Messages in this thread |  | | Date | Wed, 20 Sep 2000 08:27:18 -0500 | Subject | Re: __ucmpdi2 | From | Peter Samuelson <> |
| |
[Jeremy Higdon] > Is there a little FAQ of C constructions you should not use if you > are writing kernel code? It took a little doing to figure out that > it was the switch that was causing trouble and not the shifts and > arithmetic operations, and I'd like to avoid that in the future :-).
Here are some of the big ones:
* 64-bit division. Most such is by powers of two so you can use shifts instead. Probably the compiler should optimize this case automatically, but gcc doesn't, currently.
* floating point. Use fixed point integers. Floating point registers (at least on i386) are not saved/restored in the right places (because doing so is very expensive on some CPUs), so you'll clobber someone else's context.
* MMX. See floating point.
* excess stack usage. The kernel stack is very very small, so don't waste it. Large automatic arrays are BAD. Use globals or dynamic allocation.
* C++ exceptions, run-time types (RTTI) and other features-of-the-month. Stick to C and you'll be OK.
* global variables initialized to 0. Not necessary, as uninitialized globals are zeroed automatically, and that way they don't take up space in the on-disk image. This is a concern for embedded systems.
Also make sure to read Documentation/CodingStyle for, well, coding style.
Peter - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org Please read the FAQ at http://www.tux.org/lkml/
|  |