[U-Boot-Users] X-Scale mtune parameters?
G'Day everyone,
Quick question as to why we are using mtune=strongarm1100 for the xscale? (and -march=armv4).
I've been using -march=armv5 and -mtune=xscale successfully for the past few weeks using arm-linux-gcc version 3.3.2. These are more appropriate settings for the pxa.
(This is PLATFORM_CPPFLAGS, line 28 in cpu/pxa/config.mk)
(Point of the question: can I patch this?).
Cheers,
Dave. - David Snowdon PhD Student National ICT Australia (NICTA) Phone: 9385 7355 Email: daves@cse.unsw.edu.au
Dear David,
in message 9AE6C62E-0098-11D9-9008-000393C7979E@student.unsw.edu.au you wrote:
Quick question as to why we are using mtune=strongarm1100 for the xscale? (and -march=armv4).
Historical reasons?
I've been using -march=armv5 and -mtune=xscale successfully for the past few weeks using arm-linux-gcc version 3.3.2. These are more appropriate settings for the pxa.
(This is PLATFORM_CPPFLAGS, line 28 in cpu/pxa/config.mk)
(Point of the question: can I patch this?).
Will this work for people using a GCC 2.95.x based toolchain?
Best regards,
Wolfgang Denk
G'Day Wolfgang,
I've been using -march=armv5 and -mtune=xscale successfully for the past few weeks using arm-linux-gcc version 3.3.2. These are more appropriate settings for the pxa.
(This is PLATFORM_CPPFLAGS, line 28 in cpu/pxa/config.mk)
(Point of the question: can I patch this?).
Will this work for people using a GCC 2.95.x based toolchain?
No, I don't believe so (at least, the 2.95.x toolchain that I have here doesn't like it).
Is it necessary to keep this compatibility? (What is the motivation?).
Cheers,
Dave.
In message 8C35D986-00BA-11D9-9008-000393C7979E@student.unsw.edu.au you wrote:
Will this work for people using a GCC 2.95.x based toolchain?
No, I don't believe so (at least, the 2.95.x toolchain that I have here doesn't like it).
Is it necessary to keep this compatibility? (What is the motivation?).
I see no reason to break compatibility without need. It should be possible to detect the version of GCC at runtime and make appropriate adjustments to the compiler options.
Best regards,
Wolfgang Denk
participants (2)
-
David Snowdon -
Wolfgang Denk