public inbox for devel@edk2.groups.io
 help / color / mirror / Atom feed
From: "Yao, Jiewen" <jiewen.yao@intel.com>
To: Laszlo Ersek <lersek@redhat.com>,
	"Vanguput, Narendra K" <narendra.k.vanguput@intel.com>,
	"edk2-devel@lists.01.org" <edk2-devel@lists.01.org>
Cc: "Dong, Eric" <eric.dong@intel.com>, "Ni, Ray" <ray.ni@intel.com>
Subject: Re: [PATCH v2] UefiCpuPkg\CpuSmm: Save & restore CR2 on-demand paging in SMM
Date: Thu, 7 Mar 2019 18:18:46 +0000	[thread overview]
Message-ID: <74D8A39837DF1E4DA445A8C0B3885C503F558C5B@shsmsx102.ccr.corp.intel.com> (raw)
In-Reply-To: <428b8706-59f8-fcfc-e06c-57ab4f13c328@redhat.com>

Good catch Laszo!!!

I found PcdCpuSmmStaticPageTable is [PcdsFixedAtBuild, PcdsPatchableInModule, PcdsDynamic, PcdsDynamicEx].
I think it should only be static, but I am wrong. Thanks to point it out.
Then I think we need get the PCD value at the entrypoint.


Another option is just to move the CR2 access from C code to ASM code, to isolate the CR2 access in C code.


Thank you
Yao Jiewen


> -----Original Message-----
> From: Laszlo Ersek [mailto:lersek@redhat.com]
> Sent: Thursday, March 7, 2019 9:58 AM
> To: Vanguput, Narendra K <narendra.k.vanguput@intel.com>;
> edk2-devel@lists.01.org
> Cc: Dong, Eric <eric.dong@intel.com>; Ni, Ray <ray.ni@intel.com>; Yao,
> Jiewen <jiewen.yao@intel.com>
> Subject: Re: [PATCH v2] UefiCpuPkg\CpuSmm: Save & restore CR2
> on-demand paging in SMM
> 
> On 03/07/19 12:14, nkvangup wrote:
> > BZ: https://bugzilla.tianocore.org/show_bug.cgi?id=1593
> >
> > For every SMI occurrence, save and restore CR2 register only when SMM
> > on-demand paging support is enabled in 64 bit operation mode.
> >
> > Contributed-under: TianoCore Contribution Agreement 1.1
> > Signed-off-by: Vanguput Narendra K <narendra.k.vanguput@intel.com>
> > Cc: Eric Dong <eric.dong@intel.com>
> > Cc: Ray Ni <ray.ni@intel.com>
> > Cc: Laszlo Ersek <lersek@redhat.com>
> > Cc: Yao Jiewen <jiewen.yao@intel.com>
> > ---
> >  UefiCpuPkg/PiSmmCpuDxeSmm/MpService.c | 20 ++++++++++++--------
> >  1 file changed, 12 insertions(+), 8 deletions(-)
> 
> (1) There is an open question about the usefulness of this patch in
> <https://bugzilla.tianocore.org/show_bug.cgi?id=1593#c1>. It should be
> answered in the BZ, or the same description should be included in the
> commit message.
> 
> (2) Also, the commit message should refer to the BZ.
> 
> 
> > diff --git a/UefiCpuPkg/PiSmmCpuDxeSmm/MpService.c
> b/UefiCpuPkg/PiSmmCpuDxeSmm/MpService.c
> > index 3b0b3b52ac..5be4a2b020 100644
> > --- a/UefiCpuPkg/PiSmmCpuDxeSmm/MpService.c
> > +++ b/UefiCpuPkg/PiSmmCpuDxeSmm/MpService.c
> > @@ -1111,10 +1111,12 @@ SmiRendezvous (
> >
> >    ASSERT(CpuIndex < mMaxNumberOfCpus);
> >
> > -  //
> > -  // Save Cr2 because Page Fault exception in SMM may override its value
> > -  //
> > -  Cr2 = AsmReadCr2 ();
> > +  if ((sizeof (UINTN) == sizeof (UINT64)) && (!PcdGetBool
> (PcdCpuSmmStaticPageTable))) {
> 
> (3) It doesn't look like a good idea to me to call PcdGetBool() in the
> SmiRendezvous() function.
> 
> If the PCD is not fixed-at-build (but dynamic), then we'll end up
> calling a PI protocol member from a function that is by definition
> executed by multiple processors at the same time.
> 
> "X64/PageTbl.c" already defines the global variable
> "mCpuSmmStaticPageTable", setting it from the PCD on the call stack of
> the entry point function of the driver. That is safe -- we can call PI /
> UEFI protocols in the entry point functions of a DXE_SMM_DRIVER.
> 
> Now, the fact that "mCpuSmmStaticPageTable" is only defined in the X64
> build (that is, in "X64/PageTbl.c"), is actually quite informative. It
> means that, instead of this conditional code in "MpService.c", we should
> introduce two new helper functions, "SaveCr2" and "RestoreCr2". And we
> should provide separate implementations for IA32 and X64. For IA32, the
> function should do nothing. For X64, the function should depend on
> "mCpuSmmStaticPageTable", and massage CR2 as necessary.
> 
> However: that *still* depends on whether this change is useful. I
> realize the CR2 manipulation may not be overly useful on IA32 (we can't
> address >4GB memory, so demand paging for >4GB makes no sense), but its
> performance hit should be negligible. Again, back to point (1): what is
> the actual issue with the current code?
> 
> Thanks
> Laszlo
> 
> > +    //
> > +    // Save Cr2 because Page Fault exception in SMM may override its
> value
> > +    //
> > +    Cr2 = AsmReadCr2 ();
> > +  }
> >
> >    //
> >    // Perform CPU specific entry hooks
> > @@ -1253,10 +1255,12 @@ SmiRendezvous (
> >
> >  Exit:
> >    SmmCpuFeaturesRendezvousExit (CpuIndex);
> > -  //
> > -  // Restore Cr2
> > -  //
> > -  AsmWriteCr2 (Cr2);
> > +  if ((sizeof (UINTN) == sizeof (UINT64)) && (!PcdGetBool
> (PcdCpuSmmStaticPageTable))) {
> > +    //
> > +    // Restore Cr2
> > +    //
> > +    AsmWriteCr2 (Cr2);
> > +  }
> >  }
> >
> >  /**
> >


      parent reply	other threads:[~2019-03-07 18:18 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-03-07 11:14 [PATCH v2] UefiCpuPkg\CpuSmm: Save & restore CR2 on-demand paging in SMM nkvangup
2019-03-07 14:38 ` Yao, Jiewen
2019-03-07 17:57 ` Laszlo Ersek
2019-03-07 18:10   ` Kinney, Michael D
2019-03-07 18:24     ` Kinney, Michael D
2019-03-07 18:18   ` Yao, Jiewen [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-list 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=74D8A39837DF1E4DA445A8C0B3885C503F558C5B@shsmsx102.ccr.corp.intel.com \
    --to=devel@edk2.groups.io \
    /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