linuxppc-dev.lists.ozlabs.org archive mirror
 help / color / mirror / Atom feed
From: Nicholas Piggin <npiggin@gmail.com>
To: Michael Ellerman <mpe@ellerman.id.au>
Cc: linuxppc-dev@lists.ozlabs.org,
	"Aneesh Kumar K . V" <aneesh.kumar@linux.vnet.ibm.com>,
	Benjamin Herrenschmidt <benh@kernel.crashing.org>,
	Paul Mackerras <paulus@ozlabs.org>,
	Michael Neuling <mikey@neuling.org>
Subject: Re: [PATCH v2] powerpc/64s: ISAv3 initialize MMU registers before setting partition table
Date: Tue, 5 Dec 2017 15:53:17 +1000	[thread overview]
Message-ID: <20171205155317.6df2a891@roar.ozlabs.ibm.com> (raw)
In-Reply-To: <87d13tswmt.fsf@concordia.ellerman.id.au>

On Tue, 05 Dec 2017 14:04:42 +1100
Michael Ellerman <mpe@ellerman.id.au> wrote:

> Hi Nick,
> 
> Sorry I didn't reply sooner.
> 
> Nicholas Piggin <npiggin@gmail.com> writes:
> 
> > kexec can leave MMU registers set when booting into a new kernel, PIDR
> > in particular. The boot sequence does not zero PIDR, so it only gets
> > set when CPUs first switch to a userspace processes (until then it's
> > running a kernel thread with effective PID = 0).
> >
> > This leaves a window where a process table entry and page tables are
> > set up due to user processes running on other CPUs, that happen to
> > match with a stale PID. The CPU with that PID may cause speculative
> > accesses that address quadrant 0, which will result in cached
> > translations and PWC for that process, on a CPU which is not in the
> > mm_cpumask and so they will not get invalidated properly.
> >
> > The most common result is the kernel hanging in infinite page fault
> > loops soon after kexec (usually in schedule_tail, which is usually the
> > first non-speculative quardant 0 access to a new PID) due to a stale
> > PWC. However being a stale translation erorr, it could result in
> > anything up to security and data corruption errors.
> >
> > Fix this by zeroing out PIDR before setting PTCR.
> >
> > LPIDR is also not initialized, and may cause a similar issue with
> > speculative access to quadrant 1/2. This has not been observed, but
> > LPIDR is cleared to prevent that possibility.  
> 
> Isn't LPID initialised in __setup_cpu_power9() and __restore_cpu_power9() ?
> 
> eg:
> 
> _GLOBAL(__setup_cpu_power9)
> 	mflr	r11
> 	bl	__init_FSCR
> 	bl	__init_PMU
> 	bl	__init_hvmode_206
> 	mtlr	r11
> 	beqlr
> 	li	r0,0
> 	mtspr	SPRN_PSSCR,r0
> 	mtspr	SPRN_LPID,r0
> 
> 
> Similarly, shouldn't we be doing the PID initialisation there as well?

Hmm, yes I must have missed that! Yes that would be the best place to
put it.

Thanks,
Nick

      reply	other threads:[~2017-12-05  5:53 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-12-04  2:40 [PATCH v2] powerpc/64s: ISAv3 initialize MMU registers before setting partition table Nicholas Piggin
2017-12-05  3:04 ` Michael Ellerman
2017-12-05  5:53   ` Nicholas Piggin [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=20171205155317.6df2a891@roar.ozlabs.ibm.com \
    --to=npiggin@gmail.com \
    --cc=aneesh.kumar@linux.vnet.ibm.com \
    --cc=benh@kernel.crashing.org \
    --cc=linuxppc-dev@lists.ozlabs.org \
    --cc=mikey@neuling.org \
    --cc=mpe@ellerman.id.au \
    --cc=paulus@ozlabs.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;
as well as URLs for NNTP newsgroup(s).