Libre RISC-V Mailing list Archives (accessible via public-inbox)
 help / color / Atom feed
* Re: [libre-riscv-dev] daily kan-ban update 25may2020
From: Luke Kenneth Casson Leighton @ 2020-05-25 22:45 UTC (permalink / raw)
  To: Libre-RISCV General Development
In-Reply-To: <EF27EE71-46C9-426B-93EC-A7DDC61E804E@getmailspring.com>

On Monday, May 25, 2020, Cole Poirier <colepoirier@gmail.com> wrote:

> On May 25 2020, at 2:35 pm, Luke Kenneth Casson Leighton <lkcl@lkcl.net>
> wrote:
> > in your wiki home page you will find a link to bugs assigned to you.  the
> > default action is basically "monitor and do that list".
>
> Thanks that makes it easy. I saw you assigned me 'Bug 337 - Convention
> for register outputs in *OutputData structures is to use "Data"', and
> looked at your comment. Is this simply reorg or is there something
> additional I should be aware of?


ah i think i kinda did that one yesterday.


>
> > one to add:
> >
> > if you can also find the "interfaces" issue, and assign it to
> > yourself, you
> > will see a list of interfaces in the last comment, including "serialized
> > irq" which you shouls find somewhere from raptor engineering.
> >
> > can you track the source code down, add them to the wiki  under the right
> > subdurectory (shakti mclass) and also crossreference them in the
> bugtracker?
>
>
> Are you referring to "Bug 304 - Define minimum viable interface set for
> 180nm ASIC", comment #6?


yes.


> ```
> Luke Kenneth Casson Leighton 2020-05-20 23:14:26 BST
>
> conversation notes from Tim (Raptor Engineering)
>
> priority:
>
> * 2x SPI master
> * 4x I2C master
> * 2x UART
> * 4x LPC master
> * EINT* Serialsed IRQ (LPC, from PCIe) - 15 or so IRQs, these are
> PC>PCIe>OpenPower mappings
>
> Quad SPI raptorengineering
> LPC bridge github raptorengineering
> ```
>
> I searched for "raptor serialized irq, and only found references to
> raptors lpc-spi-bridge-fpga gitlab repo. Is this what we need
> https://gitlab.raptorengineering.com/raptor-engineering-public/lpc-spi-
> bridge-fpga?


that sort of thing.

we need to find the full implementation, an actual serialized-irq master



> An in terms of adding the source code to the wiki,


> do you want me to
> just copy and paste the 1000+ line verilog file or files to the
> soc/libreriscv/shakti/m_class directory?


no!  a wiki is not a source code repository, it is for documentation



> Or do you want me to do
> something similar to what you did for sdram.mdwn in the same directory?


yes.


>
> ```
> # SDRAM
>
> * <https://bitbucket.org/casl/c-class/src/3fba75dfbd0c64815eb8ec6dc96566
> 6812c44bae/src/peripherals/sdram/?at=master>
> * <https://opencores.org/projects/sdr_ctrl>
> ```
>
> Cole
>
> _______________________________________________
> libre-riscv-dev mailing list
> libre-riscv-dev@lists.libre-riscv.org
> http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev
>


-- 
---
crowd-funded eco-conscious hardware: https://www.crowdsupply.com/eoma68
_______________________________________________
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

* Re: [libre-riscv-dev] daily kan-ban update 25may2020
From: Cole Poirier @ 2020-05-25 22:11 UTC (permalink / raw)
  To: Libre-RISCV General Development; +Cc: Libre-RISCV General Development
In-Reply-To: <CAPweEDy4iGcxj9J3p9MdJDhMku6_f6Xj4zMusU-AxHWa+_G8Hg@mail.gmail.com>

On May 25 2020, at 2:35 pm, Luke Kenneth Casson Leighton <lkcl@lkcl.net> wrote:
> in your wiki home page you will find a link to bugs assigned to you.  the
> default action is basically "monitor and do that list".

Thanks that makes it easy. I saw you assigned me 'Bug 337 - Convention
for register outputs in *OutputData structures is to use "Data"', and
looked at your comment. Is this simply reorg or is there something
additional I should be aware of?
 
> one to add:
> 
> if you can also find the "interfaces" issue, and assign it to
> yourself, you
> will see a list of interfaces in the last comment, including "serialized
> irq" which you shouls find somewhere from raptor engineering.
> 
> can you track the source code down, add them to the wiki  under the right
> subdurectory (shakti mclass) and also crossreference them in the bugtracker?


Are you referring to "Bug 304 - Define minimum viable interface set for
180nm ASIC", comment #6?
```
Luke Kenneth Casson Leighton 2020-05-20 23:14:26 BST

conversation notes from Tim (Raptor Engineering)

priority:

* 2x SPI master
* 4x I2C master
* 2x UART
* 4x LPC master
* EINT* Serialsed IRQ (LPC, from PCIe) - 15 or so IRQs, these are
PC>PCIe>OpenPower mappings

Quad SPI raptorengineering
LPC bridge github raptorengineering
```

I searched for "raptor serialized irq, and only found references to
raptors lpc-spi-bridge-fpga gitlab repo. Is this what we need
https://gitlab.raptorengineering.com/raptor-engineering-public/lpc-spi-bridge-fpga?
An in terms of adding the source code to the wiki, do you want me to
just copy and paste the 1000+ line verilog file or files to the
soc/libreriscv/shakti/m_class directory? Or do you want me to do
something similar to what you did for sdram.mdwn in the same directory?

```
# SDRAM

* <https://bitbucket.org/casl/c-class/src/3fba75dfbd0c64815eb8ec6dc965666812c44bae/src/peripherals/sdram/?at=master>
* <https://opencores.org/projects/sdr_ctrl>
```

Cole

_______________________________________________
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 316] bperm TODO
From: bugzilla-daemon @ 2020-05-25 21:36 UTC (permalink / raw)
  To: libre-riscv-dev
In-Reply-To: <bug-316-13@https.bugs.libre-soc.org/>

https://bugs.libre-soc.org/show_bug.cgi?id=316

--- Comment #100 from Luke Kenneth Casson Leighton <lkcl@lkcl.net> ---
(In reply to Cole Poirier from comment #99)
> Awesome, thank you both for your help, I clearly wasn't going to figure this
> out. What should I do now? Should I try reading and understanding Claire
> Wolf's optimized bit manip code, then try optimizing the bpermd module?

we have a working module so leave optimisation for later.

leave thus bugreport open, come back to that later

> Or
> is there something more helpful I should do?

see mailing list.

-- 
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

* Re: [libre-riscv-dev] daily kan-ban update 25may2020
From: Luke Kenneth Casson Leighton @ 2020-05-25 21:35 UTC (permalink / raw)
  To: Libre-RISCV General Development
In-Reply-To: <38DF1D4D-C203-432F-9545-341DEB3E1457@getmailspring.com>

On Monday, May 25, 2020, Cole Poirier <colepoirier@gmail.com> wrote:

> Took yesterday off. Thanks for your help Michael and Luke. Today I took
> a look at what Michael and Luke did to fix Bpermd so it passed the
> fu/logical/test_pipe_caller.py tests, it seems like there was a previous
> module in that test messing with the value of register a which caused my
> module to erroneously fail when I tried adding the modifications to
> bpermd.py that are now there from Michael.


sigh that was an annoying weird bug.

 i think 316 can be closed now

What should I do next? Should
> I try optimizing the gate count of bpermd, or is there something more
> helpful I can work on today?


in your wiki home page you will find a link to bugs assigned to you.  the
default action is basically "monitor and do that list".

one to add:

if you can also find the "interfaces" issue, and assign it to yourself, you
will see a list of interfaces in the last comment, including "serialized
irq" which you shouls find somewhere from raptor engineering.

can you track the source code down, add them to the wiki  under the right
subdurectory (shakti mclass) and also crossreference them in the bugtracker?

l.



-- 
---
crowd-funded eco-conscious hardware: https://www.crowdsupply.com/eoma68
_______________________________________________
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

* Re: [libre-riscv-dev] daily kan-ban update 25may2020
From: Cole Poirier @ 2020-05-25 21:06 UTC (permalink / raw)
  To: Libre-RISCV General Development; +Cc: Libre-RISCV General Development
In-Reply-To: <CAPweEDxS7N-V9zUuqefV+faSTuXCAFsgHNY2FErbdUO1UH04LA@mail.gmail.com>

Took yesterday off. Thanks for your help Michael and Luke. Today I took
a look at what Michael and Luke did to fix Bpermd so it passed the
fu/logical/test_pipe_caller.py tests, it seems like there was a previous
module in that test messing with the value of register a which caused my
module to erroneously fail when I tried adding the modifications to
bpermd.py that are now there from Michael. What should I do next? Should
I try optimizing the gate count of bpermd, or is there something more
helpful I can work on today?



_______________________________________________
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 316] bperm TODO
From: bugzilla-daemon @ 2020-05-25 21:03 UTC (permalink / raw)
  To: libre-riscv-dev
In-Reply-To: <bug-316-13@https.bugs.libre-soc.org/>

https://bugs.libre-soc.org/show_bug.cgi?id=316

--- Comment #99 from Cole Poirier <colepoirier@gmail.com> ---
Awesome, thank you both for your help, I clearly wasn't going to figure this
out. What should I do now? Should I try reading and understanding Claire Wolf's
optimized bit manip code, then try optimizing the bpermd module? Or is there
something more helpful I should do?

-- 
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 342] formal proof of soc.fu.compunits.FunctionUnitBaseSingle needed
From: bugzilla-daemon @ 2020-05-25 20:56 UTC (permalink / raw)
  To: libre-riscv-dev
In-Reply-To: <bug-342-13@https.bugs.libre-soc.org/>

https://bugs.libre-soc.org/show_bug.cgi?id=342

--- Comment #6 from Luke Kenneth Casson Leighton <lkcl@lkcl.net> ---
another one, a part of the API: issue_i is only asserted for one cycle.

however as an external signal i have no idea how that woule be expressed. 
Assume?

-- 
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

* Re: [libre-riscv-dev] daily kan-ban update 25may2020
From: Luke Kenneth Casson Leighton @ 2020-05-25 19:40 UTC (permalink / raw)
  To: Libre-RISCV General Development
In-Reply-To: <CAC2bXD68oyveF2S0thnAP+uodD7BchMLirEC7A15RbTKd25yvg@mail.gmail.com>

On Mon, May 25, 2020 at 8:30 PM Jacob Lifshay <programmerjake@gmail.com> wrote:
>
> On Mon, May 25, 2020, 05:31 Luke Kenneth Casson Leighton <lkcl@lkcl.net>
> wrote:
>
> > * (Jacob) DIV and MUL pipeline stages in ieee754fpu, for integration
> >    into corresponding soc.fu.*.main_stage
>
>
> Will be working on the mul and div pipelines today,

superb.

> created an enum of all the ops supported by the div pipeline on friday.

yeh saw that, especially the notes about how POWER complies with c semantics.

l.

_______________________________________________
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

* Re: [libre-riscv-dev] daily kan-ban update 25may2020
From: Jacob Lifshay @ 2020-05-25 19:30 UTC (permalink / raw)
  To: Libre-RISCV General Development
In-Reply-To: <CAPweEDxZF4PK7-dOV9H2id2BBauY8oYzu_YSaH68ZxL4Ov5NRQ@mail.gmail.com>

On Mon, May 25, 2020, 05:31 Luke Kenneth Casson Leighton <lkcl@lkcl.net>
wrote:

> * (Jacob) DIV and MUL pipeline stages in ieee754fpu, for integration
>    into corresponding soc.fu.*.main_stage


Will be working on the mul and div pipelines today, created an enum of all
the ops supported by the div pipeline on friday.

Jacob
_______________________________________________
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 342] formal proof of soc.fu.compunits.FunctionUnitBaseSingle needed
From: bugzilla-daemon @ 2020-05-25 19:09 UTC (permalink / raw)
  To: libre-riscv-dev
In-Reply-To: <bug-342-13@https.bugs.libre-soc.org/>

https://bugs.libre-soc.org/show_bug.cgi?id=342

--- Comment #5 from Luke Kenneth Casson Leighton <lkcl@lkcl.net> ---
extra little funny things that can be thrown in:

* all of rd.go, rd.req, wr.go, wr.req, and issue: if done_o is LOW then
  absolutely all of those should be LOW as well.

  the reason for the existence of done_o over busy_o is because, sigh,
  if you look closely, you can see that some of the latches are reset
  synchronously (a 1 clock delay).  this means that busy_o actually
  drops immediately on completion, and done_o drops a cycle later.

* when the write phase has started, the rd phase is *definitely* over.
  therefore, if wr.req is non-zero, rd.req and rd.go should *definitely*
  be non-zero.

-- 
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 342] formal proof of soc.fu.compunits.FunctionUnitBaseSingle needed
From: bugzilla-daemon @ 2020-05-25 19:00 UTC (permalink / raw)
  To: libre-riscv-dev
In-Reply-To: <bug-342-13@https.bugs.libre-soc.org/>

https://bugs.libre-soc.org/show_bug.cgi?id=342

--- Comment #4 from Luke Kenneth Casson Leighton <lkcl@lkcl.net> ---
michael am just going over the notes in proof_fu.py

            # Assume no instruction is issued until rd_rel is
            # released. Idk if this is valid

            with m.If(busy):
                comb += Assume(issue == 0)

ok so yes, nothing happens at the ALU - nothing is sent to the ALU - until
all incoming operands are valid, and the condition for that is:

* busy_o is set HI
* all bits of rd_rel are LOW


note btw the industry-standard convention in circuits, a "N" at the end
of a name indicates that it is inverted from what you would normally think
its function should be.  this means shadown is LOW if shadowing is requested,
and when shadown is HI, shadowing is *not* being enabled.

-- 
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

* Re: [libre-riscv-dev] daily kan-ban update 25may2020
From: Luke Kenneth Casson Leighton @ 2020-05-25 18:47 UTC (permalink / raw)
  To: Libre-RISCV General Development
In-Reply-To: <39e1b0f0-7a36-8d35-9e3b-30aefbaa645f@gmail.com>

On Mon, May 25, 2020 at 6:42 PM Cesar Strauss <cestrauss@gmail.com> wrote:

> The day before yesterday:
>
>  * Helped Luke add an optional zero_a immediate option to MultiCompUnit

i must apologise, it would seem that even just talking with you
accelerates what i am able to do.  i must try to divert my attention
away and allow you to get on with some things :)

> Yesterday:
>
>  * Continued examining the simulated output of compalu_multi.py on
> GTKWave to better understand the rel/go protocol and the overall
> behavior of MultiCompUnit. This will help with the unit tests.

the rel/go protocol is extremely simple (this for michael as well):

* when rel[x] is raised, go[x] must (not immediately, but at some
point, from 0 to N cycles in the future) be raised, and it must be
raised once and only once, for one clock cycle and one clock cycle
only.

* once go[x] is raised, rel[x] must de-assert exactly one clock cycle
from that point, and it must *remain* de-asserted until busy_o is
de-asserted.


>  * In the process, understood better how nMigen Signals and Records are
> shown (or not) in GTKWave.

sigh.  there is actually a bug in nmigen which stops top-level signals
from even being listed in gtkwave.  michael tracked down a workaround.

> Hints: (1) do not use the exact same "name"
> parameter for distinct signals. (2) Do use the "name" parameter on
> Records when necessary.
>
> This morning:
>
>  * Refined my knowledge of the parallel simulation capability of
> nMigen, taking a look at src/nmutil/test/test_buf_pipe.py.
>  * Implemented a proof of concept of a parallel test.

yes, i saw - i filled in some stub (TODO) comments.

> planned today:
>
>  * Use the parallel test capability of nMigen to simulate multiple
> concurrent ports of the Scoreboard, giving a more comprehensive coverage
> of MultiCompUnit.

excellent.  this test class actually becomes quite important as i
would like to use it as the basis for feeding data into the
soc.fu.compunits that you can see in that module.

the bit of "weirdness" about the parallel test capability of nMigen is:

1) it's not actually parallel: it's a co-iterative event-driven
simulation which steps through even "tick" (each yield) ensuring that
each and every one of them is kept in time-step depending on the clock
"pulse" that each "yielding" function returns.

2) it is slightly weird given that you can actually pass a parameter
(or parameters) to each of the functions, like this:

    test = CompUnitParallelTest(dut)
    run_simulation(dut, [test.driver(),
                         test.monitor(),
                         test.rd(0), # one read port (a)
                         test.rd(1), # one read port (b)
                         test.wr(0)  # one write port (o)
                        ],
                   vcd_name="test_compunit_parallel.vcd")

where test.rd() and test.wr() are defined like this:

class CompUnitParallelTest:
    def rd(self, rd_idx):
        # monitor self.dut.rd.req[rd_idx] and sets dut.rd.go[idx] for one cycle
        yield

so that parameter rd_idx is *DIFFERENT* for each "instance" of the
[not-]parallel tests, and we exploit this to get one function to
monitor dut.rd.req[0] and set dut.rd.go[0], another to look at
dut.rd.req[1] and dut.rd.go[1] and so on.

note, Cesar, i just put a quick extra bit into the TODO notes,

         # TODO: also when dut.rd.go is set, put the expected value into
-        # the src_i.
+        # the src_i.  use dut.get_in[rd_idx] to do so

so basically in this way, everything about the read(idx) function
involves data and signals that are indexed by "idx", and likewise for
the write(idx) function.

l.

_______________________________________________
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

* Re: [libre-riscv-dev] daily kan-ban update 25may2020
From: Luke Kenneth Casson Leighton @ 2020-05-25 18:29 UTC (permalink / raw)
  To: Libre-RISCV General Development
In-Reply-To: <ca9d1527-50f5-112b-8fe1-2be9be17e084@gmail.com>

---
crowd-funded eco-conscious hardware: https://www.crowdsupply.com/eoma68

On Mon, May 25, 2020 at 5:39 PM Michael Nolan <mtnolan2640@gmail.com> wrote:
>
> Today I'm working on understanding how the MultiCompUnit from
> compalu_multi.py works, and how to go about formally verifying it.

cool.  here's some notes:

1. i moved compunit_multi.py's docstring to here (it was a bit long):
https://libre-soc.org/3d_gpu/architecture/compunit/  there, you will
find a description of the crucial stages that a CompUnit goes through.

2. if you set n_src=1, and n_dst=1, and you pass it a CompNNNOpSubset
that has neither an imm_data nor a zero_a, you end up preeettty much
exactly with this:
https://libre-soc.org/3d_gpu/comp_unit_req_rel.jpg

3. if however you have n_src=3 and n_dst=2, and you pass in an
opsubset that has an imm_data, you end up with something like this
(ignore go_adr and go_st for now):
https://libre-soc.org/3d_gpu/ld_st_comp_unit.jpg

l.

_______________________________________________
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

* Re: [libre-riscv-dev] daily kan-ban update 25may2020
From: Cesar Strauss @ 2020-05-25 17:41 UTC (permalink / raw)
  To: libre-riscv-dev
In-Reply-To: <CAPweEDxZF4PK7-dOV9H2id2BBauY8oYzu_YSaH68ZxL4Ov5NRQ@mail.gmail.com>

The day before yesterday:

 * Helped Luke add an optional zero_a immediate option to MultiCompUnit

Yesterday:

 * Continued examining the simulated output of compalu_multi.py on
GTKWave to better understand the rel/go protocol and the overall
behavior of MultiCompUnit. This will help with the unit tests.
 * In the process, understood better how nMigen Signals and Records are
shown (or not) in GTKWave. Hints: (1) do not use the exact same "name"
parameter for distinct signals. (2) Do use the "name" parameter on
Records when necessary.

This morning:

 * Refined my knowledge of the parallel simulation capability of
nMigen, taking a look at src/nmutil/test/test_buf_pipe.py.
 * Implemented a proof of concept of a parallel test.

planned today:

 * Use the parallel test capability of nMigen to simulate multiple
concurrent ports of the Scoreboard, giving a more comprehensive coverage
of MultiCompUnit.

Regards,

Cesar

_______________________________________________
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

* Re: [libre-riscv-dev] daily kan-ban update 25may2020
From: Michael Nolan @ 2020-05-25 16:39 UTC (permalink / raw)
  To: libre-riscv-dev
In-Reply-To: <CAPweEDxZF4PK7-dOV9H2id2BBauY8oYzu_YSaH68ZxL4Ov5NRQ@mail.gmail.com>

Today I'm working on understanding how the MultiCompUnit from 
compalu_multi.py works, and how to go about formally verifying it.

--Michael


_______________________________________________
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 15:32 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 #33 from Luke Kenneth Casson Leighton <lkcl@lkcl.net> ---
(In reply to Luke Kenneth Casson Leighton from comment #32)
> (In reply to Cesar Strauss from comment #31)
> > Regarding the parallel test:
> > 
> > Pushed a proof of concept.
> > 
> > Defective as it was, I pushed it anyway, hoping for some feedback on the
> > bug. Found it myself literally seconds after pushing.
> 
> :)
> 
> ok i added some stubs here.

right.  so i've pushed another commit which shows where we would expect to
put the inputs in (constructor).  rather than have them directly as arguments
to op_sim(), all those arguments - a, b, op, inv_a, imm_ok, zero_a - they
become the arguments to the constructor.

also expected_o should be in the constructor as well.

-- 
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 15:25 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 #32 from Luke Kenneth Casson Leighton <lkcl@lkcl.net> ---
(In reply to Cesar Strauss from comment #31)
> Regarding the parallel test:
> 
> Pushed a proof of concept.
> 
> Defective as it was, I pushed it anyway, hoping for some feedback on the
> bug. Found it myself literally seconds after pushing.

:)

ok i added some stubs here.

-- 
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 15:14 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

--- Comment #31 from Cesar Strauss <cestrauss@gmail.com> ---
Regarding the parallel test:

Pushed a proof of concept.

Defective as it was, I pushed it anyway, hoping for some feedback on the bug.
Found it myself literally seconds after pushing.

-- 
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

* Re: [libre-riscv-dev] daily kan-ban update 25may2020
From: Luke Kenneth Casson Leighton @ 2020-05-25 14:27 UTC (permalink / raw)
  To: Libre-RISCV General Development
In-Reply-To: <20200525162057.c011c8aead9c705bf824db1e@platen-software.de>

On Monday, May 25, 2020, Tobias Platen <libre-soc@platen-software.de> wrote:

> yesterday:
>         begun work on DataMerger


looks good tobias.  i updated the docstring.  should actually be really
short, despite being strategically a critical component.

l.



-- 
---
crowd-funded eco-conscious hardware: https://www.crowdsupply.com/eoma68
_______________________________________________
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 352] New: virtual (dependency-tracked) regfile needed
From: bugzilla-daemon @ 2020-05-25 14:23 UTC (permalink / raw)
  To: libre-riscv-dev

https://bugs.libre-soc.org/show_bug.cgi?id=352

            Bug ID: 352
           Summary: virtual (dependency-tracked) regfile 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: ---

i was hoping this would not be needed at this stage, however it looks like it
is.

there are over 100 active SPRs in POWER9.   we cannot possibly have 
a Dependency Matrix 100-wide for that.

therefore, option 1:

* restrict the supported SPRs to a manageable subset (and remap them)

option 2:

* have a "virtual" (dependency-tracked) front-end on the Regfile.

traditional names for this are "PRF-ARF" - physical regfile / architectural
regfile.

normally, the numbering for regfiles would have fixed indexes.  here, instead,
we need *dynamically-allocated* re-mapped indexing, plus a tracking system
which will be based on a pair of "bitvectors", generated and managed by the
FU-REGs Dependency Matrix: global_read_pending and global_write_pending

on first allocation, any free bit in g_r and g_w (both must be free) is used
to indicate that that "slot" is available.  the association between the
"real" regfile (binary? unary?) index will be recorded and kept.

the association will only be dropped when both the g_r and g_w pending
bitvectors clear their bits corresponding to that PRF-ARF association.

if there are no free bits, the entire execution engine - further allocation
of instructions to Function Units - is REQUIRED to stall.

-- 
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

* Re: [libre-riscv-dev] daily kan-ban update 25may2020
From: Tobias Platen @ 2020-05-25 14:20 UTC (permalink / raw)
  To: libre-riscv-dev
In-Reply-To: <CAPweEDxZF4PK7-dOV9H2id2BBauY8oYzu_YSaH68ZxL4Ov5NRQ@mail.gmail.com>

yesterday:
	begun work on DataMerger
this morning:
	reading the whole bug report related to #216
planned this evening:
	more work on Datamerger

_______________________________________________
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 351] New: create a "block" (mass) regfile port (read and write) onto an array-based regfile
From: bugzilla-daemon @ 2020-05-25 14:07 UTC (permalink / raw)
  To: libre-riscv-dev

https://bugs.libre-soc.org/show_bug.cgi?id=351

            Bug ID: 351
           Summary: create a "block" (mass) regfile port (read and write)
                    onto an array-based regfile
           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: ---

see soc.regfile.RegFileArray

for the Condition Regfile and for the XER Regfile, a "mass-hit" read/write
port is required that looks like a standard read_port (or write_port) however
permits writing to the *entire* regfile, all at once.

this can be done as a virtual port, as follows:

* create 8x read (or write) ports
* join all 8 data together as if they were a single thing
* join all 8 "enable" lines together as if they were a single thing

it will however be critical to ensure that the alternative "view" is not
used at the same time.

-- 
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 13:55 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 #44 from Luke Kenneth Casson Leighton <lkcl@lkcl.net> ---
excellent, tobias.  i added some docstrings, please do git pull and carry on.
i just had to redo the docstring i just wrote after noticing (only now) several
days of using a MACos terminal at 86x30 not 80x30.  sigh :)

anyway: the docstring should now show clearly what's needed.

commit db13193e20441c6988e7a1b3969979461136eca3
Author: Tobias Platen <tplaten@posteo.de>
Date:   Mon May 25 15:41:49 2020 +0200

    refactoring (see #216 Comment 43)

-- 
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] daily kan-ban update 25may2020
From: Luke Kenneth Casson Leighton @ 2020-05-25 12:30 UTC (permalink / raw)
  To: Libre-RISCV General Development

yesterday:

* split out Popcount into separate module, used by Logical pipeline
* added MFMSR and MTMSR to the CSV files (they were missing)
* created a new SPR pipeline bugreport
* completed a rather large update to make every pipeline use "Data" for
  for output, because of the "ok" signalling it provides.
* documented which regfiles are needed.
* worked out a test plan for how to test the pipelines by connecting
   them to regfiles in a very, very basic way.

this morning:

* reviewed DataMerger beginnings (Tobias: see #216)
* added zero_a support to LDSTCompUnit (and a preliminary test)

planned today:

* define some regfiles and their ports
* work out how to do the link and test of pipelines to regfiles

blocked:

* (Cesar) the multi-process unit test for MultiCompUnit is critical to
   proceeding with a test of the pipelines when they are accessed
   through MultiCompUnit.  the exact same functions are needed.
* (Jacob) DIV and MUL pipeline stages in ieee754fpu, for integration
   into corresponding soc.fu.*.main_stage

_______________________________________________
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 11:57 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
----------------------------------------------------------------------------
           Assignee|lkcl@lkcl.net               |cestrauss@gmail.com

--- Comment #3 from Luke Kenneth Casson Leighton <lkcl@lkcl.net> ---
cesar this one likewise needs similar tests to the MultiCompUnit, might even
be able to share the same base classes at some point.

again it needs that "multi-process" simulation model:

* one function doing setup, setting oper_i, issue_i, then monitoring busy_o
  before exiting

* another that checks busy_o and, when set, proceeds to monitor a single
  bit of rd.req, sets the corresponding bit of rd.go for one cycle, then
  monitors busy_o before exiting

* likewise for write requests.

this is exactly the same as needed for MultiCompUnit (up to this point).
the additions obviously are to have extra functions that:

* monitor req_adr_o and set go_adr_i
* monitor req_st_o and set go_st_i
* check the address and the data afterwards

-- 
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