Verify that loading images with <4k section alignment either succeeds or fails gracefully (Bugzilla Bug 1860)

[tianocore-issues]

This issue was created automatically with bugzilla2github

Bugzilla Bug 1860

Date: 2019-05-29T06:43:55+00:00
From: @leiflindholm
To: Edhay <<edhaya.chandran>>
CC: arvin.chen, benjamin.you, edhaya.chandran, eric.jin, gaojie, @leiflindholm, samer.el-haj-mahmoud, siyuan.fu, xypron.glpk

Last updated: 2022-04-19T10:51:08+00:00

[tianocore-issues]

Comment 7851

Date: 2019-05-29 06:43:55 +0000
From: @leiflindholm

• Industry Specification: ---
• Release Observed: EDK II Master
• Releases to Fix: EDK II Master
• Target OS: Other
• Bugzilla Assignee(s): Edhay <<edhaya.chandran>>

We have seen issues in the field where platforms that enable NX attribute on the pe/coff header hang hard when presented.

So it would be beneficial if we had a test that verified graceful failure (or success) of images with <4k section alignment.

[tianocore-issues]

Comment 7916

Date: 2019-06-03 14:50:12 +0000
From: supreeth.venkatesh

Acknowledged.
Queued for development.

[tianocore-issues]

Comment 9818

Date: 2019-09-25 19:04:53 +0000
From: supreeth.venkatesh

Can you please point me to relevant section in the UEFI Specification which mandates section alignment restrictions so that additional test case can be added to existing image tests?

[tianocore-issues]

Comment 9822

Date: 2019-09-25 20:09:28 +0000
From: @leiflindholm

There is no explicit statement about the alignment.

There is an implicit requirement that if you want to use EFI_MEMORY_XP attribute, you need to keep non-executable regions in separate pages from executable regions - for example ensuring the page containing the PE header does not also contain code.

The request is for a test that ensures that if a platform enforces restrictions on alignment, then it fails gracefully when an image with smaller alignment is executed.

Today we have platforms in the wild that hang hard if you execute a program with < 4k section alignment.

[tianocore-issues]

Comment 15869

Date: 2021-05-01 04:40:51 +0000
From: Heinrich Schuchardt <<xypron.glpk>>

@leif:
Will you provide a patch?

[tianocore-issues]

Comment 16231

Date: 2021-06-01 07:54:34 +0000
From: @leiflindholm

Supreeth left ARM a while ago - re-assigning to Edhay for consideration.

[tianocore-issues]

Comment 18839

Date: 2022-04-19 10:51:08 +0000
From: Edhay <<edhaya.chandran>>

@lief: Can your team provide a patch for this solution?