From: "Vitaly Cheptsov" <cheptsov@ispras.ru>
To: Leif Lindholm <leif@nuviainc.com>
Cc: devel@edk2.groups.io, "Jiewen Yao" <jiewen.yao@intel.com>,
"Eric Dong" <eric.dong@intel.com>,
"Michael Kinney" <michael.d.kinney@intel.com>,
"Jian J Wang" <jian.j.wang@intel.com>,
"Jeff Fan" <vanjeff_919@hotmail.com>,
"Mikhail Krichanov" <krichanov@ispras.ru>,
"Marvin Häuser" <mhaeuser@posteo.de>
Subject: Re: [edk2-devel] [PATCH] UefiCpuPkg: Fix CPU stack guard support by aligning GDT buffer
Date: Fri, 5 Nov 2021 22:37:23 +0300 [thread overview]
Message-ID: <081FC9DB-3E1C-4FAB-8F4F-CD48F85F1EC4@ispras.ru> (raw)
In-Reply-To: <20211105192821.s2itdxh5t6azp4z6@leviathan>
[-- Attachment #1: Type: text/plain, Size: 4231 bytes --]
Hi Leif,
I assume you mean the commit description, because the commit message is in the topic. I believe something like that would do:
CpuExceptionHandlerLib supplies misaligned GDT to the outer world
(e.g. ArchSetupExceptionStack) when PcdCpuStackGuard is enabled.
This happens because it uses an array of UINT8 for the mNewGdt
variable, which alignment is 1 byte versus required 8 bytes. As a result
ArchSetupExceptionStack always returns EFI_INVALID_PARAMETER in OVMF Ia32
with XCODE5 and CLANGPDB at least.
Fix this by allocating extra space in mNewGdt and then aligning the pointer
upwards.
Best wishes,
Vitaly
> On 5 Nov 2021, at 22:28, Leif Lindholm <leif@nuviainc.com> wrote:
>
> UefiCpuPkg maintainers - please respond.
>
> Meanwhile, Vitaly, could you please provide a commit message?
> The BZ link is needed, but it's not a substitute.
>
> /
> Leif
>
> On Mon, Sep 20, 2021 at 17:13:47 +0300, Vitaly Cheptsov wrote:
>> REF: https://bugzilla.tianocore.org/show_bug.cgi?id=3639
>>
>>
>>
>> Cc: Jiewen Yao <jiewen.yao@intel.com>
>>
>> Cc: Eric Dong <eric.dong@intel.com>
>>
>> Cc: Michael Kinney <michael.d.kinney@intel.com>
>>
>> Cc: Jian J Wang <jian.j.wang@intel.com>
>>
>> Cc: Jeff Fan <vanjeff_919@hotmail.com>
>>
>> Cc: Mikhail Krichanov <krichanov@ispras.ru>
>>
>> Cc: Marvin Häuser <mhaeuser@posteo.de>
>>
>> Signed-off-by: Vitaly Cheptsov <cheptsov@ispras.ru>
>>
>> ---
>>
>> .../Library/CpuExceptionHandlerLib/DxeException.c | 12 +++++++-----
>>
>> 1 file changed, 7 insertions(+), 5 deletions(-)
>>
>>
>>
>> diff --git a/UefiCpuPkg/Library/CpuExceptionHandlerLib/DxeException.c b/UefiCpuPkg/Library/CpuExceptionHandlerLib/DxeException.c
>>
>> index fd59f09ecd..12874811e1 100644
>>
>> --- a/UefiCpuPkg/Library/CpuExceptionHandlerLib/DxeException.c
>>
>> +++ b/UefiCpuPkg/Library/CpuExceptionHandlerLib/DxeException.c
>>
>> @@ -22,7 +22,7 @@ EXCEPTION_HANDLER_DATA mExceptionHandlerData;
>>
>>
>>
>> UINT8 mNewStack[CPU_STACK_SWITCH_EXCEPTION_NUMBER *
>>
>> CPU_KNOWN_GOOD_STACK_SIZE];
>>
>> -UINT8 mNewGdt[CPU_TSS_GDT_SIZE];
>>
>> +UINT8 mNewGdt[CPU_TSS_GDT_SIZE + IA32_GDT_ALIGNMENT];
>>
>>
>>
>> /**
>>
>> Common exception handler.
>>
>> @@ -238,6 +238,7 @@ InitializeCpuExceptionHandlersEx (
>>
>> CPU_EXCEPTION_INIT_DATA EssData;
>>
>> IA32_DESCRIPTOR Idtr;
>>
>> IA32_DESCRIPTOR Gdtr;
>>
>> + UINT8 *Gdt;
>>
>>
>>
>> //
>>
>> // To avoid repeat initialization of default handlers, the caller should pass
>>
>> @@ -259,6 +260,7 @@ InitializeCpuExceptionHandlersEx (
>>
>> if (PcdGetBool (PcdCpuStackGuard)) {
>>
>> if (InitData == NULL) {
>>
>> SetMem (mNewGdt, sizeof (mNewGdt), 0);
>>
>> + Gdt = ALIGN_POINTER (mNewGdt, IA32_GDT_ALIGNMENT);
>>
>>
>>
>> AsmReadIdtr (&Idtr);
>>
>> AsmReadGdtr (&Gdtr);
>>
>> @@ -270,11 +272,11 @@ InitializeCpuExceptionHandlersEx (
>>
>> EssData.X64.StackSwitchExceptionNumber = CPU_STACK_SWITCH_EXCEPTION_NUMBER;
>>
>> EssData.X64.IdtTable = (VOID *)Idtr.Base;
>>
>> EssData.X64.IdtTableSize = Idtr.Limit + 1;
>>
>> - EssData.X64.GdtTable = mNewGdt;
>>
>> - EssData.X64.GdtTableSize = sizeof (mNewGdt);
>>
>> - EssData.X64.ExceptionTssDesc = mNewGdt + Gdtr.Limit + 1;
>>
>> + EssData.X64.GdtTable = Gdt;
>>
>> + EssData.X64.GdtTableSize = CPU_TSS_GDT_SIZE;
>>
>> + EssData.X64.ExceptionTssDesc = Gdt + Gdtr.Limit + 1;
>>
>> EssData.X64.ExceptionTssDescSize = CPU_TSS_DESC_SIZE;
>>
>> - EssData.X64.ExceptionTss = mNewGdt + Gdtr.Limit + 1 + CPU_TSS_DESC_SIZE;
>>
>> + EssData.X64.ExceptionTss = Gdt + Gdtr.Limit + 1 + CPU_TSS_DESC_SIZE;
>>
>> EssData.X64.ExceptionTssSize = CPU_TSS_SIZE;
>>
>>
>>
>> InitData = &EssData;
>>
>> --
>>
>> 2.30.1 (Apple Git-130)
>>
>>
>>
>>
>>
>>
>>
>>
[-- Attachment #2: Message signed with OpenPGP --]
[-- Type: application/pgp-signature, Size: 833 bytes --]
next prev parent reply other threads:[~2021-11-05 19:37 UTC|newest]
Thread overview: 10+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-09-20 14:13 [PATCH] UefiCpuPkg: Fix CPU stack guard support by aligning GDT buffer Vitaly Cheptsov
2021-09-20 16:15 ` [edk2-devel] " Vitaly Cheptsov
2021-10-24 10:59 ` Vitaly Cheptsov
2021-11-05 19:28 ` Leif Lindholm
2021-11-05 19:37 ` Vitaly Cheptsov [this message]
2021-11-08 14:04 ` Leif Lindholm
2021-11-05 19:42 ` Michael D Kinney
2021-11-05 20:36 ` Vitaly Cheptsov
2021-11-05 21:39 ` Michael D Kinney
2021-11-06 4:42 ` Vitaly Cheptsov
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=081FC9DB-3E1C-4FAB-8F4F-CD48F85F1EC4@ispras.ru \
--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