Libre RISC-V Mailing list Archives (accessible via public-inbox)
 help / color / Atom feed
* [libre-riscv-dev] [Bug 348] New: POWER9 SPR pipeline needed
From: bugzilla-daemon @ 2020-05-24 21:46 UTC (permalink / raw)
  To: libre-riscv-dev

https://bugs.libre-soc.org/show_bug.cgi?id=348

            Bug ID: 348
           Summary: POWER9 SPR pipeline needed
           Product: Libre-SOC's first SoC
           Version: unspecified
          Hardware: PC
                OS: Mac OS
            Status: CONFIRMED
          Severity: enhancement
          Priority: ---
         Component: Source Code
          Assignee: lkcl@lkcl.net
          Reporter: lkcl@lkcl.net
                CC: libre-riscv-dev@lists.libre-riscv.org
   NLnet milestone: ---

OP_MTSPR, OP_MFSPR operations to be added to an SPR pipeline.
with some of the SPRs being in a separate regfile (XER, LR, CTR, SRR1, SRR2)
this pipeline is slightly more involved than it seems.   it also requires
quite a lot of regfile ports.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev

^ permalink raw reply

* [libre-riscv-dev] [Bug 348] POWER9 SPR pipeline needed
From: bugzilla-daemon @ 2020-05-24 21:49 UTC (permalink / raw)
  To: libre-riscv-dev
In-Reply-To: <bug-348-13@https.bugs.libre-soc.org/>

https://bugs.libre-soc.org/show_bug.cgi?id=348

--- Comment #1 from Luke Kenneth Casson Leighton <lkcl@lkcl.net> ---
OP_MFSPR - note that XER is "constructed"

report "MFSPR to SPR " & integer'image(decode_spr_num(e_in.insn)) &
    "=" & to_hstring(a_in);
if is_fast_spr(e_in.read_reg1) then
    result := a_in;
    if decode_spr_num(e_in.insn) = SPR_XER then
        -- bits 0:31 and 35:43 are treated as reserved and return 0s
        -- when read using mfxer
        result(63 downto 32) := (others => '0');
        result(63-32) := v.e.xerc.so;
        result(63-33) := v.e.xerc.ov;
        result(63-34) := v.e.xerc.ca;
        result(63-35 downto 63-43) := "000000000";
        result(63-44) := v.e.xerc.ov32;
        result(63-45) := v.e.xerc.ca32;
else
    case decode_spr_num(e_in.insn) is
    when SPR_TB =>              result := ctrl.tb;
    when SPR_DEC =>             result := ctrl.dec;
    when others =>              result := (others => '0');
    end case;

-- 
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev

^ permalink raw reply

* [libre-riscv-dev] [Bug 349] New: privileged-instruction decoding function needed
From: bugzilla-daemon @ 2020-05-24 21:51 UTC (permalink / raw)
  To: libre-riscv-dev

https://bugs.libre-soc.org/show_bug.cgi?id=349

            Bug ID: 349
           Summary: privileged-instruction decoding function needed
           Product: Libre-SOC's first SoC
           Version: unspecified
          Hardware: PC
                OS: Mac OS
            Status: CONFIRMED
          Severity: enhancement
          Priority: ---
         Component: Source Code
          Assignee: lkcl@lkcl.net
          Reporter: lkcl@lkcl.net
                CC: libre-riscv-dev@lists.libre-riscv.org
   NLnet milestone: ---

certain operations are privileged.  in the decoder these need to be
marked as such, and a trap raised instead.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev

^ permalink raw reply

* [libre-riscv-dev] [Bug 349] privileged-instruction decoding function needed
From: bugzilla-daemon @ 2020-05-24 21:51 UTC (permalink / raw)
  To: libre-riscv-dev
In-Reply-To: <bug-349-13@https.bugs.libre-soc.org/>

https://bugs.libre-soc.org/show_bug.cgi?id=349

--- Comment #1 from Luke Kenneth Casson Leighton <lkcl@lkcl.net> ---
from https://github.com/antonblanchard/microwatt/blob/master/execute1.vhdl

    type privilege_level is (USER, SUPER);
    type op_privilege_array is array(insn_type_t) of privilege_level;
    constant op_privilege: op_privilege_array := (
        OP_ATTN => SUPER,
        OP_MFMSR => SUPER,
        OP_MTMSRD => SUPER,
        OP_RFID => SUPER,
        others => USER
        );

    function instr_is_privileged(op: insn_type_t; insn: std_ulogic_vector(31
downto 0))
        return boolean is
    begin
        if op_privilege(op) = SUPER then
            return true;
        elsif op = OP_MFSPR or op = OP_MTSPR then
            return insn(20) = '1';
        else
            return false;
    end;

-- 
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev

^ permalink raw reply

* [libre-riscv-dev] [Bug 350] New: LDSTCompUnit also needs to support zeroing on RA
From: bugzilla-daemon @ 2020-05-25  0:13 UTC (permalink / raw)
  To: libre-riscv-dev

https://bugs.libre-soc.org/show_bug.cgi?id=350

            Bug ID: 350
           Summary: LDSTCompUnit also needs to support zeroing on RA
           Product: Libre-SOC's first SoC
           Version: unspecified
          Hardware: Other
                OS: Linux
            Status: CONFIRMED
          Severity: enhancement
          Priority: ---
         Component: Source Code
          Assignee: lkcl@lkcl.net
          Reporter: lkcl@lkcl.net
                CC: libre-riscv-dev@lists.libre-riscv.org
   NLnet milestone: ---

https://libre-soc.org/openpower/isa/fixedload/

noted that zeroing on RA is possible

-- 
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev

^ permalink raw reply

* [libre-riscv-dev] [Bug 350] LDSTCompUnit also needs to support zeroing on RA
From: bugzilla-daemon @ 2020-05-25  0:14 UTC (permalink / raw)
  To: libre-riscv-dev
In-Reply-To: <bug-350-13@https.bugs.libre-soc.org/>

https://bugs.libre-soc.org/show_bug.cgi?id=350

Luke Kenneth Casson Leighton <lkcl@lkcl.net> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           See Also|                            |https://bugs.libre-soc.org/
                   |                            |show_bug.cgi?id=318

-- 
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev

^ permalink raw reply

* [libre-riscv-dev] [Bug 318] fix LDSTCompUnit
From: bugzilla-daemon @ 2020-05-25  0:14 UTC (permalink / raw)
  To: libre-riscv-dev
In-Reply-To: <bug-318-13@https.bugs.libre-soc.org/>

https://bugs.libre-soc.org/show_bug.cgi?id=318

Luke Kenneth Casson Leighton <lkcl@lkcl.net> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           See Also|                            |https://bugs.libre-soc.org/
                   |                            |show_bug.cgi?id=350

-- 
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev

^ permalink raw reply

* [libre-riscv-dev] [Bug 350] LDSTCompUnit also needs to support zeroing on RA
From: bugzilla-daemon @ 2020-05-25  0:14 UTC (permalink / raw)
  To: libre-riscv-dev
In-Reply-To: <bug-350-13@https.bugs.libre-soc.org/>

https://bugs.libre-soc.org/show_bug.cgi?id=350

Luke Kenneth Casson Leighton <lkcl@lkcl.net> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           See Also|                            |https://bugs.libre-soc.org/
                   |                            |show_bug.cgi?id=336
    NLnet milestone|---                         |NLnet.2019.02

-- 
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev

^ permalink raw reply

* [libre-riscv-dev] [Bug 336] ALU CompUnit needs to recognise that RA (src1) can be zero
From: bugzilla-daemon @ 2020-05-25  0:14 UTC (permalink / raw)
  To: libre-riscv-dev
In-Reply-To: <bug-336-13@https.bugs.libre-soc.org/>

https://bugs.libre-soc.org/show_bug.cgi?id=336

Luke Kenneth Casson Leighton <lkcl@lkcl.net> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           See Also|                            |https://bugs.libre-soc.org/
                   |                            |show_bug.cgi?id=350

-- 
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev

^ permalink raw reply

* [libre-riscv-dev] [Bug 335] Formal Correctness Proof for Branch pipeline
From: bugzilla-daemon @ 2020-05-25  0:22 UTC (permalink / raw)
  To: libre-riscv-dev
In-Reply-To: <bug-335-13@https.bugs.libre-soc.org/>

https://bugs.libre-soc.org/show_bug.cgi?id=335

--- Comment #11 from Luke Kenneth Casson Leighton <lkcl@lkcl.net> ---
(In reply to Michael Nolan from comment #10)
> (In reply to Luke Kenneth Casson Leighton from comment #9)
> > hm i feel the need to assert that when the cases are not being enabled,
> > the registers not in use (not changed) should be Asserted 0
> > (self.o.{something}.ok == 0)
> > 
> > this because if they are not zero they will cause a request for a
> > register port write at the MultiCompUnit, which, because their data
> > (self.o.{something}.data) was not initialised, would cause regfile
> > corruption.
> 
> Good idea - done. Caught a bug where a regular branch would write to ctr if
> the branch immediate was just right (or wrong).

these are not fully showing up because we do not have writes to regfile(s) yet.
 if we put together the regfiles and a quick link to the pipelines through the
CompUnits interface, we pretty much have a working CPU done and can catch these
things, because SPRs,  XER, etc will all be spuriously modified.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev

^ permalink raw reply

* [libre-riscv-dev] [Bug 306] Formal Correctness Proof for ALU pipeline
From: bugzilla-daemon @ 2020-05-25  0:42 UTC (permalink / raw)
  To: libre-riscv-dev
In-Reply-To: <bug-306-13@https.bugs.libre-soc.org/>

https://bugs.libre-soc.org/show_bug.cgi?id=306

--- Comment #4 from Luke Kenneth Casson Leighton <lkcl@lkcl.net> ---
hmm

the same thing we noted in Branch, for CMPEQB is also not being picked up n the
main_stage.py

https://libre-soc.org/openpower/isa/comparefixed/

cmpeqb is only supposed to alter the CR, not the main register RT.

the formal proof is not picking that up.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev

^ permalink raw reply

* [libre-riscv-dev] [Bug 336] ALU CompUnit needs to recognise that RA (src1) can be zero
From: bugzilla-daemon @ 2020-05-25  0:56 UTC (permalink / raw)
  To: libre-riscv-dev
In-Reply-To: <bug-336-13@https.bugs.libre-soc.org/>

https://bugs.libre-soc.org/show_bug.cgi?id=336

--- Comment #21 from Cesar Strauss <cestrauss@gmail.com> ---
(In reply to Luke Kenneth Casson Leighton from comment #1)
> * third - and this is the kicker - that the FSM for src1_i needs a Mux to
> select RA or zero *and* it needs to *no longer* request a READ for src1
> (rd.req[0]) if the ra_is_zero flag is set.
> 
> * additionally that if the z flag is set, the rd_done part of the FSM does
> NOT wait for GO_READ on the src1 port (rd.go[0])

Just a reminder that the FSM still needs adjustment for rd.rel/rd.go. Checked
in GTKWave that rd.rel is still being set unconditionally after issue_i.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev

^ permalink raw reply

* [libre-riscv-dev] [Bug 336] ALU CompUnit needs to recognise that RA (src1) can be zero
From: bugzilla-daemon @ 2020-05-25  1:19 UTC (permalink / raw)
  To: libre-riscv-dev
In-Reply-To: <bug-336-13@https.bugs.libre-soc.org/>

https://bugs.libre-soc.org/show_bug.cgi?id=336

--- Comment #22 from Luke Kenneth Casson Leighton <lkcl@lkcl.net> ---
(In reply to Cesar Strauss from comment #21)

> Just a reminder that the FSM still needs adjustment for rd.rel/rd.go.
> Checked in GTKWave that rd.rel is still being set unconditionally after
> issue_i.

if rd.rel[0] is being set when zero_a is on, yes that's bad.

rd.go[N] is an external response to rd.req[N].  this is why it is critical to
get rd.req and wr.req bits set right because the external user of the CompUnits
kniws nothing about *why* they are set, they just respond.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev

^ permalink raw reply

* [libre-riscv-dev] [Bug 336] ALU CompUnit needs to recognise that RA (src1) can be zero
From: bugzilla-daemon @ 2020-05-25  1:27 UTC (permalink / raw)
  To: libre-riscv-dev
In-Reply-To: <bug-336-13@https.bugs.libre-soc.org/>

https://bugs.libre-soc.org/show_bug.cgi?id=336

--- Comment #23 from Luke Kenneth Casson Leighton <lkcl@lkcl.net> ---

 255         # all request signals gated by busy_o.  prevents picker problems
 256         m.d.comb += self.busy_o.eq(opc_l.q) # busy out
 257         bro = Repl(self.busy_o, self.n_src)
 258         m.d.comb += self.rd.rel.eq(src_l.q & bro) # src1/src2 req rel
 259 

ah.

right.

try this, at line 258

sls = []
for l in sl:
   sls.append(l[2]) # get latch conditions
self.rd.rel.eq(Cat(*sls) & bro)

that is bringing in the immediate selection.  i think.  from the muxes.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev

^ permalink raw reply

* [libre-riscv-dev] [Bug 336] ALU CompUnit needs to recognise that RA (src1) can be zero
From: bugzilla-daemon @ 2020-05-25  1:31 UTC (permalink / raw)
  To: libre-riscv-dev
In-Reply-To: <bug-336-13@https.bugs.libre-soc.org/>

https://bugs.libre-soc.org/show_bug.cgi?id=336

--- Comment #24 from Luke Kenneth Casson Leighton <lkcl@lkcl.net> ---
m.d.comb += src_sel.eq(Mux(op_is_imm, self.opc_l.q, self.src_l.q[i]))
sl[i][2] = src_sel

hmm you see how that interacts, from the src_srl?

-- 
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev

^ permalink raw reply

* [libre-riscv-dev] [Bug 336] ALU CompUnit needs to recognise that RA (src1) can be zero
From: bugzilla-daemon @ 2020-05-25  2:21 UTC (permalink / raw)
  To: libre-riscv-dev
In-Reply-To: <bug-336-13@https.bugs.libre-soc.org/>

https://bugs.libre-soc.org/show_bug.cgi?id=336

--- Comment #25 from Luke Kenneth Casson Leighton <lkcl@lkcl.net> ---
(In reply to Luke Kenneth Casson Leighton from comment #24)
> m.d.comb += src_sel.eq(Mux(op_is_imm, self.opc_l.q, self.src_l.q[i]))
> sl[i][2] = src_sel
> 
> hmm you see how that interacts, from the src_srl?

not quite correct, did something different, do git pull see if you can
get the test to work.  will take a look in morning.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev

^ permalink raw reply

* [libre-riscv-dev] [Bug 336] ALU CompUnit needs to recognise that RA (src1) can be zero
From: bugzilla-daemon @ 2020-05-25  2:29 UTC (permalink / raw)
  To: libre-riscv-dev
In-Reply-To: <bug-336-13@https.bugs.libre-soc.org/>

https://bugs.libre-soc.org/show_bug.cgi?id=336

--- Comment #26 from Luke Kenneth Casson Leighton <lkcl@lkcl.net> ---
interesting!

    # XXX - immediate and zero is not a POWER mode (and won't work anyway)
    # reason: no actual operands.
    #result = yield from op_sim(dut, 5, 2, InternalOp.OP_ADD, zero_a=1,
    #                                imm=8, imm_ok=1)
    #assert result == 8

this test is invalid, because it effectively is a "update register with
immediate", disabling both A and B regs.

i believe there are other opcodes to cover this, therefore it is not a
valid test.   which is good because the FSM doesn't cope with it
(it relies on at least one RD.GO coming in, and if both operands are
immediates, that can never happen)

-- 
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev

^ permalink raw reply

* [libre-riscv-dev] [Bug 325] create POWER9 TRAP pipeline
From: bugzilla-daemon @ 2020-05-25  2:41 UTC (permalink / raw)
  To: libre-riscv-dev
In-Reply-To: <bug-325-13@https.bugs.libre-soc.org/>

https://bugs.libre-soc.org/show_bug.cgi?id=325

--- Comment #11 from Luke Kenneth Casson Leighton <lkcl@lkcl.net> ---
and probably addpcis:

D <- d0||d1||d2
RT <- NIA + EXTS(D || 16 0)

-- 
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev

^ permalink raw reply

* [libre-riscv-dev] [Bug 350] LDSTCompUnit also needs to support zeroing on RA
From: bugzilla-daemon @ 2020-05-25  4:03 UTC (permalink / raw)
  To: libre-riscv-dev
In-Reply-To: <bug-350-13@https.bugs.libre-soc.org/>

https://bugs.libre-soc.org/show_bug.cgi?id=350

--- Comment #1 from Luke Kenneth Casson Leighton <lkcl@lkcl.net> ---
added, untested, however may have a block when both zero and immediate set.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev

^ permalink raw reply

* [libre-riscv-dev] [Bug 336] ALU CompUnit needs to recognise that RA (src1) can be zero
From: bugzilla-daemon @ 2020-05-25  4:05 UTC (permalink / raw)
  To: libre-riscv-dev
In-Reply-To: <bug-336-13@https.bugs.libre-soc.org/>

https://bugs.libre-soc.org/show_bug.cgi?id=336

--- Comment #27 from Luke Kenneth Casson Leighton <lkcl@lkcl.net> ---
argh yes it is a valid test.  there is an add zero immediate instruction.

so this means rd_done will never be set, and the FSM never proceeds to
completion on 2-operand CompUnits

-- 
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev

^ permalink raw reply

* [libre-riscv-dev] [Bug 336] ALU CompUnit needs to recognise that RA (src1) can be zero
From: bugzilla-daemon @ 2020-05-25  7:28 UTC (permalink / raw)
  To: libre-riscv-dev
In-Reply-To: <bug-336-13@https.bugs.libre-soc.org/>

https://bugs.libre-soc.org/show_bug.cgi?id=336

--- Comment #28 from Cesar Strauss <cestrauss@gmail.com> ---
Don't lose sleep about it, I can handle the FSM.(In reply to Luke Kenneth
Casson Leighton from comment #25)
> not quite correct, did something different, do git pull see if you can
> get the test to work.  will take a look in morning.

Don't lose sleep about it, I can handle the FSM. It's my bread and butter.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev

^ permalink raw reply

* [libre-riscv-dev] [Bug 336] ALU CompUnit needs to recognise that RA (src1) can be zero
From: bugzilla-daemon @ 2020-05-25  9:42 UTC (permalink / raw)
  To: libre-riscv-dev
In-Reply-To: <bug-336-13@https.bugs.libre-soc.org/>

https://bugs.libre-soc.org/show_bug.cgi?id=336

--- Comment #29 from Cesar Strauss <cestrauss@gmail.com> ---
(In reply to Luke Kenneth Casson Leighton from comment #22)
> (In reply to Cesar Strauss from comment #21)
> 
> > Just a reminder that the FSM still needs adjustment for rd.rel/rd.go.
> > Checked in GTKWave that rd.rel is still being set unconditionally after
> > issue_i.
> 
> if rd.rel[0] is being set when zero_a is on, yes that's bad.
> 
> rd.go[N] is an external response to rd.req[N].  this is why it is critical
> to get rd.req and wr.req bits set right because the external user of the
> CompUnits kniws nothing about *why* they are set, they just respond.

In this case, "external user" are the dependency matrices, right?

I've read https://libre-soc.org/3d_gpu/architecture/6600scoreboard/.

As far as I understand it, the Dependency Matrices also need to know themselves
that the operand is immediate, and not set the corresponding FU-Reg Latch,
correct? Is this already implemented?

-- 
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev

^ permalink raw reply

* [libre-riscv-dev] [Bug 216] LOAD STORE buffer needed
From: bugzilla-daemon @ 2020-05-25 10:36 UTC (permalink / raw)
  To: libre-riscv-dev
In-Reply-To: <bug-216-13@https.bugs.libre-soc.org/>

https://bugs.libre-soc.org/show_bug.cgi?id=216

--- Comment #43 from Luke Kenneth Casson Leighton <lkcl@lkcl.net> ---
Created attachment 63
  --> https://bugs.libre-soc.org/attachment.cgi?id=63&action=edit
over-run 80 chars

tobias, hi, code-review:

* please keep to under 80 chars.  the best way to do this is to make sure
  that the editor window is only 80 chars wide.

* please do not make multi-purpose commits.  the commit contains about a
  hundred lines of undocumented whitespace changes... *AND* contains
  code changes

* range(0, array_size) does not need the 0,.  just range(array_size)

* the 2D array can actually be done as a 1D array of Signals of length N

        for i in range(0, array_size):
            ul2 = []
            for j in range(0, array_size):
                ul2.append(Signal())

   can just be:

        for i in range(array_size):
            ul2.append(Signal())

* Signals all need to be reset_less=True (with very few exceptions)

* Signals in loops do not end up being named properly ($1, $2) which
  gets really annoying.

  add a parameter name="addr_match_%d" % j

* the traditional industry-standard short-name in hardware for "byte enable"
  is just "en".

        layout = (('data', 128),
                  ('en', 16)
                  )

-- 
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev

^ permalink raw reply

* [libre-riscv-dev] [Bug 336] ALU CompUnit needs to recognise that RA (src1) can be zero
From: bugzilla-daemon @ 2020-05-25 10:43 UTC (permalink / raw)
  To: libre-riscv-dev
In-Reply-To: <bug-336-13@https.bugs.libre-soc.org/>

https://bugs.libre-soc.org/show_bug.cgi?id=336

--- Comment #30 from Luke Kenneth Casson Leighton <lkcl@lkcl.net> ---
(In reply to Cesar Strauss from comment #28)
> Don't lose sleep about it, I can handle the FSM.(In reply to Luke Kenneth
> Casson Leighton from comment #25)
> > not quite correct, did something different, do git pull see if you can
> > get the test to work.  will take a look in morning.
> 
> Don't lose sleep about it, I can handle the FSM. It's my bread and butter.

:)

i realised that actually there's probably nothing wrong with the modification
i made: it's the unit test that must not check a flag (rd.req) that is never
going to get set.


(In reply to Cesar Strauss from comment #29)
> (In reply to Luke Kenneth Casson Leighton from comment #22)
> > if rd.rel[0] is being set when zero_a is on, yes that's bad.
> > 
> > rd.go[N] is an external response to rd.req[N].  this is why it is critical
> > to get rd.req and wr.req bits set right because the external user of the
> > CompUnits kniws nothing about *why* they are set, they just respond.
> 
> In this case, "external user" are the dependency matrices, right?

yes.

> I've read https://libre-soc.org/3d_gpu/architecture/6600scoreboard/.
> 
> As far as I understand it, the Dependency Matrices also need to know
> themselves that the operand is immediate, 

not quite.  immediates are immediate, therefore there is no concept of a
"dependency".  

therefore, technically you are correct, but it is by way of "total omission
of involvement".

> and not set the corresponding FU-Reg Latch, correct?

by way of the DMs not in the slightest remote bit knowing *anything* about
immediates.... yes.

> Is this already implemented?

yes.  the DMs manage "dependencies".  an immediate is not a Dependency
Hazard (not a read or write register).

-- 
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev

^ permalink raw reply

* [libre-riscv-dev] [Bug 350] LDSTCompUnit also needs to support zeroing on RA
From: bugzilla-daemon @ 2020-05-25 11:46 UTC (permalink / raw)
  To: libre-riscv-dev
In-Reply-To: <bug-350-13@https.bugs.libre-soc.org/>

https://bugs.libre-soc.org/show_bug.cgi?id=350

--- Comment #2 from Luke Kenneth Casson Leighton <lkcl@lkcl.net> ---
(In reply to Luke Kenneth Casson Leighton from comment #1)
> added, untested, however may have a block when both zero and immediate set.

it's fine.  unit test modified as follows:

+    rd = 0b00
+    if not imm_ok:
+        rd |= 0b10
+    if not zero_a:
+        rd |= 0b01
+
+    if rd:
+        yield dut.rd.go.eq(rd)
+        yield from wait_for(dut.rd.rel)
+        yield dut.rd.go.eq(0)

-- 
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
libre-riscv-dev mailing list
libre-riscv-dev@lists.libre-riscv.org
http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev

^ permalink raw reply

page:   latest

Libre RISC-V Mailing list Archives (accessible via public-inbox)

Archives are clonable: git clone --mirror https://inbox.libre-soc.org/libre-riscv-dev

Example config snippet for mirrors


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git