1
0
Commit Graph

1066 Commits

Author SHA1 Message Date
Andrew Waterman
b92c73e361 Add LR/SC to blocking D$ 2016-05-24 15:05:41 -07:00
Andrew Waterman
0d93d1a1a0 Clean up pending store logic a bit 2016-05-24 15:05:41 -07:00
Andrew Waterman
0b8de578d4 Add additional D$ store buffering to prevent structural hazards 2016-05-24 15:05:41 -07:00
Andrew Waterman
354cb2d5ec Don't stall I$ response when resolving a branch misprediction
This avoids a fetch bubble.

Not clear if this is the best way to do it.  Perhaps this change should
instead be made to Frontend (i.e., ignore resp.ready when req.valid is
high), but that might exacerbate a critical path.
2016-05-24 15:05:41 -07:00
Andrew Waterman
d7790ac6a4 WIP on blocking D$ 2016-05-24 15:05:41 -07:00
Andrew Waterman
335e2c8a1e Support disabling atomics extension 2016-05-24 15:05:41 -07:00
Andrew Waterman
765b90f6a4 Stall on D$ lockups less conservatively 2016-05-24 15:05:41 -07:00
Andrew Waterman
a3061047e3 Instantiate blocking D$ when NMSHRS=0 2016-05-24 15:05:41 -07:00
Andrew Waterman
80482890fd Don't rely on tag value for nacks 2016-05-24 15:05:41 -07:00
Wesley W. Terpstra
e19c5e5d2c IOMSHR: support atomic operations 2016-05-24 15:00:50 -07:00
Andrew Waterman
7bc38383de add (non-working) blocking data cache 2016-05-20 18:59:05 -07:00
Howard Mao
f228309bd1 add assertion to make sure SimpleHellaCacheIF doesn't get exception 2016-05-20 16:30:27 -07:00
Andrew Waterman
4aef567a80 Fix MMIO bug: replay_next wasn't set 2016-05-13 17:59:53 -07:00
Andrew Waterman
742c05d6a7 Pipeline D$->I$ control paths
These stretch the miss latency by a cycle in exchange for slack.
The current implementation also adds a cycle to mul/div latency,
which can be worked around for more hardware (possibly gated by
the FastMulDiv option).
2016-05-13 17:07:28 -07:00
Colin Schmidt
8fa2de0816 chisel3 fix to RoCC connections honor last connect 2016-05-05 18:09:48 -07:00
Andrew Waterman
9dd23a603a Remove HTIF port 2016-05-03 13:41:58 -07:00
Andrew Waterman
5352497edb MPRV takes effect regardless of privilege mode 2016-05-02 19:53:25 -07:00
Howard Mao
5cbcc41515 get rid of unused imports 2016-05-02 18:23:46 -07:00
Andrew Waterman
f784f4da93 Rename PRCICoreIO to PRCITileIO 2016-05-02 18:08:01 -07:00
Andrew Waterman
000e20f937 Remove MIPI; make mip.MSIP read-only
The PRCI block outside the core will provide IPIs eventually
2016-05-02 15:18:41 -07:00
Andrew Waterman
83fa489cef Stop using HTIF CSR port
The port itself is still present to keep other stuff compiling.
2016-05-02 14:40:52 -07:00
Albert Ou
0ff4fd0ccd Fix IOMSHR to send finishes for stores 2016-04-30 22:20:29 -07:00
Andrew Waterman
491184a8f8 ERET -> xRET; remove mcfgaddr 2016-04-30 17:32:51 -07:00
Andrew Waterman
5af98145b9 don't signal bad physical address on TLB miss 2016-04-30 17:31:46 -07:00
Andrew Waterman
cae4265f3b Change mcfgaddr pointer 2016-04-28 16:14:05 -07:00
Andrew Waterman
739cf07637 Remove mtime/mtimecmp
The RTC is now a device that lives on the MMIO bus.
2016-04-27 14:54:51 -07:00
Andrew Waterman
fb5c38c186 Handle invalidate_lr in cache arbiter, not tile 2016-04-27 11:22:04 -07:00
Andrew Waterman
b99db83e67 Avoid needless Vec generation 2016-04-27 00:28:39 -07:00
Andrew Waterman
8acec8eb36 Remove dead code from BTB 2016-04-27 00:28:12 -07:00
Andrew Waterman
fe8c91f620 Fix IOMSHR state machine bug
Sending the finish too early causes the CPU response to get dropped.

attn @zhemao
2016-04-26 15:32:25 -07:00
Andrew Waterman
5fd5b58743 Remove stats CSR 2016-04-26 15:31:32 -07:00
Andrew Waterman
d93677a343 Support larger cache sets when not using VM 2016-04-26 15:31:32 -07:00
Yunsup Lee
5dbf9640e2 Use TLB flush signal to I$ explicitly 2016-04-22 15:41:31 -07:00
Andrew Waterman
84fd45fd77 Pass TLB flush signal to I$ explicitly 2016-04-22 15:20:17 -07:00
Howard Mao
b7527268bb use address map instead of MMIOBase to find size of memory 2016-04-21 18:44:39 -07:00
Christopher Celio
2d6f35525e Added Field[Int] to SFMALatency/DFMALatency params 2016-04-06 14:50:57 -07:00
Andrew Waterman
51e0870e23 Separate I$ and D$ interface signals that span clock cycles
For example, Decopuled[HellaCacheReq].bits.kill doesn't make sense,
since it doesn't come the same cycle as ready/valid.
2016-04-01 19:30:39 -07:00
Andrew Waterman
dc662f28a0 Specify width on s1_pc to avoid width inference problem 2016-04-01 17:28:42 -07:00
Andrew Waterman
72f7f71eb5 No need to allow finishes to be sent in s_refill_resp state
This is a hold-over from when writebacks needed finish messages.
2016-04-01 16:19:57 -07:00
Henry Cook
78bc18736e LRSC startvation fix: HellaCache generates its own Finish messages again. 2016-04-01 16:04:25 -07:00
Andrew Waterman
37b9051762 No need to validate npc if BTB is disabled 2016-04-01 15:54:57 -07:00
Andrew Waterman
4480d1e817 Don't compile BTB when nEntries=0 2016-04-01 15:14:45 -07:00
Andrew Waterman
d406dc1231 Remove vestigial BTB enable option 2016-04-01 15:14:34 -07:00
Henry Cook
54dd82ff76 bugfix for WB data buffer 2016-03-31 17:53:49 -07:00
Christopher Celio
1792d01ce1 fix leaky assert in nbdcache
Squash of #33.
2016-03-31 15:56:14 -07:00
Andrew Waterman
adb7eacf6e Fix Chisel3 build for XLen=32 2016-03-30 22:48:51 -07:00
Andrew Waterman
70664bbca0 Fix Chisel3 build for UseVM=false 2016-03-30 22:48:31 -07:00
Andrew Waterman
8ad8e8a691 Add partial Sv48/Sv57 support
Right now, we don't support Sv39 and Sv48 at the same time, which needs
to change.
2016-03-30 11:02:22 -07:00
Andrew Waterman
e652821962 Use correct kind of TileLink arbiter
It was "correct" before, but broke Chisel3 build.
2016-03-28 22:53:47 -07:00
Andrew Waterman
5ce3527b88 Merge pull request #32 from ucb-bar/pr-btb-masking
separate btb response mask from the frontend mask
2016-03-26 18:15:14 -07:00