public inbox for linux-arm-kernel@lists.infradead.org
 help / color / mirror / Atom feed
From: Mark Rutland <mark.rutland@arm.com>
To: "崇文 李" <chongwen.li@outlook.com>
Cc: "linux-arm-kernel@lists.infradead.org"
	<linux-arm-kernel@lists.infradead.org>
Subject: Re: Potential bug of arm64 SMCCC call
Date: Mon, 25 Mar 2024 17:00:47 +0000	[thread overview]
Message-ID: <ZgGtv2fgn05dLMWq@FVFF77S0Q05N> (raw)
In-Reply-To: <DS7PR19MB4407C29AC8FF38BA192DCDFFE6362@DS7PR19MB4407.namprd19.prod.outlook.com>

On Mon, Mar 25, 2024 at 01:36:37PM +0000, 崇文 李 wrote:
> Hello. I saw a potential bug when I was checking arm64 SMCCC call.
 
Hi,

> In arm-smccc.h [see link 0], from line 481 to line 497 resides the
> definition of "__arm_smccc_1_1", which is in turn used to implement
> the RMI interface in rmi_cmds.h [see link 1].
> 
> In the definition of "__arm_smccc_1_1", there is no clobbering of
> registers x4 -- x17.  However, according to Arm SMC Calling Convention
> [see link 2, page 10 of 19], registers x4 -- x17 are scratch registers. 
> Upon returning from an arm64 smc instruction, x4 -- x17 registers may 
> be overwritten with mysterious values from the firmware.

I think you're looking at a very old version of the SMCCC specification. The
__arm_smccc_1_1 helper is only intended to be used for SMCCCv1.1+, which is
newer than the specification that you link to below.

Please see a more recent version, e.g. ARM DEN 0028F: 

  https://developer.arm.com/documentation/den0028/f/?lang=en

If you look in "13 Appendix F: Changelog", you'll see the key change in
SMCCCv1.1:

| Result register set:
| Mandates preservation of registers X4—X17 across a SMC or HVC call.

... and this is described earlier in the document.

> So it is safer to mark x4--x17 registers as clobbered registers in the
> "asm" statement inside the definition of "__arm_smccc_1_1", so the
> compiler avoids using x4--x17 registers around arm64 smc instructions.

As above, this should not be necessary. These SMCCC invocation functions should
only be used with SMCCCv1.1+ compliant firmware, and in those cases x4-x17 are
preserved.

Note that specific SMCCC calls imply that the SMCCC version is 1.1 or later.

Thanks,
Mark

> 
> Best regards.
> 
> 
> 
> Links
> ============
> 
> [0] arm-smccc.h in arm's linux kernel repository
>    https://gitlab.arm.com/linux-arm/linux-cca/-/blob/cca-host/rmm-v1.0-eac5/include/linux/arm-smccc.h?ref_type=heads
> 
> [1] rmi_cmds.h in arm's linux kernel repository
>    https://gitlab.arm.com/linux-arm/linux-cca/-/blob/cca-host/rmm-v1.0-eac5/arch/arm64/include/asm/rmi_cmds.h?ref_type=heads
> 
> [2] Arm SMC Calling Convention
>    https://documentation-service.arm.com/static/5f8ea482f86e16515cdbe3c6?token=
> _______________________________________________
> linux-arm-kernel mailing list
> linux-arm-kernel@lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

      reply	other threads:[~2024-03-25 17:01 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-03-25 13:36 Potential bug of arm64 SMCCC call 崇文 李
2024-03-25 17:00 ` Mark Rutland [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=ZgGtv2fgn05dLMWq@FVFF77S0Q05N \
    --to=mark.rutland@arm.com \
    --cc=chongwen.li@outlook.com \
    --cc=linux-arm-kernel@lists.infradead.org \
    /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