Cheri-research/Toooba
1
0
Fork 0
Code Issues 1 Pull Requests Actions Packages Projects Releases Wiki Activity
471 Commits 7 Branches 0 Tags
c515fb4518d3f4fa2c7b417ea1803deafae03d05
Commit Graph

318 Commits

Author SHA1 Message Date
jon
c515fb4518 Remove commented-out code. 
Enable building with open-source BSC.

Resolve a couple issues brought up by Jess.

Remove a stray space (as noted by Jess).
 2021-01-14 17:10:04 +00:00
jon
a6771219ba Fold the fragment count into the standard pipeline structs. 2021-01-14 12:29:38 +00:00
jon
fc968c867f PC compression scheme for the instruction fetch pipeline. 
Store the upper bits of the PC in a table and only handle indices and
lsbs in the main pipeline.
This eliminates redundancy between PCs and predicted PCs, and even more
between fragments of instructions.
 2021-01-14 11:41:11 +00:00
jon
dd82b2703c Record an exception from either half of an instruction, with preference 
for the first half.
 2021-01-13 15:33:08 +00:00
jon
0d93ed3ad2 Merge branch 'CHERI' into ifetch-cleanup 2021-01-13 15:21:20 +00:00
Jessica Clarke
4a0c8fe7d2 PLIC: Update comment after previous commit 2020-12-21 13:37:01 +00:00
Jessica Clarke
ac6244181a PLIC: Allow multiple interrupts to be claimed but not completed 
Whilst the current PLIC spec is hard to decipher, the older v1.9
privileged spec is clearer. The intent is that each interrupt is
independent and so claims and completions can be overlapped (e.g. as
done by FreeBSD if deferring an interrupt handler). This is
approximately expressed in the spec by:

  The PLIC can perform a claim at any time and the claim operation is
  not affected by the setting of the priority threshold register.

  ...

  The PLIC does not check whether the completion ID is the same as the
  last claim ID for that target. If the completion ID does not match an
  interrupt source that is currently enabled for the target, the
  completion is silently ignored.

Thost last two sentences make it sound like the completion ID not
matching the last claim ID is a thing to ignore, but in fact they are
entirely independent statements. This can be confirmed by reading the
source for the Rocket and QEMU PLICs, which have none of the stricter
requirements of the Bluespec PLIC implementation.

This should fix running FreeBSD after r362977, which started to
correctly delay the completion of interrupts until after deferred
handlers had completed (as opposed to after they had been queued).
 2020-12-21 13:32:06 +00:00
Jessica Clarke
6b55d30a8e PLIC: Restore upstream whitespace 2020-12-21 13:32:06 +00:00
Jessica Clarke
e3f94224de Revert "Cause two surprising conditions to not throw slave errors as this is not" 
This reverts commit 4131dcd073.
 2020-12-21 13:03:13 +00:00
Jonathan Woodruff
92ca72a6e2 Fix build error by providing default. 2020-12-18 17:55:18 +00:00
Jonathan Woodruff
84d60d03b1 Merge pull request #6 from nwf/capdirty 
CHERI ISAv8 CapDirty fault priority fix
 2020-12-18 16:48:56 +00:00
jon
f86d4d9dc0 Add counter support to ITLB. 2020-12-17 17:57:43 +00:00
jon
4ddcaff497 Support for data TLB counters. 2020-12-17 16:52:45 +00:00
Jonathan Woodruff
4dd0a73051 Port to internal AXI interfaces for the debug module. 2020-12-17 15:02:05 +00:00
jon
bab989d6d2 Support for stat counters from the tag controller. 
The order of the events is very not tested; 50% chance of being correct.
 2020-12-17 12:38:52 +00:00
jon
4131dcd073 Cause two surprising conditions to not throw slave errors as this is not 
specified and causes kernel panics.  Hopefully the details of what
should happen in these cases is correct, but we shall see...
 2020-12-17 12:37:45 +00:00
jon
fa9931b64c Implement a couple Cache counters in the LLCache. 
These aren't too useful... We don't seem to have normal counters for
loads and stores in the baseline.
 2020-12-15 18:16:21 +00:00
jon
17a7a32092 Use DRegOR instead of immitating latching behaviour. 2020-12-15 16:18:36 +00:00
jon
b6a397df52 Support for ICache stat counters. 2020-12-15 14:49:16 +00:00
jon
73d25bf8f5 Support performance counters (hopefully) in the caches. The DCache 
should be fully wired up.
 2020-12-14 18:10:06 +00:00
Jessica Clarke
f6569e2d14 Actually do the intended fix, not just inverting the variable 2020-12-14 10:51:57 +00:00
Jessica Clarke
295b878836 Fix negated CYCLE/INSTRET logic for ASR and make less confusing 2020-12-14 10:50:23 +00:00
jon
6768bf92c7 iMove the store count to the pipeline write. 
It appeared to be counting capabilities in store counter flushes, which,
FI think, is not quite what was intended in this counter.
 2020-12-11 17:16:38 +00:00
jon
19a07882ab CSome style cleanups from chat with Alexandre. 2020-12-11 16:44:59 +00:00
jon
1bae03da9c Implement all the remaining core (not cache) memory stat counters. 2020-12-11 15:48:10 +00:00
jon
d6243be1dd Add a few more core performance counters. 2020-12-04 18:34:56 +00:00
jon
bdc2d494f2 Change name something that makes sense. 2020-12-04 17:12:26 +00:00
jon
3f60c046f7 Fix typo. 2020-12-04 16:51:44 +00:00
jon
f42c3f3577 Merge branch 'cheri-perfmon' into CHERI 2020-12-04 16:29:51 +00:00
jon
5665d2e9cd A more effective Fence count (as some fences go through as the Nop 
iType).
 2020-12-04 16:28:19 +00:00
jon
1fb388fb18 Count more things correctly. 2020-12-03 17:54:46 +00:00
jon
4fcc18635a Supporting most architectural counters, wiring them out from 
CommitStage.
Also support multiple events per cycle.
 2020-12-02 15:28:41 +00:00
jon
e99e1e6274 Update CSR names to fullish standard names. 2020-12-02 10:18:25 +00:00
jon
259d34618c A design that actually passes one performance monitor trace from 
TestRig.  The example counted redirections, which happend to match
between Flute and Toooba for this example.
 2020-12-01 18:02:11 +00:00
Alexandre Joannou
98a9c076cd Update BlueStuff API and Bump BlueStuff and TagController 2020-12-01 14:22:12 +00:00
Jonathan Woodruff
0289bfe17d Fix a couple errors due to changes somehow merged over from the rich_predicted_state branch. 2020-11-09 11:14:11 +00:00
Nathaniel Filardo
6e4c378d4d CHERI ISAv8 CapDirty fault priority fix 
Raise a store page fault if the PTE is missing R or W, even if the store is a
cap store and the PTE is also missing CW, as per the sail spec.

Co-authored-by: Jessica Clarke <jrtc27@jrtc27.com>
 2020-11-06 19:15:47 +00:00
Jonathan Woodruff
69c697daf7 Changes needed to build for FPGA. 2020-11-06 11:44:33 +00:00
Peter Rugg
bf77ac74ba Fix performance reporting requests for multicore 2020-11-04 12:11:31 +00:00
Jonathan Woodruff
29168574ca Revert "Sort of support 2 cores for starting and halting." 
This reverts commit b4b7718aa2.
 2020-10-16 15:44:53 +01:00
jon
b4b7718aa2 Sort of support 2 cores for starting and halting. 2020-10-08 15:09:23 +01:00
Nathaniel Filardo
d4308f0f03 Slightly rework CHERI PTE to add CapDirty flag 
Permissions are consistently checked before dirty flags, now, too.
 2020-09-30 15:40:50 +01:00
jon
9508eada79 Change to allow 1-way superscalar to build without error. 2020-09-17 14:40:46 +01:00
jon
1568fcf485 Enable larger than 2x superscalar by using explicit ports into the BTB. 
(Previously we ran into the limit of ports on the regFiles holding state
in the BTB.  This uses a vector of regFiles with a vector of interfaces,
which is logically the same but might produce multiple copies of state.)
 2020-09-16 11:42:19 +01:00
jon
96afb0a7f5 FIFO optimisation both for build time (and possibly condition 
complexity?) and for throughput.
 2020-08-13 17:33:46 +01:00
jon
4c74d27e4e Merge branch 'CHERI' into ifetch-cleanup 2020-08-11 15:31:30 +01:00
jon
7a93056b8b Increase the capacity of f32d to enable full throughput. 
I don't know why this is necessary, but this allows performance parity
with the baseline.
 2020-08-11 15:30:03 +01:00
jon
9230fcc3b4 Merge branch 'CHERI' into ifetch-cleanup 2020-08-11 15:25:12 +01:00
jon
00ffdd7228 Enable to build with latest BSC compiler. 2020-08-11 14:46:28 +01:00
jon
3b97fbe255 Rename last->prev for clarity. 2020-07-31 13:46:33 +01:00