Documentation on MOVTS/MOVFS encoding

Any technical questions about the Epiphany chip and Parallella HW Platform.

Moderator: aolofsson

Re: Documentation on MOVTS/MOVFS encoding

Postby timpart » Mon Dec 23, 2013 5:00 pm

I may be able to test experimental features. I'm happy writing Epiphany assembler. Presumably you don't expect them to cause any permanent damage to the chip if they don't work out?

Work can be hectic at times, leaving me tired in the evenings, so I don't always have as much time to spend on hobbies as I'd like to. Hopefully I'll have my boards by the end of January.

Tim
timpart
 
Posts: 302
Joined: Mon Dec 17, 2012 3:25 am
Location: UK

Re: Documentation on MOVTS/MOVFS encoding

Postby alexrp » Sat Dec 28, 2013 7:10 pm

Andreas, one thing: In binutils, 0xF0410 is understood as "iab" by the assembler. Given that it's usable in assembly language, can you clarify what it is?

Similarly, 'memconfig' and 'memreserve' are understood by the assembler, but are not documented.
alexrp
 
Posts: 154
Joined: Mon Dec 17, 2012 3:22 am
Location: Thisted, Denmark

Re: Documentation on MOVTS/MOVFS encoding

Postby alexrp » Mon Dec 30, 2013 7:11 am

And another: In the 16-bit encoding, there is no M0-M1. Does this just mean that the register group is 0b00?
alexrp
 
Posts: 154
Joined: Mon Dec 17, 2012 3:22 am
Location: Thisted, Denmark

Re: Documentation on MOVTS/MOVFS encoding

Postby aolofsson » Thu Jan 02, 2014 4:00 am

Correct!
User avatar
aolofsson
 
Posts: 1005
Joined: Tue Dec 11, 2012 6:59 pm
Location: Lexington, Massachusetts,USA

Re: Documentation on MOVTS/MOVFS encoding

Postby alexrp » Sat Jan 11, 2014 6:11 pm

@aolofsson so is the "iab" register part of the user/kernel mode stuff, or something else? It's a supported register by the toolchain, so I need to have some idea how to handle it in my simulator. If it's part of the user/kernel mode stuff, I can wait, though.
alexrp
 
Posts: 154
Joined: Mon Dec 17, 2012 3:22 am
Location: Thisted, Denmark

Re: Documentation on MOVTS/MOVFS encoding

Postby mhonman » Sat Jan 11, 2014 11:17 pm

Re experimental feature testing, I'm also up for that!
mhonman
 
Posts: 112
Joined: Thu Apr 25, 2013 2:22 pm

Re: Documentation on MOVTS/MOVFS encoding

Postby aolofsson » Wed Jan 15, 2014 3:47 am

alexrp wrote:@aolofsson so is the "iab" register part of the user/kernel mode stuff, or something else? It's a supported register by the toolchain, so I need to have some idea how to handle it in my simulator. If it's part of the user/kernel mode stuff, I can wait, though.


alex,
The IAB register is the holding buffer that receives the instruction to be decoded by the core. When fetching an instruction from external memory (outside the core) the address of this register is used as a "return to sender" address. Once the instruction arrives in the register, the instruction decode starts. I don't really see any reason to expose it to the user, but it was convenient to make it memory mapped for initial debugging purposes. It doesn't belong in the tool chain and we will remove it (although I could swear we removed it three times already..)
Andreas
User avatar
aolofsson
 
Posts: 1005
Joined: Tue Dec 11, 2012 6:59 pm
Location: Lexington, Massachusetts,USA

Re: Documentation on MOVTS/MOVFS encoding

Postby alexrp » Wed Jan 15, 2014 4:35 am

OK, thanks! I'll just simply make it a reserved register area in my simulator.
alexrp
 
Posts: 154
Joined: Mon Dec 17, 2012 3:22 am
Location: Thisted, Denmark

Re: Documentation on MOVTS/MOVFS encoding

Postby mhonman » Wed Jan 15, 2014 9:37 am

aolofsson wrote:When fetching an instruction from external memory (outside the core) the address of this register is used as a "return to sender" address. Once the instruction arrives in the register, the instruction decode starts.


Would that be an 8-byte "instruction" fetch as per arch. ref. section 6.2 "Mesh-node crossbar switch"?
mhonman
 
Posts: 112
Joined: Thu Apr 25, 2013 2:22 pm

Previous

Return to Epiphany and Parallella Q & A

Who is online

Users browsing this forum: No registered users and 6 guests

cron