Commit e4412739472b ("Documentation: raise minimum supported version of
binutils to 2.25") allows us to remove the checks for old binutils.

There is no more user for ld-ifversion. Remove it as well.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Reviewed-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://msgid.link/20230119082250.151485-1-masahiroy@kernel.org

binutils v2.37 drops unused section symbols, which prevents recordmcount
from capturing mcount locations in sections that have no non-weak
symbols. This results in a build failure with a message such as:
	Cannot find symbol for section 12: .text.perf_callchain_kernel.
	kernel/events/callchain.o: failed

The change to binutils was reverted for v2.38, so this behavior is
specific to binutils v2.37:
https://sourceware.org/git/?p=binutils-gdb.git;a=commit;h=c09c8b42021180eee9495bd50d8b35e683d3901b

Objtool is able to cope with such sections, so this issue is specific to
recordmcount.

Fail the build and print a warning if binutils v2.37 is detected and if
we are using recordmcount.

Cc: stable@vger.kernel.org
Suggested-by: Joel Stanley <joel@jms.id.au>
Signed-off-by: Naveen N Rao <naveen@kernel.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://msgid.link/20230530061436.56925-1-naveen@kernel.org

x86 removed -pipe in commit 437e88ab8f9e2 ("x86/build: Remove -pipe from
KBUILD_CFLAGS") and the newer arm64 and riscv seem to have never used it,
so that seems to be the way the world's going.

Compile performance building defconfig on a POWER10 PowerNV system
was in the noise after 10 builds each. No point in adding options unless
they help something, so remove it.

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://msgid.link/20230606064830.184083-1-npiggin@gmail.com

Rather than trying to keep multiple configs up to date, make
pseries_defconfig an alias for ppc64le_guest_defconfig.

NOTE, pseries_defconfig was a big endian config, but this commit
switches it to little endian.

Almost all distros are ppc64le these days, so little endian is much more
likely to be what a user wants when they build for "pseries".

For an actual big endian guest, use ppc64_guest_defconfig.

Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://msgid.link/20230414132415.821564-32-mpe@ellerman.id.au

Rather than trying to keep multiple configs up to date, make
pseries_le_defconfig an alias for ppc64le_guest_defconfig.

ppc64le_guest_defconfig should work in all cases that
pseries_le_defconfig currently does, but if not we can update it.

Move pseries_le_defconfig down in the Makefile, so it appears after
ppc64le_guest_defconfig in the help output.

Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://msgid.link/20230414132415.821564-31-mpe@ellerman.id.au

Incorporate the generic kvm_guest.config into the powerpc guest configs,
ppc64[le]_guest_defconfig.

This brings in some useful options, in particular 9P support, and also
means future additions to the generic file will be automatically picked
up by the powerpc configs.

Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://msgid.link/20230414132415.821564-30-mpe@ellerman.id.au

Currently none of the generated defconfigs appear in the help output,
because the help text discovers defconfigs by looking for actual files
named "*_defconfig".

Collect the generated defconfig names into a variable and then print
those out in archhelp.

Output looks like eg:

  pseries_le_defconfig        - Build for pseries_le
  ppc64le_defconfig           - Build for ppc64le
  ppc64le_guest_defconfig     - Build for ppc64le_guest
  ...
  ppc64_randconfig            - Build for ppc64_randconfig

  adder875_defconfig          - Build for adder875
  amigaone_defconfig          - Build for amigaone

Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
[mpe: Fix PHONY bug which broke in-tree build, thanks rmclure]
Link: https://msgid.link/20230329072334.2023357-2-mpe@ellerman.id.au

It's not necessary to prefix every command in archhelp with "@" (to
suppress echoing the command), because that is done by the top level
Makefile when it evaluates archhelp.

Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://msgid.link/20230329072334.2023357-1-mpe@ellerman.id.au

Build modules using PCREL addressing when CONFIG_PPC_KERNEL_PCREL=y.

- The module loader must handle several new relocation types:

  * R_PPC64_REL24_NOTOC is a function call handled like R_PPC_REL24, but
    does not restore r2 upon return. The external function call stub is
    changed to use pcrel addressing to load the function pointer rather
    than based on the module TOC.

  * R_PPC64_GOT_PCREL34 is a reference to external data. A GOT table
    must be built by hand, because the linker adds this during the final
    link (which is not done for kernel modules). The GOT table is built
    similarly to the way the external function call stub table is. This
    section is called .mygot because .got has a special meaning for the
    linker and can become upset.

  * R_PPC64_PCREL34 is used for local data addressing, but there is a
    special case where the percpu section is moved at load-time to the
    percpu area which is out of range of this relocation. This requires
    the PCREL34 relocations are converted to use GOT_PCREL34 addressing.

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
[mpe: Some coding style & formatting fixups]
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://msgid.link/20230408021752.862660-7-npiggin@gmail.com

PC-Relative or PCREL addressing is an extension to the ELF ABI which
uses Power ISA v3.1 PC-relative instructions to calculate addresses,
rather than the traditional TOC scheme.

Add an option to build vmlinux using pcrel addressing. Modules continue
to use TOC addressing.

- TOC address helpers and r2 are poisoned with -1 when running vmlinux.
  r2 could be used for something useful once things are ironed out.

- Assembly must call C functions with @notoc annotation, or the linker
  complains aobut a missing nop after the call. This is done with the
  CFUNC macro introduced earlier.

- Boot: with the exception of prom_init, the execution branches to the
  kernel virtual address early in boot, before any addresses are
  generated, which ensures 34-bit pcrel addressing does not miss the
  high PAGE_OFFSET bits. TOC relative addressing has a similar
  requirement. prom_init does not go to the virtual address and its
  addresses should not carry over to the post-prom kernel.

- Ftrace trampolines are converted from TOC addressing to pcrel
  addressing, including module ftrace trampolines that currently use the
  kernel TOC to find ftrace target functions.

- BPF function prologue and function calling generation are converted
  from TOC to pcrel.

- copypage_64.S has an interesting problem, prefixed instructions have
  alignment restrictions so the linker can add padding, which makes the
  assembler treat the difference between two local labels as
  non-constant even if alignment is arranged so padding is not required.
  This may need toolchain help to solve nicely, for now move the prefix
  instruction out of the alternate patch section to work around it.

This reduces kernel text size by about 6%.

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://msgid.link/20230408021752.862660-6-npiggin@gmail.com