Waterman
604abd5b07 Only report ECC errors when the RAM was actually read 2017-08-12 15:28:03 -07:00
Andrew Waterman
18fb052fc9 DRY 2017-08-12 15:27:30 -07:00
Andrew Waterman
176110b6d3 Don't trigger ECC writebacks when a release is in flight 2017-08-12 15:23:57 -07:00
Wesley W. Terpstra
f191bb994c PatternPusher: can now expect a certain output (#952) 2017-08-11 18:10:27 -07:00
Wesley W. Terpstra
baf769f924 tilelink: add PatternPusher, a device to inject a fixed traffic pattern (#950) 2017-08-11 15:07:10 -07:00
Andrew Waterman
a3358f34a0 Fix priority inversion for two back-to-back divides (#948)
If the first one is killed for some unrelated reason (e.g. write port
hazard), the second one will still issue to the div-sqrt unit.  While
it will itself later be killed, the fact that the later instruction
acquires a resource needed by the former instruction leads to deadlock.
2017-08-10 17:12:09 -07:00
Andrew Waterman
0a591c5b5b Roll back use of UIntToOH1 (#946)
These appear to be equivalent, but the old one seems to fail in Vivado and
this one seems to pass.  This is not yet conclusive.
2017-08-09 18:39:47 -07:00
Andrew Waterman
721770244e Fix IBuf bug
Don't examine a packet's xcpt signal if it might be invalid.  In this case,
the correct fix is to not examine xcpt at all; the deleted code was vestigial.
(Note, the other use of xcpt(j+1) in this code is indeed safe.)
2017-08-09 09:47:51 -07:00
Wesley W. Terpstra
a9b1410f01 BusBlocker: parameterize page granularity 2017-08-08 17:10:01 -07:00
Wesley W. Terpstra
010ba94474 BusBlocker: rename a variable 2017-08-08 17:00:22 -07:00
Wesley W. Terpstra
6d6fc38787 BusBlocker: lock bit should affect the prior PMP address, not next 2017-08-08 17:00:12 -07:00
Andrew Waterman
809c7e8551 Don't merge stores that manifest WAW hazards
The following sequence would drop the first store when eccBytes=4:

    sb x0, 0(t0)
    nop
    sb x0, 4(t0)
    nop
    sb x0, 1(t0)

Because the first and second store are to different ECC granules, the
hazard check correctly allowed the second one to proceed, but the third
was merged with the second, even though it conflicted with the first.
So, don't allow the third to be merged with the second, since the second
stored to a different ECC granule.
2017-08-08 15:19:05 -07:00
Wesley W. Terpstra
3ef6e4c9f2 Merge pull request #939 from freechipsproject/bus-blocker
tilelink: PMP controlled BusBlocker prevents bus accesses
2017-08-08 15:06:55 -07:00
Andrew Waterman
82e13443b2 Merge pull request #937 from freechipsproject/critical-paths
Perform tag error detectoin/correction in same cycle as RAM
2017-08-08 15:03:28 -07:00
Wesley W. Terpstra
8f261adc6b BusBlocker: change default policy to deny 2017-08-08 14:19:59 -07:00
Wesley W. Terpstra
0d76e96b88 tilelink: PMP controlled BusBlocker prevents bus accesses 2017-08-08 13:28:01 -07:00
Andrew Waterman
7935c61c19 Don't report to the DTIM that data is cacheable
Otherwise, it will attempt to perform AMOs where they're unsupported!
2017-08-08 11:55:04 -07:00
Andrew Waterman
74d309c18e Make I vs. D a static property of TLB, not an input pin
The microarchitecture doesn't really support unified TLBs, so don't fake it.
2017-08-08 11:54:47 -07:00
Andrew Waterman
e92981b0bd DRY 2017-08-08 11:46:38 -07:00
Andrew Waterman
62ccba304c Perform tag error detectoin/correction in same cycle as RAM
The tag RAMs tend to be fast, so take up some of the slack.
This makes s2_nack faster.
2017-08-08 10:21:30 -07:00
Palmer Dabbelt
6d1d285464 Merge pull request #933 from freechipsproject/cinst
Print out the compressed instruction when executing one
2017-08-07 21:40:10 -07:00
Palmer Dabbelt
cc1e2af336 Merge pull request #934 from freechipsproject/critical-paths
Revert "Remove one gate from D$ ECC check"
2017-08-07 19:41:08 -07:00
Henry Cook
c8f8806df0 Merge pull request #932 from freechipsproject/tl-bus-delayer
tilelink: allow insertion of TLDelayer on TLBus outward node
2017-08-07 19:01:39 -07:00
Henry Cook
c4092dd0cc tilelink: improve entropy of bus delayer 2017-08-07 17:36:07 -07:00
Andrew Waterman
402907990c Revert "Remove one gate from D$ ECC check"
This reverts commit 7d94074b05, which
works fine with optimistic behavioral RAMs but not real ones.
2017-08-07 17:33:20 -07:00
Henry Cook
2910d6fa2a tilelink: make bus xbar protected so it can be suggestNamed 2017-08-07 17:30:24 -07:00
Palmer Dabbelt
fc0d5fcf98 Print out the compressed instruction when executing one 2017-08-07 17:21:53 -07:00
Wesley W. Terpstra
e27072e063 Merge pull request #931 from freechipsproject/fix-ram-model-source-reuse
Fix ram model source reuse
2017-08-07 16:56:13 -07:00
Henry Cook
c457c9cb9f tilelink: allow insertion of TLDelayer on TLBus outward node 2017-08-07 16:43:06 -07:00
Wesley W. Terpstra
f8b45564d1 tilelink: RAMModel must support source reuse
If a multibeat response comes back, the source might be reused.
If response reordering has made the multibeat response invalid,
we need to remember this even if the valid bit is cleared on reuse.
2017-08-07 16:01:15 -07:00
Yunsup Lee
558fc7f293 maskrom: retain data for d channel is not ready 2017-08-07 12:17:10 -07:00
Andrew Waterman
7fd8bb1159 Merge pull request #928 from freechipsproject/critical-paths
Critical paths
2017-08-06 18:50:59 -07:00
Andrew Waterman
658e36f98b Reduce fanout on frontend io.cpu.req.valid signal 2017-08-06 17:38:51 -07:00
Andrew Waterman
7d94074b05 Remove one gate from D$ ECC check
The D$ corrects via writeback, so which word the error was in doesn't
matter, as the entire line is corrected.
2017-08-06 17:36:53 -07:00
Wesley W. Terpstra
d03fdc4f30 diplomacy: seal the LazyModuleImpLike trait (#927)
This makes sure that all the base classes call instantiate()
2017-08-06 17:32:23 -07:00
Yunsup Lee
aa60c6944b diplomacy: provide default clock/reset for LazyRawModuleImp 2017-08-06 13:40:07 -07:00
Andrew Waterman
83875e3a0c Only flush D$ on FENCE.I if it won't always be probed on I$ miss 2017-08-05 14:22:40 -07:00
Andrew Waterman
991e16de92 Remove probe address mux from TLB response path 2017-08-05 12:57:38 -07:00
Andrew Waterman
b9b4142bb4 Get s2_nack off the critical path
We were using it to compute the next PC on flush vs. replay (which require
PC+4 and PC, respectively).  This fix gets rid of the adder altogether by
reusing the M-stage PC in the flush case, which by construction holds PC+4.
2017-08-05 00:30:36 -07:00
Andrew Waterman
bc298bf146 Optimize ShiftQueue for late-arriving deq.ready 2017-08-04 22:06:37 -07:00
Andrew Waterman
6112adfbb0 Get L2 TLB tag/parity check off the D$ arbitration path 2017-08-04 17:01:51 -07:00
Andrew Waterman
8d97684555 Fix L2 TLB perfctr
It was counting conflict misses but not cold misses.
2017-08-04 17:01:31 -07:00
Andrew Waterman
df7f09b9ce Get I$ ECC check further off critical path 2017-08-04 16:59:21 -07:00
Andrew Waterman
4bfbe75d74 Avoid pipeline replays when fetch queue is full 2017-08-04 16:59:21 -07:00
Andrew Waterman
a45997d03f Separate I$ parity error from miss signal
Handle parity errors with a pipeline flush rather than a faster
frontend replay, reducing a critical path.
2017-08-04 16:59:21 -07:00
Andrew Waterman
06a831310b Shave a gate delay off I$ backpressure path
The deleted code was a holdover from Hwacha's vector fences.
2017-08-04 13:12:43 -07:00
Andrew Waterman
ecc2ee366c Shave a few gate delays off IBuf control logic
It takes a while for the pipeline to compute the stall signal, so avoid
using it until the last logic levels in the clock cycle.
2017-08-04 13:12:43 -07:00
Andrew Waterman
7937db0c84 Merge pull request #919 from freechipsproject/imiss-perf-counter
Fix I$ miss perfctr
2017-08-04 01:04:23 -07:00
Megan Wachs
017ac130c1 Merge pull request #922 from freechipsproject/bigger_tl_xbar
TLXbar: Allow more masters and slaves and issue a warning.
2017-08-03 16:52:56 -07:00
Megan Wachs
50c85f1b62 TLXbar: Allow more masters and slaves and issue a warning. 2017-08-03 15:46:06 -07:00
Andrew Waterman
ba4eecc0f0 Use UIntToOH1 (#921)
Closes #920
2017-08-03 14:55:39 -07:00
Andrew Waterman
f483bab4aa Fix I$ miss perfctr
The old version was counting prefetches, too.
2017-08-03 00:52:12 -07:00
Andrew Waterman
1be1433f04 Merge pull request #918 from freechipsproject/icache-prefetch
Icache prefetch
2017-08-02 21:22:20 -07:00
Andrew Waterman
d66e8f8e80 Merge pull request #914 from freechipsproject/critical-paths
Fix some critical paths
2017-08-02 19:05:31 -07:00
Megan Wachs
3fc7100048 Merge pull request #917 from freechipsproject/fuzzer_order
TLFuzzer: Allow Ordered clients to be created as well by the fuzzer
2017-08-02 18:39:59 -07:00
Andrew Waterman
2537d0d54e Optionally prefetch next I$ line into L2$ on miss 2017-08-02 17:10:56 -07:00
Andrew Waterman
744cdb2f72 Make TLB report when it's safe to prefetch within a page 2017-08-02 17:09:38 -07:00
Megan Wachs
595415d207 TLFuzzer: Correct the number of ordered clients created 2017-08-02 15:48:21 -07:00
Megan Wachs
fc5c04ed4b TLFuzzer: Allow Ordered clients to be created as well by the fuzzer 2017-08-02 14:44:18 -07:00
Andrew Waterman
7d2dd3769f Optimize a hazard check critical path 2017-08-02 14:27:25 -07:00
Megan Wachs
85bdae0fa8 diplomacy: Pretty Print for TransferSizes 2017-08-02 11:40:50 -07:00
Andrew Waterman
2eb239d03f Add option to retime D$ way mux into subsequent pipeline stage 2017-08-01 23:59:20 -07:00
Andrew Waterman
9464c6db40 Mitigate(?) frontend critical path 2017-08-01 18:51:17 -07:00
Andrew Waterman
735701382f Mitigate some I$ response valid critical paths 2017-08-01 18:51:17 -07:00
Andrew Waterman
2ecea2ef60 Don't use a pipe queue on D$ TL A-channel
This cuts an I$->D$ path.
2017-08-01 15:17:07 -07:00
Yunsup Lee
6ef8ee5d4d tilelink: add mask rom 2017-07-31 21:34:04 -07:00
Yunsup Lee
4b33249812 Merge pull request #911 from freechipsproject/fix-dcache-bug
Fix D$ ready-valid signaling bug
2017-07-31 19:14:16 -07:00
Wesley W. Terpstra
42ff74bd34 Merge pull request #910 from freechipsproject/tilelink-map
Tilelink map
2017-07-31 18:33:09 -07:00
Andrew Waterman
e140893a01 Use 1-entry queue on processor-side E-channel
The cache can't sink a grant every cycle, so extra E buffering doesn't help.
2017-07-31 18:06:54 -07:00
Andrew Waterman
5681693ccc Fix a D$ ready-valid signaling regression
I broke this in 66d06460fa.
2017-07-31 18:05:14 -07:00
Wesley W. Terpstra
d7fd9d2b82 tilelink: Filter, add another case 2017-07-31 16:51:26 -07:00
Yunsup Lee
71a250b071 Merge pull request #909 from freechipsproject/tile-buffer
add optional tile boundary buffers
2017-07-31 16:46:22 -07:00
Wesley W. Terpstra
b126105230 tilelink: add TLMap to make it possible to move slaves 2017-07-31 16:39:00 -07:00
Wesley W. Terpstra
13d3ffbcaa tilelink: Filter now support arbitrary filter functions 2017-07-31 16:38:38 -07:00
Yunsup Lee
7adfd5c431 Merge pull request #906 from freechipsproject/critical-paths
Mitigate I$->D$->I$ critical path
2017-07-31 16:14:11 -07:00
Yunsup Lee
f473e6bad0 tile: add optional boundary buffers 2017-07-31 15:57:22 -07:00
Yunsup Lee
cb3529bbc3 util: tweak rational crossings to avoid mux in source 2017-07-31 15:10:15 -07:00
Henry Cook
11332c1226 dcache: break potential combinatorial loop by making pstore_drain_on_miss more conservative 2017-07-31 14:03:30 -07:00
Andrew Waterman
d811692c3b Mitigate I$->D$->I$ critical path
This seemingly irrelevant change shaves several gate delays off the I$
tl.a.valid path.
2017-07-31 01:43:04 -07:00
Andrew Waterman
ac4339a8e7 Pass D$ backpressure to D-channel, rather than asserting 2017-07-29 11:48:36 -07:00
Andrew Waterman
edcd2c696c Avoid needless stall on E-channel back pressure 2017-07-29 11:47:58 -07:00
Wesley W. Terpstra
8e2e931770 Merge pull request #903 from freechipsproject/monitor-probes
tilelink: use the Monitor to enforce Probe sourcing
2017-07-29 01:12:08 -07:00
Wesley W. Terpstra
56e28026a6 TLError: does not need to be fast; cut the loop
The SystemBus already has a flow buffer on outputs.
2017-07-29 00:22:21 -07:00
Wesley W. Terpstra
540256e24a systembus: all slaves should have an output buffer 2017-07-29 00:13:33 -07:00
Wesley W. Terpstra
eadf4e9fcc Revert "tile: add option for tile boundary buffers"
This reverts commit b64b87ad07.

The crossings already have buffering in those places where it was
appropriate. Adding more does not help flow through paths.
2017-07-29 00:03:24 -07:00
Wesley W. Terpstra
68064ba260 systembus: don't double down on buffers
The order should be:
  master => buffer|xing => fifofixer => splitter => xbar
2017-07-29 00:02:12 -07:00
Yunsup Lee
140086e2c5 Merge pull request #902 from freechipsproject/perf-improvements
Perf improvements
2017-07-28 20:12:10 -07:00
Wesley W. Terpstra
a0db929003 tilelink: use the Monitor to enforce Probe sourcing 2017-07-28 18:08:00 -07:00
Megan Wachs
573890e102 Merge pull request #900 from freechipsproject/more_verbose_requires
diplomacy: More verbose require
2017-07-28 13:23:33 -07:00
Andrew Waterman
fdb8935712 Improve fidelity of two perf counters 2017-07-28 13:14:04 -07:00
Andrew Waterman
4c82f6b77e Don't refill BTB on not-taken branches 2017-07-28 13:13:52 -07:00
Andrew Waterman
2e8b02e780 Merge D$ store hits when ECC is enabled
This avoids pipeline flushes due to subword WAW hazards, as with
consecutive byte stores.
2017-07-28 12:56:36 -07:00
Andrew Waterman
838864870e Bypass TLB refill signal to halve L2 TLB hit time
The 4-cycle hit time is 1 cycle too long to avoid a second
pipeline replay, so it was effectively 9 cycles instead of 4.
2017-07-28 12:56:36 -07:00
Andrew Waterman
ae1f7a95f6 Don't nack misses when there's a pending store
That effectively increased the miss latency by 5 cycles when there was
a store hit followed by a load miss.  Since pending stores are drained
when releaseInFlight, the check I removed was redundant.
2017-07-28 12:56:36 -07:00
Henry Cook
7eeb9dfd88 Merge pull request #899 from freechipsproject/wrapper-dedup
Stabilize tile wrappers for downstream tools
2017-07-28 10:52:59 -07:00
Megan Wachs
f61fe2be1e diplomacy: More verbose require 2017-07-28 10:05:45 -07:00
Wesley W. Terpstra
5f81c2243f tilelink: add BusBypass, useful to turn devices off 2017-07-27 20:16:30 -07:00
Wesley W. Terpstra
9a36755b6a tilelink: CacheCork uses constructor helpers 2017-07-27 18:38:15 -07:00
Wesley W. Terpstra
45189c3e30 tilelink: CacheCork now supports errors and BtoT upgrade
- Acquire.BtoT succeeds with toT instantly
- AccessAckData.error causes Grant.toN.error
2017-07-27 18:38:13 -07:00
Wesley W. Terpstra
2e4f1611ed tilelink: Error device supports Acquire
We need this if we want to divert traffic to it from a TL-C slave.
2017-07-27 18:32:58 -07:00
Henry Cook
b64b87ad07 tile: add option for tile boundary buffers 2017-07-27 17:30:51 -07:00
Henry Cook
289ef30dbc coreplex: change AsynchronousCrossing.sync default to 3 2017-07-27 15:44:51 -07:00
Henry Cook
266ed56e8d tile: turn off more slave port monitors 2017-07-27 15:28:53 -07:00
Henry Cook
9a483af6e8 coreplex: naming of tile wrappers 2017-07-27 15:16:48 -07:00
Henry Cook
33852ef965 coreplex: remove superfluous sink and source from wrapper 2017-07-27 14:23:03 -07:00
Wesley W. Terpstra
651da73d89 tilelink: it is now legal to support Acquire for UNCACHED regions
These cases exist:
  GET_EFFECTS, PUT_EFFECTS, UNCACHEABLE && !supportsAcquire: MMIO
  UNCACHED && !supportsAcquire: speculation ok and may be cached
  UNCACHED && supportsAcquire: LLC/CacheCork applied (slave never probes)
  CACHED, TRACKED && supportsAcquire: slave might probe
2017-07-27 11:11:22 -07:00
Wesley W. Terpstra
0ab5cb67b3 tilelink: fix RAMModel handling of AMOs on early source reuse (#897) 2017-07-27 11:07:13 -07:00
Wesley W. Terpstra
9804bdc34e tilelink: remove obsolete addr_lo signal (#895)
When we first implemented TL, we thought this was helpful, because
it made WidthWidgets stateless in all cases. However, it put too
much burden on all other masters and slaves, none of which benefitted
from this signal. Furthermore, even with addr_lo, WidthWidgets were
information lossy because when they widen, they have no information
about what to fill in the new high bits of addr_lo.
2017-07-26 16:01:21 -07:00
Wesley W. Terpstra
d096d5d1c4 tilelink: fix AtomicAutomata bug wrt early source reuse
The new fuzzer already found it's first victim.
2017-07-26 12:52:29 -07:00
Wesley W. Terpstra
6550ae2e31 tilelink: increase Fuzzer source reuse aggression 2017-07-26 12:37:31 -07:00
Wesley W. Terpstra
1efdca106c tilelink: RAMModel support early reuse of source 2017-07-26 12:37:31 -07:00
Wesley W. Terpstra
138276fd87 tilelink: SourceShrinker should work also for 0 latency 2017-07-26 12:37:31 -07:00
Wesley W. Terpstra
b2edca2a6b tilelink: cut WidthWidget from dependency on addr_lo 2017-07-26 10:31:09 -07:00
Wesley W. Terpstra
ede87c1f73 tilelink: rewrite WidthWidget beat splitter
- split the data based on the address, not the mask
  (the first version of TileLink did not have low address bits)
- the dependency on addr_lo is now exposed and easy to replace
2017-07-26 10:24:16 -07:00
Wesley W. Terpstra
0f5065fbf3 tilelink: WidthWidget rewrite beat merging
- errors are properly OR reduced
- registers latched only as needed (was previously a shift register)
- combines beats without inspecting address (removes addr_lo dependency)
2017-07-26 10:24:12 -07:00
Wesley W. Terpstra
f0ffb7e31e tilelink: initialize toggle in Fragmenter (#894)
No strictly necessary, because the initial value does not matter, but good hygiene since it drives a cycle of logic.
2017-07-26 10:21:31 -07:00
Andrew Waterman
5a5b78b15e Improve L2 TLB coding style 2017-07-26 02:22:43 -07:00
Andrew Waterman
5a9c673f41 Fix L2 TLB response bug
Sometimes, it would inform the L1 TLB that the translation was for
a superpage, even though that's never the case.
2017-07-26 02:20:41 -07:00
Andrew Waterman
acca0fccf5 Fix BTB not being refilled on some indirect jumps
We are overloading the BTB-hit signal to mean that any part of the frontend
changed the control-flow, not just the BTB.  That's the right thing to do for
most of the control logic, but it means the BTB sometimes won't get refilled
when we'd like it to.  This commit makes the frontend use an invalid BTB entry
number when it, rather than the BTB, changes the control flow.  Since the
entry number is invalid, the BTB will treat it as a miss and refill itself.

This is kind of a hack, but a more palatable fix requires reworking the RVC
IBuf, which I don't have time for right now.
2017-07-26 02:13:43 -07:00