Re: IDE/SUPPORT_VLB_SYNC in 2.5.x

[Martin Dalecki]

Geert Uytterhoeven wrote:
> SUPPORT_VLB_SYNC is now unconditionally hardcoded to 1 in
> drivers/ide/ide-taskfile.c.
> 
> While most architectures disable it, which no longer works:
> 
> | tux$ find include -type f | xargs grep SUPPORT_VLB_SYNC
> | include/asm-cris/ide.h:#undef SUPPORT_VLB_SYNC
> | include/asm-cris/ide.h:#define SUPPORT_VLB_SYNC 0
> | include/asm-m68k/ide.h:#undef SUPPORT_VLB_SYNC
> | include/asm-m68k/ide.h:#define SUPPORT_VLB_SYNC 0
> | include/asm-mips/ide.h:#undef  SUPPORT_VLB_SYNC
> | include/asm-mips/ide.h:#define SUPPORT_VLB_SYNC 0
> | include/asm-ppc/ide.h:#undef	SUPPORT_VLB_SYNC
> | include/asm-ppc/ide.h:#define SUPPORT_VLB_SYNC	0
> | include/asm-sparc/ide.h:#undef  SUPPORT_VLB_SYNC
> | include/asm-sparc/ide.h:#define SUPPORT_VLB_SYNC 0
> | include/asm-sparc64/ide.h:#undef  SUPPORT_VLB_SYNC
> | include/asm-sparc64/ide.h:#define SUPPORT_VLB_SYNC 0
> | include/linux/ide.h:#ifndef SUPPORT_VLB_SYNC		/* 1 to support weird 32-bit chips */
> | include/linux/ide.h:#define SUPPORT_VLB_SYNC	1	/* 0 to reduce kernel size */
> 
> Wouldn't it be better to enable it on architectures which can have a VESA local
> bus (ia32 only?) only?

Thank you for pointing it out. Of course it just shouldn't be enabled
unconditionally there. Apparently it "slipped in" during some
compiling for "code coverage". My appologies for the inconvenience.
It will be disabled in the next patch round.