From: "Laszlo Ersek" <lersek@redhat.com>
To: devel@edk2.groups.io, ard.biesheuvel@linaro.org
Cc: Leif Lindholm <leif@nuviainc.com>,
Ashish Singhal <ashishsingha@nvidia.com>
Subject: Re: [edk2-devel] [PATCH v3 0/3] ArmPkg/ArmMmuLib AARCH64: correctness fix
Date: Wed, 25 Mar 2020 20:49:01 +0100 [thread overview]
Message-ID: <4e92b528-921a-6159-55b8-bb0d6d7dea7d@redhat.com> (raw)
In-Reply-To: <20200325152940.1492-1-ard.biesheuvel@linaro.org>
On 03/25/20 16:29, Ard Biesheuvel wrote:
> The new ArmMmuLib code is easier to reason about, so that is what I did:
> currently, when we create mappings that cover existing table entries, we
> may end up overwriting those with block entries without taking the mapping
> attributes of the original table entries into account. So let's fix this.
>
> I honestly don't know whether the original code was better at dealing with
> this: I do remember some changes from Heyi that may have been related, but
> the old code is not easy to follow. In any case, I didn't manage to hit this
> case in practice, given that we typically start out with large mappings, and
> break them down later (to set permissions), rather than the other way around.
>
> Patch #1 adds some helpers to hide the insane way the type bits change
> meaning when you change to level 3.
>
> Patch #2 ensures that we only replace (and free) table entries with block
> entries if it is guaranteed that doing so will not lose any attribute
> information.
>
> Changes since v2:
> - add patch to limit recursion to levels < 3 in FreePageTablesRecursive()
>
> Changes since v1:
> - zero newly allocated pages before splitting a block entry into a table
> entry, to avoid garbage in that page being misidentified as entry type
> attributes - this should fix the crash observed by Laszlo
>
> Cc: Laszlo Ersek <lersek@redhat.com>
> Cc: Leif Lindholm <leif@nuviainc.com>
> Cc: Ashish Singhal <ashishsingha@nvidia.com>
>
> Ard Biesheuvel (3):
> ArmPkg/ArmMmuLib AARCH64: limit recursion when freeing page tables
> ArmPkg/ArmMmuLib AARCH64: use helpers to determine table entry types
> ArmPkg/ArmMmuLib AARCH64: preserve attributes when replacing a table
> entry
>
> .../Library/ArmMmuLib/AArch64/ArmMmuLibCore.c | 83 +++++++++++++++----
> 1 file changed, 68 insertions(+), 15 deletions(-)
>
Tested-by: Laszlo Ersek <lersek@redhat.com>
Thanks!
Laszlo
next prev parent reply other threads:[~2020-03-25 19:49 UTC|newest]
Thread overview: 12+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-03-25 15:29 [PATCH v3 0/3] ArmPkg/ArmMmuLib AARCH64: correctness fix Ard Biesheuvel
2020-03-25 15:29 ` [PATCH v3 1/3] ArmPkg/ArmMmuLib AARCH64: limit recursion when freeing page tables Ard Biesheuvel
2020-03-25 15:47 ` Ashish Singhal
2020-03-26 10:22 ` Leif Lindholm
2020-03-26 10:25 ` Ard Biesheuvel
2020-03-25 15:29 ` [PATCH v3 2/3] ArmPkg/ArmMmuLib AARCH64: use helpers to determine table entry types Ard Biesheuvel
2020-03-25 15:47 ` Ashish Singhal
2020-03-25 15:29 ` [PATCH v3 3/3] ArmPkg/ArmMmuLib AARCH64: preserve attributes when replacing a table entry Ard Biesheuvel
2020-03-25 15:47 ` Ashish Singhal
2020-03-25 15:48 ` [PATCH v3 0/3] ArmPkg/ArmMmuLib AARCH64: correctness fix Ashish Singhal
2020-03-25 19:49 ` Laszlo Ersek [this message]
2020-03-26 10:35 ` [edk2-devel] " 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=4e92b528-921a-6159-55b8-bb0d6d7dea7d@redhat.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