OK, I've spotted my mistake - but further digging is keeping me confused about what the expected behaviour is, especially with modern gcc's * Dr. David Alan Gilbert (linux@xxxxxxxxxxx) wrote: > Hi, > I noticed the apparently simple entry in the TODO list saying: > > From: Jeff Garzik > > 1) The string form > > [const] char *foo = "blah"; > creates two variables in the final assembly output, a > static string, and a char pointer to the static string. > The alternate string form > > [const] char foo[] = "blah"; > is better because it declares a single variable. > > 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. Using the "foo[]" form with > regular 'ole local variables also makes the assembly shorter. <snip my original message> I've been trying this out on a little standalone hello world. I'd originally replaced char *foo = "blah" by char foo[] = "blah" and been surprised at the extra code generated - but actually thinking about it, inside a function this makes sense - every time through the compiler has to refresh the array since you might have changed it - so I'm not sure that substitution ever makes sense. Life gets more interesting with const - and I don't understand why. On gcc 4.2 if I declare const char foo[] = "blah" then blah ends up in .rodata.str, if you use it multiple times it gets shared. However with 4.3.4 or 4.4.1 (all ubuntu packaged versions) the const char foo[]="blah" case generates code to write the string to the stack on each function call - why??? Is this a compiler bug or is there a reason for this change? Finally, static const char foo[] = "blah"; This ends up putting the constants in a .rodata section rather than .rodata.str and doesn't share multiple uses of the same string within the same .c file. These were all done for local variables; with global variables it doesn't look much better; const char foo[] does put things in .rodata but again none-shared, as does static const char foo[]. All these tests were done on x64. So, too summarise I'm a bit confused, I think the advise to change char* foo="hello" to char foo[]="hello" appears to be bogus without the const. With the const then I'm not sure the newer compilers are behaving as expected. Sharing only seems to happen with the char* foo="hello". Maybe in the case of initdata some of this helps, but I can't see outside that - but I can see it causing code duplication or with the newer compilers const char foo[] generating more code for no good reason. Dave -- -----Open up your eyes, open up your mind, open up your code ------- / Dr. David Alan Gilbert | Running GNU/Linux on Alpha,68K| Happy \ \ gro.gilbert @ treblig.org | MIPS,x86,ARM,SPARC,PPC & HPPA | In Hex / \ _________________________|_____ http://www.treblig.org |_______/ -- To unsubscribe from this list: send the line "unsubscribe kernel-janitors" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
