On Fri, Oct 07, 2005 at 07:06:21PM +0100, Felix Oxley wrote: > > For variables marked __initdata, the "*foo" form causes only the > > pointer, not the string itself, to be dropped from the kernel image, > > which is a bug.
Not a bug. A missing feature, perhaps, but not a bug. > > http://lists.osdl.org/pipermail/kernel-janitors/2005-May/004146.html He did something wrong. The size numbers he gives are obviously wrong -- the data field is 0 in cases that cannot be true. I don't replicate his results either. E.g. text data bss dec hex filename 11 8 0 19 13 s1.o 0 11 0 11 b s2.o That said, some ABIs over-align arrays, and so you might get more padding with arrays than raw strings. But it'll be less than the size of an extra pointer variable. > The second part of the task was descibed thus: > > 2) "unsigned int" is preferred to "int", it generates better asm code on > > all platforms except sh5. This replacement needs to be done manually, > > because often 'int' is required due to negative values -Exxx commonly > > passed as error values. > > Can anyone comment on the accuracy of this? Definitely false for mips and alpha, so I really doubt the accuracy of their survey of all platforms. My guess is that it allows for the possibility of better code for x86_64 (but is not totally realized; there's a large outstanding gcc patch that attempts to fix this), but is mostly a wash elsewhere. r~
