LLD for x86_64

[tpimh]

Currently I have a patch which allows linking x86_64 kernel with LLD: fulldiff.patch.txt
Unfortunately there is no conditional detection of linker features, unsupported flags are just removed. This patch needs to be improved.

[nathanchance]

If we strip that patch down to just the LLD bits, we could theoretically add this to the script to only apply that patch when building with LLD.

  for folder in ../patches/${ARCH} ../patches/${ARCH}/${LD}; done
    [[ -d ${folder} ]] && git apply -3 "${folder}"/*.patch
  done

[tpimh]

These are the LLD specific parts of the patch:

diff --git a/arch/x86/boot/Makefile b/arch/x86/boot/Makefile
index 9b5adae..e2839b5 100644
--- a/arch/x86/boot/Makefile
+++ b/arch/x86/boot/Makefile
@@ -100,7 +100,7 @@ $(obj)/zoffset.h: $(obj)/compressed/vmlinux FORCE
 AFLAGS_header.o += -I$(objtree)/$(obj)
 $(obj)/header.o: $(obj)/zoffset.h
 
-LDFLAGS_setup.elf	:= -T
+LDFLAGS_setup.elf	:= -m elf_i386 -T
 $(obj)/setup.elf: $(src)/setup.ld $(SETUP_OBJS) FORCE
 	$(call if_changed,ld)
 
diff --git a/arch/x86/entry/vdso/Makefile b/arch/x86/entry/vdso/Makefile
index 141d415..0624bf2 100644
--- a/arch/x86/entry/vdso/Makefile
+++ b/arch/x86/entry/vdso/Makefile
@@ -47,7 +47,7 @@ targets += $(vdso_img_sodbg) $(vdso_img-y:%=vdso%.so)
 CPPFLAGS_vdso.lds += -P -C
 
 VDSO_LDFLAGS_vdso.lds = -m elf_x86_64 -soname linux-vdso.so.1 --no-undefined \
-			-z max-page-size=4096 -z common-page-size=4096
+			-z max-page-size=4096
 
 $(obj)/vdso64.so.dbg: $(obj)/vdso.lds $(vobjs) FORCE
 	$(call if_changed,vdso)
@@ -98,7 +98,7 @@ CFLAGS_REMOVE_vvar.o = -pg
 
 CPPFLAGS_vdsox32.lds = $(CPPFLAGS_vdso.lds)
 VDSO_LDFLAGS_vdsox32.lds = -m elf32_x86_64 -soname linux-vdso.so.1 \
-			   -z max-page-size=4096 -z common-page-size=4096
+			   -z max-page-size=4096
 
 # x32-rebranded versions
 vobjx32s-y := $(vobjs-y:.o=-x32.o)
 diff --git a/arch/x86/kernel/vmlinux.lds.S b/arch/x86/kernel/vmlinux.lds.S
 index 0d618ee..ee3b5c7 100644
 --- a/arch/x86/kernel/vmlinux.lds.S
 +++ b/arch/x86/kernel/vmlinux.lds.S
 @@ -401,7 +401,7 @@ SECTIONS
   * Per-cpu symbols which need to be offset from __per_cpu_load
   * for the boot processor.
   */
 -#define INIT_PER_CPU(x) init_per_cpu__##x = x + __per_cpu_load
 +#define INIT_PER_CPU(x) init_per_cpu__##x = ABSOLUTE(x) + __per_cpu_load
  INIT_PER_CPU(gdt_page);
  INIT_PER_CPU(irq_stack_union);
  
 diff --git a/arch/x86/purgatory/Makefile b/arch/x86/purgatory/Makefile
 index 3cf302b..6c167b9 100644
 --- a/arch/x86/purgatory/Makefile
 +++ b/arch/x86/purgatory/Makefile
 @@ -9,7 +9,7 @@ PURGATORY_OBJS = $(addprefix $(obj)/,$(purgatory-y))
  $(obj)/sha256.o: $(srctree)/lib/sha256.c FORCE
  	$(call if_changed_rule,cc_o_c)
  
 -LDFLAGS_purgatory.ro := -e purgatory_start -r --no-undefined -nostdlib -z nodefaultlib
 +LDFLAGS_purgatory.ro := -e purgatory_start -r --no-undefined -nostdlib
  targets += purgatory.ro
  
  KASAN_SANITIZE	:= n
 diff --git a/arch/x86/realmode/rm/Makefile b/arch/x86/realmode/rm/Makefile
 index 4463fa7..96cb20d 100644
 --- a/arch/x86/realmode/rm/Makefile
 +++ b/arch/x86/realmode/rm/Makefile
 @@ -47,7 +47,7 @@ $(obj)/pasyms.h: $(REALMODE_OBJS) FORCE
  targets += realmode.lds
  $(obj)/realmode.lds: $(obj)/pasyms.h
  
 -LDFLAGS_realmode.elf := --emit-relocs -T
 +LDFLAGS_realmode.elf := -m elf_i386 --emit-relocs -T
  CPPFLAGS_realmode.lds += -P -C -I$(objtree)/$(obj)
  
  targets += realmode.elf

[arndb]

On Tue, Nov 6, 2018 at 8:30 AM Twisted Pair in my Hair ***@***.***> wrote: diff --git a/arch/x86/entry/vdso/Makefile b/arch/x86/entry/vdso/Makefile index 141d415..0624bf2 100644 --- a/arch/x86/entry/vdso/Makefile +++ b/arch/x86/entry/vdso/Makefile @@ -47,7 +47,7 @@ targets += $(vdso_img_sodbg) $(vdso_img-y:%=vdso%.so) CPPFLAGS_vdso.lds += -P -C VDSO_LDFLAGS_vdso.lds = -m elf_x86_64 -soname linux-vdso.so.1 --no-undefined \ - -z max-page-size=4096 -z common-page-size=4096 + -z max-page-size=4096

Is it just that lld does not support that option, or does it do something wrong? It looks like this should be -z max-page-size=4096 $(call ld-option, -z common-page-size=4096) to make it use that specifically when the linker takes the option but not complain about it otherwise.

diff --git a/arch/x86/purgatory/Makefile b/arch/x86/purgatory/Makefile index 3cf302b..6c167b9 100644 --- a/arch/x86/purgatory/Makefile +++ b/arch/x86/purgatory/Makefile @@ -9,7 +9,7 @@ PURGATORY_OBJS = $(addprefix $(obj)/,$(purgatory-y)) $(obj)/sha256.o: $(srctree)/lib/sha256.c FORCE $(call if_changed_rule,cc_o_c) -LDFLAGS_purgatory.ro := -e purgatory_start -r --no-undefined -nostdlib -z nodefaultlib +LDFLAGS_purgatory.ro := -e purgatory_start -r --no-undefined -nostdlib

Same here.

--- a/arch/x86/kernel/vmlinux.lds.S +++ b/arch/x86/kernel/vmlinux.lds.S @@ -401,7 +401,7 @@ SECTIONS * Per-cpu symbols which need to be offset from __per_cpu_load * for the boot processor. */ -#define INIT_PER_CPU(x) init_per_cpu__##x = x + __per_cpu_load +#define INIT_PER_CPU(x) init_per_cpu__##x = ABSOLUTE(x) + __per_cpu_load INIT_PER_CPU(gdt_page); INIT_PER_CPU(irq_stack_union);

Not sure about this one. Does this have any effect on gnu ld? The rest looks like it can just be applied. Arnd

[tpimh]

Hi, Arnd!

Here are the related issues:

• elf_i386: -m linker option has higher priority than OUTPUT_FORMAT directive linux#30
• common-page-size: LLD doesn't support common-page-size=value keyword linux#31 (same feature detection was suggested by Nick there)
• nodefaultlib: LLD doesn't support -z nodefaultlib flag linux#94
• vmlinux.lds.S: Malformed linker script: vmlinux.lds linux#29 (doesn't break ld.bfd builds, probably should be submitted to upstream)