EVT::isRound on non-8-bit byte targets

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|

EVT::isRound on non-8-bit byte targets

Sam Cristall
I'm new to LLVM dev, but I have been working with a target with a
minimum addressable byte of 16-bits.  I found that in
DAGCombiner::visitAND, EVT::isRound could create i8 loads on my 16-bit
target which are ultimately invalid.  EVT::isRound appears to use a
hard-coded 8, rather than pulling the targets BitsPerByte field.  Is this
a potential bug or is there a better way to address this?  Hard coding a
16 in the isRound field fixes the issue for me.

Cheers,
Sam
_______________________________________________
LLVM Developers mailing list
[hidden email]         http://llvm.cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev
Reply | Threaded
Open this post in threaded view
|

Re: EVT::isRound on non-8-bit byte targets

Duncan Sands
Hi Sam,

On 09/07/13 17:01, Sam Cristall wrote:
> I'm new to LLVM dev, but I have been working with a target with a
> minimum addressable byte of 16-bits.  I found that in
> DAGCombiner::visitAND, EVT::isRound could create i8 loads on my 16-bit
> target which are ultimately invalid.  EVT::isRound appears to use a
> hard-coded 8, rather than pulling the targets BitsPerByte field.  Is this
> a potential bug or is there a better way to address this?  Hard coding a
> 16 in the isRound field fixes the issue for me.

last time I checked LLVM did not have a BitsPerByte field.  Are you working
with a private copy of LLVM where someone has added support for non-octet
bytes?  If so, I guess they forgot to adjust isRound.

Ciao, Duncan.

_______________________________________________
LLVM Developers mailing list
[hidden email]         http://llvm.cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev
Reply | Threaded
Open this post in threaded view
|

Re: EVT::isRound on non-8-bit byte targets

Sam Cristall
Hi Duncan,

It appears you are correct -- I didn't realize this was a wart of my
fork, thank you for your time!

Cheers,
Sam
_______________________________________________
LLVM Developers mailing list
[hidden email]         http://llvm.cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev