From: "H.J. Lu" <hjl.tools@gmail.com>
To: binutils@sourceware.org, Cary Coutant <ccoutant@gmail.com>
Subject: Re: [PATCH] gold: Remove BND from 64-bit x86-64 IBT PLT
Date: Thu, 8 Dec 2022 08:12:50 -0800 [thread overview]
Message-ID: <CAMe9rOoZf0BcAYCs=+fB-YvNJrGY+E_c1TL2gvCfu7ZOueV_2Q@mail.gmail.com> (raw)
In-Reply-To: <20221206214444.799449-1-hjl.tools@gmail.com>
On Tue, Dec 6, 2022 at 1:44 PM H.J. Lu <hjl.tools@gmail.com> wrote:
>
> Since MPX support has been removed from x86-64 psABI, remove BND from
> 64-bit IBT PLT by using 32-bit IBT PLT.
>
> PR gold/29851
> * x86_64.cc (Output_data_plt_x86_64_ibt<32>::first_plt_entry):
> Renamed to ...
> (Output_data_plt_x86_64_ibt<size>::first_plt_entry): This.
> (Output_data_plt_x86_64_ibt<64>::first_plt_entry): Removed.
> (Output_data_plt_x86_64_ibt<size>::do_fill_first_plt_entry):
> Drop the size == 32 check.
> (Output_data_plt_x86_64_ibt<32>::plt_entry): Renamed to ...
> (Output_data_plt_x86_64_ibt<size>::plt_entry): This.
> (Output_data_plt_x86_64_ibt<64>::plt_entry): Removed.
> (Output_data_plt_x86_64_ibt<32>::aplt_entry): Renamed to ...
> (Output_data_plt_x86_64_ibt<size>::aplt_entry): This.
> (Output_data_plt_x86_64_ibt<64>::aplt_entry): Removed.
> (Output_data_plt_x86_64_ibt<size>::do_fill_plt_entry): Drop the
> size == 32 check.
> (Output_data_plt_x86_64_ibt<size>::fill_aplt_entry): Likewise.
> ---
> gold/x86_64.cc | 57 ++++++++------------------------------------------
> 1 file changed, 9 insertions(+), 48 deletions(-)
>
> diff --git a/gold/x86_64.cc b/gold/x86_64.cc
> index 86b232a5e2d..83684b90fcf 100644
> --- a/gold/x86_64.cc
> +++ b/gold/x86_64.cc
> @@ -2328,12 +2328,10 @@ Output_data_plt_x86_64_ibt<size>::set_final_data_size()
>
> // The first entry in the IBT PLT.
>
> -template<>
> +template<int size>
> const unsigned char
> -Output_data_plt_x86_64_ibt<32>::first_plt_entry[plt_entry_size] =
> +Output_data_plt_x86_64_ibt<size>::first_plt_entry[plt_entry_size] =
> {
> - // MPX isn't supported for x32, so we don't need the BND prefix.
> - // From AMD64 ABI Draft 0.98, page 76
> 0xff, 0x35, // pushq contents of memory address
> 0, 0, 0, 0, // replaced with address of .got + 8
> 0xff, 0x25, // jmp indirect
> @@ -2341,18 +2339,6 @@ Output_data_plt_x86_64_ibt<32>::first_plt_entry[plt_entry_size] =
> 0x90, 0x90, 0x90, 0x90 // noop (x4)
> };
>
> -template<>
> -const unsigned char
> -Output_data_plt_x86_64_ibt<64>::first_plt_entry[plt_entry_size] =
> -{
> - // Use the BND prefix so that IBT is compatible with MPX.
> - 0xff, 0x35, // pushq contents of memory address
> - 0, 0, 0, 0, // replaced with address of .got + 8
> - 0xf2, 0xff, 0x25, // bnd jmp indirect
> - 0, 0, 0, 0, // replaced with address of .got + 16
> - 0x0f, 0x1f, 0x00 // nop
> -};
> -
> template<int size>
> void
> Output_data_plt_x86_64_ibt<size>::do_fill_first_plt_entry(
> @@ -2362,7 +2348,7 @@ Output_data_plt_x86_64_ibt<size>::do_fill_first_plt_entry(
> {
> // Offsets to the addresses needing relocation.
> const unsigned int roff1 = 2;
> - const unsigned int roff2 = (size == 32) ? 8 : 9;
> + const unsigned int roff2 = 8;
>
> memcpy(pov, first_plt_entry, plt_entry_size);
> // We do a jmp relative to the PC at the end of this instruction.
> @@ -2376,9 +2362,9 @@ Output_data_plt_x86_64_ibt<size>::do_fill_first_plt_entry(
>
> // Subsequent entries in the IBT PLT.
>
> -template<>
> +template<int size>
> const unsigned char
> -Output_data_plt_x86_64_ibt<32>::plt_entry[plt_entry_size] =
> +Output_data_plt_x86_64_ibt<size>::plt_entry[plt_entry_size] =
> {
> // From AMD64 ABI Draft 1.0-rc1, Chapter 13.
> 0xf3, 0x0f, 0x1e, 0xfa, // endbr64
> @@ -2389,24 +2375,11 @@ Output_data_plt_x86_64_ibt<32>::plt_entry[plt_entry_size] =
> 0x90, 0x90 // nop
> };
>
> -template<>
> -const unsigned char
> -Output_data_plt_x86_64_ibt<64>::plt_entry[plt_entry_size] =
> -{
> - // From AMD64 ABI Draft 1.0-rc1, Chapter 13.
> - 0xf3, 0x0f, 0x1e, 0xfa, // endbr64
> - 0x68, // pushq immediate
> - 0, 0, 0, 0, // replaced with offset into relocation table
> - 0xf2, 0xe9, // bnd jmpq relative
> - 0, 0, 0, 0, // replaced with offset to start of .plt
> - 0x90 // nop
> -};
> -
> // Entries in the IBT Additional PLT.
>
> -template<>
> +template<int size>
> const unsigned char
> -Output_data_plt_x86_64_ibt<32>::aplt_entry[aplt_entry_size] =
> +Output_data_plt_x86_64_ibt<size>::aplt_entry[aplt_entry_size] =
> {
> // From AMD64 ABI Draft 1.0-rc1, Chapter 13.
> 0xf3, 0x0f, 0x1e, 0xfa, // endbr64
> @@ -2416,18 +2389,6 @@ Output_data_plt_x86_64_ibt<32>::aplt_entry[aplt_entry_size] =
> 0x90, 0x90 // nop
> };
>
> -template<>
> -const unsigned char
> -Output_data_plt_x86_64_ibt<64>::aplt_entry[aplt_entry_size] =
> -{
> - // From AMD64 ABI Draft 1.0-rc1, Chapter 13.
> - 0xf3, 0x0f, 0x1e, 0xfa, // endbr64
> - 0xf2, 0xff, 0x25, // bnd jmpq indirect
> - 0, 0, 0, 0, // replaced with address of symbol in .got
> - 0x0f, 0x1f, 0x04, 0x00, // nop
> - 0x90, // nop
> -};
> -
> template<int size>
> unsigned int
> Output_data_plt_x86_64_ibt<size>::do_fill_plt_entry(
> @@ -2440,7 +2401,7 @@ Output_data_plt_x86_64_ibt<size>::do_fill_plt_entry(
> {
> // Offsets to the addresses needing relocation.
> const unsigned int roff1 = 5;
> - const unsigned int roff2 = (size == 32) ? 10 : 11;
> + const unsigned int roff2 = 10;
>
> memcpy(pov, plt_entry, plt_entry_size);
> elfcpp::Swap_unaligned<32, false>::writeval(pov + roff1, plt_index);
> @@ -2459,7 +2420,7 @@ Output_data_plt_x86_64_ibt<size>::fill_aplt_entry(
> unsigned int plt_index)
> {
> // Offset to the address needing relocation.
> - const unsigned int roff = (size == 32) ? 6 : 7;
> + const unsigned int roff = 6;
>
> // Check PC-relative offset overflow in PLT entry.
> uint64_t plt_got_pcrel_offset = (got_address + got_offset
> --
> 2.38.1
>
Hi Cary,
Is it OK for gold?
Thanks.
--
H.J.
next prev parent reply other threads:[~2022-12-08 16:13 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-12-06 21:44 H.J. Lu
2022-12-08 8:21 ` [PATCH] ld, gold: remove support for -z bndplt (MPX prefix) Martin Liška
2022-12-13 0:02 ` Cary Coutant
2022-12-13 12:54 ` Martin Liška
2022-12-13 17:14 ` H.J. Lu
2022-12-08 16:12 ` H.J. Lu [this message]
2022-12-12 20:34 ` [PATCH] gold: Remove BND from 64-bit x86-64 IBT PLT H.J. Lu
2022-12-12 23:57 ` Cary Coutant
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='CAMe9rOoZf0BcAYCs=+fB-YvNJrGY+E_c1TL2gvCfu7ZOueV_2Q@mail.gmail.com' \
--to=hjl.tools@gmail.com \
--cc=binutils@sourceware.org \
--cc=ccoutant@gmail.com \
/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
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).