Messages in this thread | | | From | (Linus Torvalds) | Subject | Re: Linux/IA-64 byte order | Date | 9 Mar 1999 00:36:24 GMT |
| |
In article <199903090013.LAA21453@vindaloo.atnf.CSIRO.AU>, Richard Gooch <rgooch@atnf.csiro.au> wrote: > Hi, all. I've been discussing the byte order that Linux/IA-64 will >have with David Mosberger from HP. I'm arguing for big-endian to be >used.
Not a chance in hell.
The whole point of IA-64 is to run x86 binaries while still being able to do large data-sets (if you didn't want to run x86 binaries you should just buy an alpha instead, and forget about IA-64).
And I'm not going to accept a IA-64 port that does big-endian IA-64 mode and little-endian x86 mode. That's just too ugly for words.
>I implore you: please reconsider your decision. Don't punish Linux >because of the x86 legacy.
Buy an ultra64 if you need big-endian and 64 bits. Really.
We're going to run x86 binaries on any merced we ever have in the near future - we may certainly run native merced binaries too, but running x86 binaries is a given. Which would mean that if the Linux/merced port was natively big-endian, it would have to do a lot of extra conversions from x86 user space to kernel space, for no good technical reason.
Mixing endianness on the fly is certainly possible, but stupid unless you have some REALLY good reason for it. And quite frankly, there are NEVER any good technical reasons for considering one endianness over another (it's a completely arbitrary thing).
Linus
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |