From: Richard Earnshaw <Richard.Earnshaw@foss.arm.com>
To: "Victor L. Do Nascimento" <victor.donascimento@arm.com>,
newlib@sourceware.org
Subject: Re: [PATCH v5a 8/8] newlib: libc: setjmp M-profile PACBTI-enablement
Date: Tue, 10 Jan 2023 11:24:25 +0000 [thread overview]
Message-ID: <21874616-3f79-dfe5-6b14-283117bd9b2d@foss.arm.com> (raw)
In-Reply-To: <yw8jcz7n5ww7.fsf@arm.com>
On 09/01/2023 19:30, Victor L. Do Nascimento wrote:
> This patch revision provides the following improvements to proposed
> patch implementation:
> * Load ip directly when restoring PAC code from stack for
> authentication in longjmp epilogue.
> * Improve code readability by defining the REGS_SAVE_SIZE macro to
> map onto the otherwise obscure 104 constant used in str/ldr
> instructions.
>
> Add function prologue/epilogue to conditionally add BTI landing pads
> and/or PAC code generation & authentication instructions depending on
> compilation flags.
> ---
> newlib/libc/machine/arm/setjmp.S | 43 ++++++++++++++++++++++++++++++++
> 1 file changed, 43 insertions(+)
>
> diff --git a/newlib/libc/machine/arm/setjmp.S b/newlib/libc/machine/arm/setjmp.S
> index d814afea8..820230458 100644
> --- a/newlib/libc/machine/arm/setjmp.S
> +++ b/newlib/libc/machine/arm/setjmp.S
> @@ -22,6 +22,11 @@
> #define SIZE(x)
> #endif
>
> +/* Jump buffer allocation sizes. */
> +#define CORE_REGS_SAVE_SIZE 40
> +#define FP_REGS_SAVE_SIZE 64
> +#define REGS_SAVE_SIZE (CORE_REGS_SAVE_SIZE + FP_REGS_SAVE_SIZE)
Nearly there, but I think the names could be a bit clearer. I'd suggest:
JMPBUF_CORE_REGS_SIZE
JMPBUF_FP_REGS_SIZE
JMPBUF_PAC
I'd also write the values as the number of regs saved multiplied by the
size of a register (in bytes), so (10 * 4) and (8 * 8) for the sizes and
finally write the last one as
(JMPBUF_CORE_REGS_SIZE + JMPBUF_FP_REGS_SIZE + 0)
the "+ 0" to make it clear that it's after the other registers; if we
ever need to add more registers in future they can then follow the same
style.
R.
> +
> /* Arm/Thumb interworking support:
>
> The interworking scheme expects functions to use a BX instruction
> @@ -155,6 +160,8 @@ SYM (.arm_start_of.\name):
> .align 2
> MODE
> .globl SYM (\name)
> + .fnstart
> + .cfi_startproc
> TYPE (\name)
> SYM (\name):
> PROLOGUE \name
> @@ -162,6 +169,8 @@ SYM (\name):
>
> .macro FUNC_END name
> RET
> + .cfi_endproc
> + .fnend
> SIZE (\name)
> .endm
>
> @@ -171,6 +180,21 @@ SYM (\name):
>
> FUNC_START setjmp
>
> +#if __ARM_FEATURE_PAC_DEFAULT
> +# if __ARM_FEATURE_BTI_DEFAULT
> + pacbti ip, lr, sp
> +# else
> + pac ip, lr, sp
> +# endif /* __ARM_FEATURE_BTI_DEFAULT */
> + mov r3, ip
> + str r3, [r0, #REGS_SAVE_SIZE]
> + .cfi_register 143, 12
> +#else
> +# if __ARM_FEATURE_BTI_DEFAULT
> + bti
> +# endif /* __ARM_FEATURE_BTI_DEFAULT */
> +#endif /* __ARM_FEATURE_PAC_DEFAULT */
> +
> /* Save all the callee-preserved registers into the jump buffer. */
> #ifdef __thumb2__
> mov ip, sp
> @@ -184,6 +208,10 @@ SYM (\name):
>
> /* When setting up the jump buffer return 0. */
> mov r0, #0
> +#if __ARM_FEATURE_PAC_DEFAULT
> + mov ip, r3
> + aut ip, lr, sp
> +#endif /* __ARM_FEATURE_PAC_DEFAULT */
>
> FUNC_END setjmp
>
> @@ -193,6 +221,16 @@ SYM (\name):
>
> FUNC_START longjmp
>
> +#if __ARM_FEATURE_BTI_DEFAULT
> + bti
> +#endif /* __ARM_FEATURE_BTI_DEFAULT */
> +
> +#if __ARM_FEATURE_PAC_DEFAULT
> + /* Keep original jmpbuf address for retrieving pac-code
> + for authentication. */
> + mov r2, r0
> +#endif /* __ARM_FEATURE_PAC_DEFAULT */
> +
> /* If we have stack extension code it ought to be handled here. */
>
> /* Restore the registers, retrieving the state when setjmp() was called. */
> @@ -212,5 +250,10 @@ SYM (\name):
> it eq
> moveq r0, #1
>
> +#if __ARM_FEATURE_PAC_DEFAULT
> + ldr ip, [r2, #REGS_SAVE_SIZE]
> + aut ip, lr, sp
> +#endif /* __ARM_FEATURE_PAC_DEFAULT */
> +
> FUNC_END longjmp
> #endif
prev parent reply other threads:[~2023-01-10 11:24 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-01-09 19:30 Victor L. Do Nascimento
2023-01-10 11:24 ` Richard Earnshaw [this message]
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=21874616-3f79-dfe5-6b14-283117bd9b2d@foss.arm.com \
--to=richard.earnshaw@foss.arm.com \
--cc=newlib@sourceware.org \
--cc=victor.donascimento@arm.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).