linuxppc-dev.lists.ozlabs.org archive mirror
 help / color / mirror / Atom feed
From: Michael Neuling <mikey@neuling.org>
To: Nicholas Piggin <npiggin@gmail.com>, linuxppc-dev@lists.ozlabs.org
Cc: Sam Bobroff <sam.bobroff@au1.ibm.com>,
	Michael Ellerman <mpe@ellerman.id.au>
Subject: Re: [PATCH] powerpc/tm: do not use r13 for tabort_syscall
Date: Mon, 22 Aug 2016 12:09:45 +1000	[thread overview]
Message-ID: <1471831785.14506.36.camel@neuling.org> (raw)
In-Reply-To: <1469172468-12892-1-git-send-email-npiggin@gmail.com>

On Fri, 2016-07-22 at 17:27 +1000, Nicholas Piggin wrote:
> tabort_syscall runs with RI=3D1, so a nested recoverable machine
> check will load the paca into r13 and overwrite what we loaded
> it with, because exceptions returning to privileged mode do not
> restore r13.
>=20
> This has survived testing with sc instruction inside transaction
> (bare sc, not glibc syscall because glibc can tabort before sc).
> Verified the transaction is failing failing with with
> TM_CAUSE_SYSCALL.
>=20
> Signed-off-by: Nick Piggin <npiggin@gmail.com>
> Cc: Michael Neuling <mikey@neuling.org>

FWIW

Acked-by: Michael Neuling <mikey@neuling.org>

> Cc: Sam Bobroff <sam.bobroff@au1.ibm.com>
> Cc: Michael Ellerman <mpe@ellerman.id.au>
>=20
> ---
>=20
> =C2=A0arch/powerpc/kernel/entry_64.S | 20 ++++++++++----------
> =C2=A01 file changed, 10 insertions(+), 10 deletions(-)
>=20
> diff --git a/arch/powerpc/kernel/entry_64.S
> b/arch/powerpc/kernel/entry_64.S
> index 73e461a..387dee3 100644
> --- a/arch/powerpc/kernel/entry_64.S
> +++ b/arch/powerpc/kernel/entry_64.S
> @@ -368,13 +368,13 @@ END_FTR_SECTION_IFSET(CPU_FTR_HAS_PPR)
> =C2=A0tabort_syscall:
> =C2=A0	/* Firstly we need to enable TM in the kernel */
> =C2=A0	mfmsr	r10
> -	li	r13, 1
> -	rldimi	r10, r13, MSR_TM_LG, 63-MSR_TM_LG
> -	mtmsrd	r10, 0
> +	li	r9,1
> +	rldimi	r10,r9,MSR_TM_LG,63-MSR_TM_LG
> +	mtmsrd	r10,0
> =C2=A0
> =C2=A0	/* tabort, this dooms the transaction, nothing else */
> -	li	r13, (TM_CAUSE_SYSCALL|TM_CAUSE_PERSISTENT)
> -	TABORT(R13)
> +	li	r9,(TM_CAUSE_SYSCALL|TM_CAUSE_PERSISTENT)
> +	TABORT(R9)
> =C2=A0
> =C2=A0	/*
> =C2=A0	=C2=A0* Return directly to userspace. We have corrupted user regis=
ter
> state,
> @@ -382,11 +382,11 @@ tabort_syscall:
> =C2=A0	=C2=A0* resume after the tbegin of the aborted transaction with th=
e
> =C2=A0	=C2=A0* checkpointed register state.
> =C2=A0	=C2=A0*/
> -	li	r13, MSR_RI
> -	andc	r10, r10, r13
> -	mtmsrd	r10, 1
> -	mtspr	SPRN_SRR0, r11
> -	mtspr	SPRN_SRR1, r12
> +	li	r9,MSR_RI
> +	andc	r10,r10,r9
> +	mtmsrd	r10,1
> +	mtspr	SPRN_SRR0,r11
> +	mtspr	SPRN_SRR1,r12
> =C2=A0
> =C2=A0	rfid
> =C2=A0	b	.	/* prevent speculative execution */

  parent reply	other threads:[~2016-08-22  2:09 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-07-22  7:27 [PATCH] powerpc/tm: do not use r13 for tabort_syscall Nicholas Piggin
2016-07-25  0:57 ` Michael Neuling
2016-08-22  2:09 ` Michael Neuling [this message]
  -- strict thread matches above, loose matches on Subject: below --
2016-07-25  4:26 Nicholas Piggin

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=1471831785.14506.36.camel@neuling.org \
    --to=mikey@neuling.org \
    --cc=linuxppc-dev@lists.ozlabs.org \
    --cc=mpe@ellerman.id.au \
    --cc=npiggin@gmail.com \
    --cc=sam.bobroff@au1.ibm.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 NNTP newsgroup(s).