Linux-EFI Archive on lore.kernel.org
 help / color / Atom feed
From: Ard Biesheuvel <ardb@kernel.org>
To: Atish Patra <Atish.Patra@wdc.com>
Cc: "linux-efi@vger.kernel.org" <linux-efi@vger.kernel.org>,
	"mingo@kernel.org" <mingo@kernel.org>,
	"nivedita@alum.mit.edu" <nivedita@alum.mit.edu>,
	"lukas@wunner.de" <lukas@wunner.de>
Subject: Re: [PATCH 03/19] efi/libstub: Use hidden visiblity for all source files
Date: Tue, 25 Feb 2020 00:36:45 +0100
Message-ID: <CAKv+Gu9bPsiJLQqsfXwDMNe-VCsVCqDNvZAiCnon18kYG4M9qA@mail.gmail.com> (raw)
In-Reply-To: <d3402c42836aca94d1a2ff385c088af605578068.camel@wdc.com>

On Tue, 25 Feb 2020 at 00:15, Atish Patra <Atish.Patra@wdc.com> wrote:
>
> On Mon, 2020-02-10 at 17:02 +0100, Ard Biesheuvel wrote:
> > Instead of setting the visibility pragma for a small set of symbol
> > declarations that could result in absolute references that we cannot
> > support in the stub, declare hidden visibility for all code in the
> > EFI stub, which is more robust and future proof.
> >
> > To ensure that the #pragma is taken into account before any other
> > includes are processed, put it in a header file of its own and
> > include it via the compiler command line using the -include option.
> >
> > Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
> > ---
> >  arch/arm64/include/asm/efi.h              | 3 ---
> >  drivers/firmware/efi/libstub/Makefile     | 2 +-
> >  drivers/firmware/efi/libstub/arm64-stub.c | 8 +-------
> >  drivers/firmware/efi/libstub/hidden.h     | 6 ++++++
> >  4 files changed, 8 insertions(+), 11 deletions(-)
> >
> > diff --git a/arch/arm64/include/asm/efi.h
> > b/arch/arm64/include/asm/efi.h
> > index 44531a69d32b..56ae87401a26 100644
> > --- a/arch/arm64/include/asm/efi.h
> > +++ b/arch/arm64/include/asm/efi.h
> > @@ -107,9 +107,6 @@ static inline void free_screen_info(struct
> > screen_info *si)
> >  {
> >  }
> >
> > -/* redeclare as 'hidden' so the compiler will generate relative
> > references */
> > -extern struct screen_info screen_info
> > __attribute__((__visibility__("hidden")));
> > -
> >  static inline void efifb_setup_from_dmi(struct screen_info *si,
> > const char *opt)
> >  {
> >  }
> > diff --git a/drivers/firmware/efi/libstub/Makefile
> > b/drivers/firmware/efi/libstub/Makefile
> > index f14b7636323a..4efdbd711e8e 100644
> > --- a/drivers/firmware/efi/libstub/Makefile
> > +++ b/drivers/firmware/efi/libstub/Makefile
> > @@ -25,7 +25,7 @@ cflags-$(CONFIG_ARM)                := $(subst
> > $(CC_FLAGS_FTRACE),,$(KBUILD_CFLAGS)) \
> >  cflags-$(CONFIG_EFI_ARMSTUB) += -I$(srctree)/scripts/dtc/libfdt
> >
> >  KBUILD_CFLAGS                        := $(cflags-y)
> > -DDISABLE_BRANCH_PROFILING \
> > -                                -D__NO_FORTIFY \
> > +                                -include hidden.h -D__NO_FORTIFY \
> >                                  $(call cc-option,-ffreestanding) \
> >                                  $(call cc-option,-fno-stack-
> > protector) \
> >                                  -D__DISABLE_EXPORTS
> > diff --git a/drivers/firmware/efi/libstub/arm64-stub.c
> > b/drivers/firmware/efi/libstub/arm64-stub.c
> > index 2915b44132e6..719d03a64329 100644
> > --- a/drivers/firmware/efi/libstub/arm64-stub.c
> > +++ b/drivers/firmware/efi/libstub/arm64-stub.c
> > @@ -6,17 +6,11 @@
> >   * Adapted from ARM version by Mark Salter <msalter@redhat.com>
> >   */
> >
> > -/*
> > - * To prevent the compiler from emitting GOT-indirected (and thus
> > absolute)
> > - * references to the section markers, override their visibility as
> > 'hidden'
> > - */
> > -#pragma GCC visibility push(hidden)
> > -#include <asm/sections.h>
> > -#pragma GCC visibility pop
> >
> >  #include <linux/efi.h>
> >  #include <asm/efi.h>
> >  #include <asm/memory.h>
> > +#include <asm/sections.h>
> >  #include <asm/sysreg.h>
> >
> >  #include "efistub.h"
> > diff --git a/drivers/firmware/efi/libstub/hidden.h
> > b/drivers/firmware/efi/libstub/hidden.h
> > new file mode 100644
> > index 000000000000..3493b041f419
> > --- /dev/null
> > +++ b/drivers/firmware/efi/libstub/hidden.h
> > @@ -0,0 +1,6 @@
> > +/* SPDX-License-Identifier: GPL-2.0 */
> > +/*
> > + * To prevent the compiler from emitting GOT-indirected (and thus
> > absolute)
> > + * references to any global symbols, override their visibility as
> > 'hidden'
> > + */
> > +#pragma GCC visibility push(hidden)
>
> I am getting following compiler errors because of this patch while
> cross compiling for ARM, ARM64 and RISC-V.
>
> cc1: fatal error: hidden.h: No such file or directory
>
> Adding the following line to cflags solves the issue which indicates
> that gcc is not looking at the source file directory for this
> particularly include file.
>
> -I$(srctree)/drivers/firmware/efi/libstub
>
> I might have some trick in my build setup. Just posting here to see if
> anybody else also seeing the same problem.
>

Thanks Atish. Heinrich reported the same issue. It didn't affect me or
any of the bots because it only hits if you build in-tree (i.e.,
$srctree == $objtree)

My latest efi/next branch carries a fix for this (similar to your
suggestion, but using -include)

  reply index

Thread overview: 29+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-02-10 16:02 [PATCH 00/19] EFI stub early spring cleaning part 2 Ard Biesheuvel
2020-02-10 16:02 ` [PATCH 01/19] efi/libstub/x86: Remove pointless zeroing of apm_bios_info Ard Biesheuvel
2020-02-10 16:02 ` [PATCH 02/19] efi/libstub/x86: Avoid overflowing code32_start on PE entry Ard Biesheuvel
2020-02-10 16:02 ` [PATCH 03/19] efi/libstub: Use hidden visiblity for all source files Ard Biesheuvel
2020-02-24 23:15   ` Atish Patra
2020-02-24 23:36     ` Ard Biesheuvel [this message]
2020-02-25 19:18       ` Atish Patra
2020-02-10 16:02 ` [PATCH 04/19] efi/libstub/arm: Relax FDT alignment requirement Ard Biesheuvel
2020-02-10 16:02 ` [PATCH 05/19] efi/libstub: Move memory map handling and allocation routines to mem.c Ard Biesheuvel
2020-02-10 16:02 ` [PATCH 06/19] efi/libstub: Simplify efi_high_alloc() and rename to efi_allocate_pages() Ard Biesheuvel
2020-02-10 16:02 ` [PATCH 07/19] efi/libstub/x86: Incorporate eboot.c into libstub Ard Biesheuvel
2020-02-10 16:02 ` [PATCH 08/19] efi/libstub: Use consistent type names for file I/O protocols Ard Biesheuvel
2020-02-10 16:02 ` [PATCH 09/19] efi/libstub: Move stub specific declarations into efistub.h Ard Biesheuvel
2020-02-10 16:02 ` [PATCH 10/19] efi/libstub/x86: Permit bootparams struct to be allocated above 4 GB Ard Biesheuvel
2020-02-10 16:02 ` [PATCH 11/19] efi/libstub/x86: Permit cmdline data " Ard Biesheuvel
2020-02-10 16:02 ` [PATCH 12/19] efi/libstub: Move efi_random_alloc() into separate source file Ard Biesheuvel
2020-02-10 16:02 ` [PATCH 13/19] efi/libstub: Move get_dram_base() into arm-stub.c Ard Biesheuvel
2020-02-17  1:17   ` Atish Patra
2020-02-17  8:37     ` Ard Biesheuvel
2020-02-26 23:34       ` Atish Patra
2020-02-27  7:38         ` Ard Biesheuvel
2020-02-27  7:48           ` Atish Patra
2020-02-27  7:50             ` Ard Biesheuvel
2020-02-10 16:02 ` [PATCH 14/19] efi/libstub: Move file I/O support code into separate file Ard Biesheuvel
2020-02-10 16:02 ` [PATCH 15/19] efi/libstub: Rewrite file I/O routine Ard Biesheuvel
2020-02-10 16:02 ` [PATCH 16/19] efi/libstub: Take soft and hard memory limits into account for initrd loading Ard Biesheuvel
2020-02-10 16:02 ` [PATCH 17/19] efi/libstub: Clean up command line parsing routine Ard Biesheuvel
2020-02-10 16:02 ` [PATCH 18/19] efi/libstub: Expose LocateDevicePath boot service Ard Biesheuvel
2020-02-10 16:02 ` [PATCH 19/19] efi/libstub: Make the LoadFile EFI protocol accessible Ard Biesheuvel

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=CAKv+Gu9bPsiJLQqsfXwDMNe-VCsVCqDNvZAiCnon18kYG4M9qA@mail.gmail.com \
    --to=ardb@kernel.org \
    --cc=Atish.Patra@wdc.com \
    --cc=linux-efi@vger.kernel.org \
    --cc=lukas@wunner.de \
    --cc=mingo@kernel.org \
    --cc=nivedita@alum.mit.edu \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

Linux-EFI Archive on lore.kernel.org

Archives are clonable:
	git clone --mirror https://lore.kernel.org/linux-efi/0 linux-efi/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 linux-efi linux-efi/ https://lore.kernel.org/linux-efi \
		linux-efi@vger.kernel.org
	public-inbox-index linux-efi

Example config snippet for mirrors

Newsgroup available over NNTP:
	nntp://nntp.lore.kernel.org/org.kernel.vger.linux-efi


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git