public inbox for devel@edk2.groups.io
 help / color / mirror / Atom feed
From: Ard Biesheuvel <ard.biesheuvel@linaro.org>
To: Laszlo Ersek <lersek@redhat.com>
Cc: "edk2-devel@lists.01.org" <edk2-devel@lists.01.org>,
	Shannon Zhao <zhaoshenglong@huawei.com>
Subject: Re: [PATCH] ArmVirtPkg/ArmVirtQemu: use non-accelerated CopyMem for VariableRuntimeDxe
Date: Wed, 15 Nov 2017 14:03:01 +0000	[thread overview]
Message-ID: <CAKv+Gu_hd=0oA7v8r6Yrh-C+3G0tVzfk=5D_eRAEvVAQnec+4Q@mail.gmail.com> (raw)
In-Reply-To: <a5a0c3c5-d2ee-ff77-2ef8-599b1b088d28@redhat.com>

On 15 November 2017 at 13:51, Laszlo Ersek <lersek@redhat.com> wrote:
> On 11/14/17 11:22, Ard Biesheuvel wrote:
>> The VariableRuntimeDxe driver may use CopyMem () on NOR flash regions,
>> assuming such regions always have full memory semantics. Given that
>> those regions cannot be mapped as ordinary memory on ARM (due to the
>> fact that the NOR flash requires device semantics while in write mode)
>> this prevents us from using BaseMemoryLibOptDxe in VariableRuntimeDxe,
>> since it may use unaligned accesses and/or DC ZVA instructions, both
>> of which are incompatible with mappings using device semantics.
>>
>> Note that there is no way we can work around this by changing the
>> mapping type between 'memory' and 'device' when switching from read to
>> write mode and back, because the runtime mapping is created by the OS,
>> and cannot be changed at will.
>>
>> So let's just switch to the unaccelerated version of BaseMemoryLib which
>> does not have the same problem.
>>
>> Contributed-under: TianoCore Contribution Agreement 1.1
>> Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
>> ---
>>  ArmVirtPkg/ArmVirtQemu.dsc       | 2 ++
>>  ArmVirtPkg/ArmVirtQemuKernel.dsc | 2 ++
>>  2 files changed, 4 insertions(+)
>>
>> diff --git a/ArmVirtPkg/ArmVirtQemu.dsc b/ArmVirtPkg/ArmVirtQemu.dsc
>> index 8a60b61f2aa6..7b220d6e3c31 100644
>> --- a/ArmVirtPkg/ArmVirtQemu.dsc
>> +++ b/ArmVirtPkg/ArmVirtQemu.dsc
>> @@ -261,6 +261,8 @@ [Components.common]
>>    MdeModulePkg/Universal/Variable/RuntimeDxe/VariableRuntimeDxe.inf {
>>      <LibraryClasses>
>>        NULL|MdeModulePkg/Library/VarCheckUefiLib/VarCheckUefiLib.inf
>> +      # don't use unaligned CopyMem () on the UEFI varstore NOR flash region
>> +      BaseMemoryLib|MdePkg/Library/BaseMemoryLib/BaseMemoryLib.inf
>>    }
>>  !if $(SECURE_BOOT_ENABLE) == TRUE
>>    MdeModulePkg/Universal/SecurityStubDxe/SecurityStubDxe.inf {
>> diff --git a/ArmVirtPkg/ArmVirtQemuKernel.dsc b/ArmVirtPkg/ArmVirtQemuKernel.dsc
>> index 9a31ec93ca06..7c032e1b07e0 100644
>> --- a/ArmVirtPkg/ArmVirtQemuKernel.dsc
>> +++ b/ArmVirtPkg/ArmVirtQemuKernel.dsc
>> @@ -252,6 +252,8 @@ [Components.common]
>>    MdeModulePkg/Universal/Variable/RuntimeDxe/VariableRuntimeDxe.inf {
>>      <LibraryClasses>
>>        NULL|MdeModulePkg/Library/VarCheckUefiLib/VarCheckUefiLib.inf
>> +      # don't use unaligned CopyMem () on the UEFI varstore NOR flash region
>> +      BaseMemoryLib|MdePkg/Library/BaseMemoryLib/BaseMemoryLib.inf
>>    }
>>  !if $(SECURE_BOOT_ENABLE) == TRUE
>>    MdeModulePkg/Universal/SecurityStubDxe/SecurityStubDxe.inf {
>>
>
> Reviewed-by: Laszlo Ersek <lersek@redhat.com>
>
> Given that we've never seen the symptom reported by Shannon, I must
> think Shannon is using some kind of new hardware. May I ask what
> hardware that is?
>

I think this is equally likely to occur in any KVM hardware
virtualization context. It is simply a regression after moving to the
OptDxe BaseMemoryLib flavor.


  reply	other threads:[~2017-11-15 13:58 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-11-14 10:22 [PATCH] ArmVirtPkg/ArmVirtQemu: use non-accelerated CopyMem for VariableRuntimeDxe Ard Biesheuvel
2017-11-15 13:51 ` Laszlo Ersek
2017-11-15 14:03   ` Ard Biesheuvel [this message]
2017-11-16  1:01     ` Shannon Zhao
2017-11-16  9:39       ` Ard Biesheuvel

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='CAKv+Gu_hd=0oA7v8r6Yrh-C+3G0tVzfk=5D_eRAEvVAQnec+4Q@mail.gmail.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