Re: broken local_t on i386

Ar Llu, 2006-06-12 am 20:11 +0200, ysgrifennodd Andi Kleen:
The segment register needs an offset. So you need the linker to generate
the offset from the base of the per cpu segment somehow. At compile time the
address is not known so it cannot be done then.

What happens if you put a section at zero and a section at non-fixed
address (aligned). In the asm macros you stick the variable in both,
using the zero based one for linker symbols and the non zero based one
for data, then discard the zero based one.

That used to work for old binutils which didn't care/spot if you were
discarding material you actually linked against. Not sure what todays
binutils does with it.

Alan

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Relevant Pages

  • Re: broken local_t on i386
    ... the offset from the base of the per cpu segment somehow. ... At compile time the ... What happens if you put a section at zero and a section at non-fixed ...
    (Linux-Kernel)
  • Re: m_hWnd getting lost its value in OnKillActive() function of Property Page.
    ... pages containing some controls on each page, to the property sheet and ... i haven't found any compile time or run time errors in the code. ... Can you just tell me what might be the reason for m_hWnd to become zero ... you could try setting a data breakpoint on the ...
    (microsoft.public.vc.mfc)
  • m_hWnd getting lost its value in OnKillActive() function of Property Page.
    ... pages containing some controls on each page, to the property sheet and ... i haven't found any compile time or run time errors in the code. ... Can you just tell me what might be the reason for m_hWnd to become zero ...
    (microsoft.public.vc.mfc)
  • Re: Reserved Words
    ... like repeat factors and format reversion, ... It wouldn't seem so hard to check at compile time, but yes, ... conditions (zero array size, zero implied-do limit), then yes, it is so ... but that's true of zillions of things that aren't constraints. ...
    (comp.lang.fortran)
  • Re: is "i0" input edit descriptor allowed by standard?
    ... On 3/20/13 2:19 PM, glen herrmannsfeldt wrote: ... it does not require that compilers have the capability to ... It doesn't notice at compile time, but at run time even before it ... There's lots of rules that say things can't be zero; but they usually only apply to some particular instance of using the thing. ...
    (comp.lang.fortran)