msp430-elf incorrect hwmult inference

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

msp430-elf incorrect hwmult inference

Peter Bigot-4
Just noticed that integer multiplication is broken in upstream msp430-elf
on msp430fr5969, apparently because gcc is hard-coding the list of MCUs and
deciding which hardware multiplier to use based on that.  None of the FR5
chips are in any of those hard-coded lists.  Interestingly, some FR2/FR4
ones are listed.

It does work if you explicitly add '-mhwmult=f5series' but the whole point
of -mmcu=foo is to set the correct flags for device foo.

This is a really lousy design choice.  TI folks: we had the toolchain fully
decoupled from the device list with msp430mcu four years ago in mspgcc,
allowing the set of supported devices to be updated without breaking the
toolchain.

Why was this solution chosen?

Peter

------------------------------------------------------------------------------
Comprehensive Server Monitoring with Site24x7.
Monitor 10 servers for $9/Month.
Get alerted through email, SMS, voice calls or mobile push notifications.
Take corrective actions from your mobile device.
http://pubads.g.doubleclick.net/gampad/clk?id=154624111&iu=/4140/ostg.clktrk
_______________________________________________
Mspgcc-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/mspgcc-users
Reply | Threaded
Open this post in threaded view
|

Re: msp430-elf incorrect hwmult inference

Peter Bigot-4
Since I haven't seen a response, I've gone ahead and made
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63953 to record the absence of
the FR5xx/6xx MCUs from the hard-coded list in upstream gcc.

Peter

On Wed, Nov 12, 2014 at 10:45 AM, Peter Bigot <[hidden email]> wrote:

> Just noticed that integer multiplication is broken in upstream msp430-elf
> on msp430fr5969, apparently because gcc is hard-coding the list of MCUs and
> deciding which hardware multiplier to use based on that.  None of the FR5
> chips are in any of those hard-coded lists.  Interestingly, some FR2/FR4
> ones are listed.
>
> It does work if you explicitly add '-mhwmult=f5series' but the whole point
> of -mmcu=foo is to set the correct flags for device foo.
>
> This is a really lousy design choice.  TI folks: we had the toolchain
> fully decoupled from the device list with msp430mcu four years ago in
> mspgcc, allowing the set of supported devices to be updated without
> breaking the toolchain.
>
> Why was this solution chosen?
>
> Peter
>
>

------------------------------------------------------------------------------
Download BIRT iHub F-Type - The Free Enterprise-Grade BIRT Server
from Actuate! Instantly Supercharge Your Business Reports and Dashboards
with Interactivity, Sharing, Native Excel Exports, App Integration & more
Get technology previously reserved for billion-dollar corporations, FREE
http://pubads.g.doubleclick.net/gampad/clk?id=157005751&iu=/4140/ostg.clktrk
_______________________________________________
Mspgcc-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/mspgcc-users