* Re: [libre-riscv-dev] daily kan-ban update 06may2020
From: Michael Nolan @ 2020-05-06 16:23 UTC (permalink / raw)
To: libre-riscv-dev
In-Reply-To: <CAPweEDyRnFhaT5tADj6ka+=Qo1RKUZjeafO9x8iFdGvXavy5yA@mail.gmail.com>
Working on branches in the power_pseudo simulator. I now have branch and
link working, move to spr, and branch to lr and ctr working. I did run
into an issue here though, in that the generated python is slightly
broken in a way I'm not entirely sure I can fix inside the code
generator. I added a way to automatically patch the generated python
files to work around this.
The issue is in the following line:
ctr_ok = BO[2] | (CTR[M:64] != 0) ^ BO[3]
It xors the result of comparing CTR to 0 (which is a bool) with BO[3]
(Which is a selectableInt)
I'm not sure how to fix this aside from having the code generator
surround every comparison expression with SelectableInt(x, bits=1)
--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
* Re: [libre-riscv-dev] daily kan-ban update 06may2020
From: Luke Kenneth Casson Leighton @ 2020-05-06 16:30 UTC (permalink / raw)
To: Libre-RISCV General Development
In-Reply-To: <d3063996-e91e-9f30-ce3a-496c5e8a27ae@gmail.com>
On Wed, May 6, 2020 at 5:23 PM Michael Nolan <mtnolan2640@gmail.com> wrote:
> I'm not sure how to fix this aside from having the code generator
> surround every comparison expression with SelectableInt(x, bits=1)
it'll look awful however that's a perfectly reasonable solution - go
for it. i'm amazed the auto-generated output is as readable as it is.
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 06may2020
From: Luke Kenneth Casson Leighton @ 2020-05-06 16:33 UTC (permalink / raw)
To: Libre-RISCV General Development
In-Reply-To: <CAPweEDw2fiu0+kZBFEyecdh0Ub_2H_tOyWbe1CzOB84_n=sLOQ@mail.gmail.com>
btw at some point, michael, we need to stop using the "standard"
python operators and use explicit functions.
oh. this is another way to fix it. add functions eq, gt, ge to
globals, and get the codegenerato to do this instead:
BO[2] | (eq(CTR[M:64],0)) ^ BO[3]
then, eq() can return a SelectableInt(x, bits=1) without having that
duplicated everywhere.
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 06may2020
From: Jacob Lifshay @ 2020-05-07 5:52 UTC (permalink / raw)
To: Libre-RISCV General Development
In-Reply-To: <CAPweEDyRnFhaT5tADj6ka+=Qo1RKUZjeafO9x8iFdGvXavy5yA@mail.gmail.com>
Still figuring out load/store code...
found an interesting article in this week's This Week in Rust about forking
a process onto a remote computer:
https://thume.ca/2020/04/18/telefork-forking-a-process-onto-a-different-computer/
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 300] Documentation for the SOC
From: bugzilla-daemon @ 2020-05-07 6:29 UTC (permalink / raw)
To: libre-riscv-dev
In-Reply-To: <bug-300-13@https.bugs.libre-soc.org/>
https://bugs.libre-soc.org/show_bug.cgi?id=300
Jacob Lifshay <programmerjake@gmail.com> changed:
What |Removed |Added
----------------------------------------------------------------------------
Resolution|WONTFIX |---
Status|RESOLVED |DEFERRED
--- Comment #2 from Jacob Lifshay <programmerjake@gmail.com> ---
deferred till after oct tapeout
--
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 06may2020
From: Luke Kenneth Casson Leighton @ 2020-05-07 10:17 UTC (permalink / raw)
To: Libre-RISCV General Development
In-Reply-To: <CAC2bXD5kmOFA=XvWyGfN4p3_WoCQEXQGRv1nhsndnwG-ZjN0VA@mail.gmail.com>
On Thu, May 7, 2020 at 6:53 AM Jacob Lifshay <programmerjake@gmail.com>
wrote:
> Still figuring out load/store code...
>
remember you don't have to do that "alone". ask questions, even if it's
"what does this class do". also that if one approach does not work (it's
been around 4 weeks now) perhaps a different approach?
>
> found an interesting article in this week's This Week in Rust about forking
> a process onto a remote computer:
>
> https://thume.ca/2020/04/18/telefork-forking-a-process-onto-a-different-computer/
ah bless 'im. he's re-inventing DCE/RPC which solved this 25+ years ago
:) i am constantly amazed by how completely unaware of fundamental
technology that already exists.
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
* [libre-riscv-dev] daily kan-ban 07may2020 update
From: Luke Kenneth Casson Leighton @ 2020-05-07 10:26 UTC (permalink / raw)
To: Libre-RISCV General Development
continuing the daily updates, so everyone knows what everyone is doing.
i worked through the new LD/ST CompUnit, got a first draft which has no
python syntax errors and generates an HDL file, today will be to get its
unit test functional (using the new PortInterface). if that works quickly
i will move on to integrating it into the score6600 code.
MarketNext are going ahead quickly with a magazine article, i will talk to
them about that today.
we had the bi-weekly call with OpenPOWER Foundation member, Tim Pearson
(Raptor CS), and he pointed us at the LPC Master code (just released) - if
we include this interface in the 180nm ASIC they may end up using it as a
"Boot Management processor" to manage larger POWER9 systems.
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
* [libre-riscv-dev] [Bug 303] New: define peripheral set for 180nm ASIC
From: bugzilla-daemon @ 2020-05-07 12:44 UTC (permalink / raw)
To: libre-riscv-dev
https://bugs.libre-soc.org/show_bug.cgi?id=303
Bug ID: 303
Summary: define peripheral set for 180nm ASIC
Product: Libre-SOC's first SoC
Version: unspecified
Hardware: PC
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: ---
we need to define a peripheral set for the ASIC, create a page
(and use the peripheral / pinmux generator to describe it)
--
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] 4 Simulators?
From: Yehowshua @ 2020-05-07 17:33 UTC (permalink / raw)
To: Libre-RISCV General Development
Hello all - I recently became aware that there a 4 different POWER simulators being worked on?
This is not a good use of resources given that we have access to a bare-metal Talos II.
We can literally run arbitrary instructions on in remotely - this means a simulator should be low priority given that we have an October deadline.
Yehowshua
_______________________________________________
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] 4 Simulators?
From: Michael Nolan @ 2020-05-07 17:38 UTC (permalink / raw)
To: libre-riscv-dev
In-Reply-To: <EDCC7F2A-1F18-4603-9423-666FCFF4C7FB@gatech.edu>
On 5/7/20 1:33 PM, Yehowshua wrote:
> Hello all - I recently became aware that there a 4 different POWER simulators being worked on?
> This is not a good use of resources given that we have access to a bare-metal Talos II.
>
> We can literally run arbitrary instructions on in remotely - this means a simulator should be low priority given that we have an October deadline.
>
> Yehowshua
> _______________________________________________
> libre-riscv-dev mailing list
> libre-riscv-dev@lists.libre-riscv.org
> http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev
>
I think we do need *a* simulator other than the talos II server. When we
have the core somewhat working, we will want to run a bunch of tests on
it, and check those against a reference. While we can use the Talos as a
reference, running code on it (I assume) will be a somewhat manual
process, not really amenable to unit testing. If we have a simulator
though, we can write (or randomly generate) a bunch of unit tests, and
run them in the simulator and our core and compare the results. So we
will need *a* simulator.
However, I do agree that we don't need 4.
--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
* Re: [libre-riscv-dev] 4 Simulators?
From: Luke Kenneth Casson Leighton @ 2020-05-07 18:23 UTC (permalink / raw)
To: Libre-RISCV General Development
In-Reply-To: <EDCC7F2A-1F18-4603-9423-666FCFF4C7FB@gatech.edu>
On Thursday, May 7, 2020, Yehowshua <yimmanuel3@gatech.edu> wrote:
> Hello all - I recently became aware that there a 4 different POWER
> simulators being worked on?
no. two.
pearpc and the autogenerated one.
This is not a good use of resources given that we have access to a
> bare-metal Talos II.
it really does not help us except when it comes to FP interoperability.
also the qemu pipe is already done and has rhe advantage of acting like a
local POWER machine instead of forcing development to have a permanent hard
internet line.
> We can literally run arbitrary instructions on in remotely - this means a
> simulator should be low priority given that we have an October deadline.
unlike in RISCV we do not have a cycle accurate simulator, we do not have
formal test suites or in fact any asseembly level test suites at all.
therefore as i outlined already a few weeks ago the strategy is to side by
side bootstrap up the following simultaneously and run the EXACT same
instructions on each:
* python simulator
* pearpc simulator pipe (when ready)
* qemu pipe
* bare metal pipe (run on the talos server)
* hardware simulation pipe
i would suggest adding power-gem5 to that list except the build
dependencies are enormous and it leaks 500mb/sec.
by doing side by side comparisons we have a means to bootstrap our way up.
cutting back actually *slows* the project by removing routes to verify the
correctness of our implementation.
also because "doing is learning" the development of the simulators teaches
us what is needed far faster than the HDL could ever be written without
that "experience"
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 07may2020 update
From: Michael Nolan @ 2020-05-07 18:25 UTC (permalink / raw)
To: libre-riscv-dev
In-Reply-To: <CAPweEDzFVoyxLpGK+9=nEniSOerHd469HQSM7qUXZ65Q6BWnbg@mail.gmail.com>
I'm continuing working on the power_pseudo simulator, I think I have the
condition registers working now. I think carry and overflow handling is
next.
--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
* Re: [libre-riscv-dev] 4 Simulators?
From: Yehowshua @ 2020-05-07 18:30 UTC (permalink / raw)
To: Libre-RISCV General Development
In-Reply-To: <CAPweEDwckA0jHpfKSy75MRbL6qGX2sc=MGt-PKQMy7nUZDkkqQ@mail.gmail.com>
So a total of two simulators - with
Tobias, Jacob, Michael and Alain working on them?
I’m guessing Tobias, Michael, and Jacob are working on the Python Simulator?
Alain is working on the PearPC simulator?
Yehowshua
_______________________________________________
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] 4 Simulators?
From: Jacob Lifshay @ 2020-05-07 18:40 UTC (permalink / raw)
To: Libre-RISCV General Development
In-Reply-To: <92D0A28A-A5D7-4389-8C58-19F217A618E3@gatech.edu>
On Thu, May 7, 2020, 11:30 Yehowshua <yimmanuel3@gatech.edu> wrote:
> So a total of two simulators - with
> Tobias, Jacob, Michael and Alain working on them?
>
> I’m guessing Tobias, Michael, and Jacob are working on the Python
> Simulator?
>
I'm not working on any simulators (except for the experimental
spectre-proof cpu design which is currently on-hold).
Jacob Lifshay
_______________________________________________
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] 4 Simulators?
From: Yehowshua @ 2020-05-07 18:41 UTC (permalink / raw)
To: Libre-RISCV General Development
In-Reply-To: <CAC2bXD7J=z5kbEBOtsmbh==awW6mBeKuFmsGJDOW6fr6y=sA6Q@mail.gmail.com>
>
> I'm not working on any simulators (except for the experimental
> spectre-proof cpu design which is currently on-hold).
>
> Jacob Lifshay
Oh - you said LD/ST here.
http://lists.libre-riscv.org/pipermail/libre-riscv-dev/2020-May/006327.html
Yehowshua
_______________________________________________
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] 4 Simulators?
From: Jacob Lifshay @ 2020-05-07 18:46 UTC (permalink / raw)
To: Libre-RISCV General Development
In-Reply-To: <1CEFB975-A644-477E-967B-50C6B0312228@gatech.edu>
On Thu, May 7, 2020, 11:42 Yehowshua <yimmanuel3@gatech.edu> wrote:
>
> >
> > I'm not working on any simulators (except for the experimental
> > spectre-proof cpu design which is currently on-hold).
> >
> > Jacob Lifshay
>
> Oh - you said LD/ST here.
>
:)
Though, I will most likely be writing a reference interpreter for
shader-compiler-ir once I'm back to working on Kazan. Sorta simulator.
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
* Re: [libre-riscv-dev] daily kan-ban 07may2020 update
From: Tobias Platen @ 2020-05-07 18:48 UTC (permalink / raw)
To: libre-riscv-dev
In-Reply-To: <1e472501-2473-6be6-f959-8c1806d7c37a@gmail.com>
On Thu, 7 May 2020 14:25:14 -0400
Michael Nolan <mtnolan2640@gmail.com> wrote:
Today I will have a short look at the changes from the last few days.
Tobias
> I'm continuing working on the power_pseudo simulator, I think I have the
> condition registers working now. I think carry and overflow handling is
> next.
>
> --Michael
>
>
> _______________________________________________
> libre-riscv-dev mailing list
> libre-riscv-dev@lists.libre-riscv.org
> http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev
--
Tobias Platen <libre-soc[at]platen-software[dot]de>
_______________________________________________
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] 4 Simulators?
From: Tobias Platen @ 2020-05-07 18:54 UTC (permalink / raw)
To: libre-riscv-dev
In-Reply-To: <92D0A28A-A5D7-4389-8C58-19F217A618E3@gatech.edu>
On Thu, 7 May 2020 14:30:07 -0400
Yehowshua <yimmanuel3@gatech.edu> wrote:
> So a total of two simulators - with
> Tobias, Jacob, Michael and Alain working on them?
>
> I’m guessing Tobias, Michael, and Jacob are working on the Python Simulator?
Yes I was working on the Python Simulator, and I was also looking into the sourcecode of the Dolphin emulator.
>
> Alain is working on the PearPC simulator?
>
> Yehowshua
> _______________________________________________
> libre-riscv-dev mailing list
> libre-riscv-dev@lists.libre-riscv.org
> http://lists.libre-riscv.org/mailman/listinfo/libre-riscv-dev
--
Tobias Platen <libre-soc[at]platen-software[dot]de>
_______________________________________________
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] 4 Simulators?
From: Luke Kenneth Casson Leighton @ 2020-05-07 18:53 UTC (permalink / raw)
To: Libre-RISCV General Development
In-Reply-To: <92D0A28A-A5D7-4389-8C58-19F217A618E3@gatech.edu>
On Thu, May 7, 2020 at 7:30 PM Yehowshua <yimmanuel3@gatech.edu> wrote:
> So a total of two simulators - with
> Tobias, Jacob, Michael and Alain working on them?
i noticed that tobias had just committed to the one that had only 1-2
day's work done on it (for michael to do a proof-of-concept and get
qemu and GNU-as up and running).
i moved it to the "unused" subdirectory.
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 07may2020 update
From: Luke Kenneth Casson Leighton @ 2020-05-07 18:55 UTC (permalink / raw)
To: Libre-RISCV General Development
In-Reply-To: <1e472501-2473-6be6-f959-8c1806d7c37a@gmail.com>
On Thursday, May 7, 2020, Michael Nolan <mtnolan2640@gmail.com> wrote:
> I'm continuing working on the power_pseudo simulator, I think I have the
> condition registers working now. I think carry and overflow handling is
> next.
great.
i could really do with understanding the relationship between condition
registers and branches.
that will allow me to start the Branch Computation Unit.
except probably first i'd need to do the Condition Regfile.
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] 4 Simulators?
From: Luke Kenneth Casson Leighton @ 2020-05-07 19:02 UTC (permalink / raw)
To: Libre-RISCV General Development
In-Reply-To: <20200507205411.809af7c024d65954ce093d1b@platen-software.de>
On Thu, May 7, 2020 at 7:54 PM Tobias Platen
<libre-soc@platen-software.de> wrote:
> Yes I was working on the Python Simulator,
the discussions in the bugtracker center around the auto-generated
one. i appreciate that there is a lot going on.
> and I was also looking into the sourcecode of the Dolphin emulator.
https://github.com/dolphin-emu/dolphin/blob/master/Source/Core/Core/PowerPC/Interpreter/Interpreter_Integer.cpp
interesting. this code *might* be a better starting point for SV
augmentation than pearpc, because c++ provides the means to do
operator-overloads (and templates).
btw do not waste time looking at the JIT version. we absolutely do
not need a JIT interpreted emulator. we need a *cycle accurate*
emulator.
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 07may2020 update
From: Jacob Lifshay @ 2020-05-07 23:23 UTC (permalink / raw)
To: Libre-RISCV General Development
In-Reply-To: <CAPweEDzFVoyxLpGK+9=nEniSOerHd469HQSM7qUXZ65Q6BWnbg@mail.gmail.com>
I attended the OpenPower virtual coffee meeting, there was an interesting
suggestion to see if we could move to a later tapeout than october, to
avoid burnout due to deadline crunch.
I also weighed in on nmigen signed/unsigned sign extension/zero
extension/truncation operators:
https://github.com/nmigen/nmigen/issues/381
still slowly getting through the load/store unit stuff, progress may be a
little slow since I have a bunch of stuff I need to do at home as well.
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
* Re: [libre-riscv-dev] daily kan-ban 07may2020 update
From: Yehowshua @ 2020-05-07 23:28 UTC (permalink / raw)
To: Libre-RISCV General Development
In-Reply-To: <CAC2bXD6EWrcrxbva=XeUokF3Csu4t0cF0LWVw8mCwZUJ55Ku7g@mail.gmail.com>
> there was an interesting
> suggestion to see if we could move to a later tapeout than october, to
> avoid burnout due to deadline crunch.
Luke and I we’re just discussing this.
Yehowshua
_______________________________________________
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] MMU + TLB idea
From: Michael Nolan @ 2020-05-07 23:39 UTC (permalink / raw)
To: Libre-RISCV General Development
I remember some time ago working on a powerpc core that (I think) had a
partly software MMU. It got me thinking: we could have a TLB and the
machinery to look up entries in it and translate the addresses in
hardware, and implement the page table walking and such in software (via
a trap). This would I think cut down the work in implementing the MMU by
making it so we don't need to implement the page table walking machinery
in hardware, and make it a bit easier to test. The downside of course
would be that TLB misses would be slower, but I think it might be a
decent trade-off for our first attempt at the MMU.
Thoughts?
--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
* Re: [libre-riscv-dev] MMU + TLB idea
From: Luke Kenneth Casson Leighton @ 2020-05-08 0:40 UTC (permalink / raw)
To: Libre-RISCV General Development
In-Reply-To: <6a63afa2-f425-9490-be9f-a05e34cadc9c@gmail.com>
On Friday, May 8, 2020, Michael Nolan <mtnolan2640@gmail.com> wrote:
> I remember some time ago working on a powerpc core that (I think) had a
> partly software MMU. It got me thinking: we could have a TLB and the
> machinery to look up entries in it and translate the addresses in hardware,
> and implement the page table walking and such in software (via a trap).
ha, funny. i went through this idea twice. once in 2012 and again 18
months ago.
> This would I think cut down the work in implementing the MMU by making it
> so we don't need to implement the page table walking machinery in hardware,
> and make it a bit easier to test.
yes.
from a unit test perspective i like it. it reduces the dependent
components for each test.
> The downside of course would be that TLB misses would be slower, but I
> think it might be a decent trade-off for our first attempt at the MMU.
yyyeahh. a TLB (basically a cache for MMU lookups) is probably the
critical component, first in the chain.
without a TLB, every MMU lookup (which is *all* memory lookups) can take 4
or more cycles on *every* memory access, both I and D.
so we will definitely need a TLB. there exists code which i moved to the
"unused" directory that implements one, so we do not need to duplicate that
work (simply adapt it to POWER).
this code was translated from ariane:
https://git.libre-soc.org/?p=soc.git;a=blob;f=src/unused/TLB/ariane/tlb.py;h=72b67a2dcc56aa76aca561f5da3345a99c36c2ab;hb=HEAD#l40
note how small the TLB is - only an 8 entry CAM! this is pretty normal, it
saves power and also means it can do single cycle requests.
the wikipedia page on TLB says that intel do i think three nested levels of
TLB Cache (!) going from 8 entries (single cycle) up to i think 2048
entries (multi cycle), to avoid missing clock cycles.
the one downside of a software MMU: now a linux kernel driver (and
potentially u-boot) supporting that design concept is on the list of
dependencies.
that in turn, to ensure a reduced dependency on the hardware team, means
implementing the software MMU in an emulator (pearpc, dolphin, qemu, gem5,
other).
this would allow the new MMU linux kernel driver to be developed and tested
in advance, then the *exact* same linux kernel binary can be run on the
newly developed hardware, and we are not trying to test 3 newly developed
"unknowns" at once (a guaranteed nightmare).
so on first glance it seems like it would save time.
unfortunately, despite the usefulness of not having an MMU at the unit test
level, as an end goal for a final product, due to the additional *software
development*, it actually *adds* time!
however if there is *already* support for a POWER "soft MMU" in the linux
kernel, we can try it out.
honestly though on balance, by conforming to the RADIX MMU already
suppported for POWER in the linux kernel, we join a "well tested" code path.
with Paul Mackerras working on that for microwatt, (right noe), if we
simply wait a while we will have working code that we can by-rote translate
to nmigen. and i kinda enjoy those mind numbing tasks :)
yes, i thought that a softmmu would save time, too, Michael :)
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
page: next prev 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