Re: [PATCH v2] powerpc/boot: Delete unneeded .globl _zimage_start
On Wed, 2020-03-25 at 16:42:57 UTC, Fangrui Song wrote: > .globl sets the symbol binding to STB_GLOBAL while .weak sets the > binding to STB_WEAK. GNU as let .weak override .globl since binutils-gdb > 5ca547dc2399a0a5d9f20626d4bf5547c3ccfddd (1996). Clang integrated > assembler let the last win but it may error in the future. > > Since it is a convention that only one binding directive is used, just > delete .globl. > > Fixes: cd197ffcf10b "[POWERPC] zImage: Cleanup and improve zImage entry point" > Fixes: ee9d21b3b358 "powerpc/boot: Ensure _zimage_start is a weak symbol" > Link: https://github.com/ClangBuiltLinux/linux/issues/937 > Signed-off-by: Fangrui Song > Cc: Alan Modra > Cc: Joel Stanley > Cc: Michael Ellerman > Cc: Nick Desaulniers > Cc: Segher Boessenkool > Cc: clang-built-li...@googlegroups.com Applied to powerpc next, thanks. https://git.kernel.org/powerpc/c/968339fad422a58312f67718691b717dac45c399 cheers
Re: [PATCH v2] powerpc/boot: Delete unneeded .globl _zimage_start
On Fri, Mar 27, 2020 at 09:50:54AM -0700, Fangrui Song wrote: > We aim for compatibility with GNU in many aspects to make it easier for > people to switch over. However, just because there is a subtle behavior > in GNU toolchain does not mean we need to emulate that behavior. It isn't subtle. It is the explicit documented behaviour. > With > all due respect, there are a large quantity of legacy behaviors we don't > want to support. And it isn't legacy at all. It is simply the defined semantics. It is semantics that real code relies on (and has for ages) as well. Segher
Re: [PATCH v2] powerpc/boot: Delete unneeded .globl _zimage_start
On Thu, Mar 26, 2020 at 03:26:12PM -0700, Fangrui Song wrote: > On 2020-03-26, Segher Boessenkool wrote: > >On Wed, Mar 25, 2020 at 09:42:57AM -0700, Fangrui Song wrote: > >>.globl sets the symbol binding to STB_GLOBAL while .weak sets the > >>binding to STB_WEAK. GNU as let .weak override .globl since binutils-gdb > >>5ca547dc2399a0a5d9f20626d4bf5547c3ccfddd (1996). Clang integrated > >>assembler let the last win but it may error in the future. > > > >GNU AS works for more than just ELF. The way the assembler language > >is defined, it is not .weak overriding .globl -- instead, .weak sets a > >symbol attribute. On an existing symbol (but it creates on if there is > >none yet). > > > >Clang is buggy if it does not allow valid (and perfectly normal) > >assembler code like this. > > https://sourceware.org/pipermail/binutils/2020-March/110399.html > > Alan: "I think it is completely fine for you to make the llvm assembler > error on inconsistent binding, or the last directive win. Either of > those behaviours is logical and good, but you quite possibly will run > into a need to fix more user assembly. This would be fine and consistent behaviour, of course. But it is not appropriate if you want to pretend to be compatible to GNU toolchains. Which is exactly why you want this kernel patch at all. And the kernel can (in this case) accommodate your buggy assembler, sure, but are you going to "fix" all other programs with this "problem" as well? Segher
Re: [PATCH v2] powerpc/boot: Delete unneeded .globl _zimage_start
On Wed, Mar 25, 2020 at 09:42:57AM -0700, Fangrui Song wrote: > .globl sets the symbol binding to STB_GLOBAL while .weak sets the > binding to STB_WEAK. GNU as let .weak override .globl since binutils-gdb > 5ca547dc2399a0a5d9f20626d4bf5547c3ccfddd (1996). Clang integrated > assembler let the last win but it may error in the future. GNU AS works for more than just ELF. The way the assembler language is defined, it is not .weak overriding .globl -- instead, .weak sets a symbol attribute. On an existing symbol (but it creates on if there is none yet). Clang is buggy if it does not allow valid (and perfectly normal) assembler code like this. Segher
