Skip to contentSkip to conversation

FAILED: patch "[PATCH] x86/retbleed: Fix return thunk alignment" failed to apply to 5.15-stable tree

PATCH [email protected] 2 participants 4 messages May 13, 2023

Subsystems (2)
Stable branch x86
lore
Share & tools
lore

[email protected]

author maintainer
#0 lore raw

The patch below does not apply to the 5.15-stable tree.
If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <[email protected]>.

To reproduce the conflict and resubmit, you may use the following commands:

git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-5.15.y
git checkout FETCH_HEAD
git cherry-pick -x 9a48d604672220545d209e9996c2a1edbb5637f6
# <resolve conflicts, build, test, etc.>
git commit -s
git send-email --to '<[email protected]>' --in-reply-to '2023051313-wrangle-brick-b43d@gregkh' --subject-prefix 'PATCH 5.15.y' HEAD^..

Possible dependencies:

9a48d6046722 ("x86/retbleed: Fix return thunk alignment")
a149180fbcf3 ("x86: Add magic AMD return-thunk")
d9e9d2300681 ("x86,objtool: Create .return_sites")
15e67227c49a ("x86: Undo return-thunk damage")
0b53c374b9ef ("x86/retpoline: Use -mfunction-return")
369ae6ffc41a ("x86/retpoline: Cleanup some #ifdefery")
a883d624aed4 ("x86/cpufeatures: Move RETPOLINE flags to word 11")
22922deae13f ("Merge tag 'objtool-core-2022-05-23' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip")

thanks,

greg k-h

------------------ original commit in Linus's tree ------------------

From 9a48d604672220545d209e9996c2a1edbb5637f6 Mon Sep 17 00:00:00 2001
From: "Borislav Petkov (AMD)" <[email protected]>
Date: Fri, 12 May 2023 23:12:26 +0200
Subject: [PATCH] x86/retbleed: Fix return thunk alignment

SYM_FUNC_START_LOCAL_NOALIGN() adds an endbr leading to this layout (leaving only the last 2 bytes of the address):

3bff <zen_untrain_ret>:

  3bff:       f3 0f 1e fa             endbr64
  3c03:       f6                      test   $0xcc,%bl

3c04 <__x86_return_thunk>:

  3c04:       c3                      ret
  3c05:       cc                      int3
  3c06:       0f ae e8                lfence

However, "the RET at __x86_return_thunk must be on a 64 byte boundary, for alignment within the BTB."

Use SYM_START instead.

Signed-off-by Borislav Petkov (AMD) <[email protected]>
Reviewed-by Thomas Gleixner <[email protected]>
Cc <[email protected]>
Signed-off-by Linus Torvalds <[email protected]>
arch/x86/lib/retpoline.S +2 -2
index 27ef53fab6bd..b3b1e376dce8 100644
@@ -144,8 +144,8 @@ SYM_CODE_END(__x86_indirect_jump_thunk_array)
  */
 	.align 64
 	.skip 63, 0xcc
-SYM_FUNC_START_NOALIGN(zen_untrain_ret);
-
+SYM_START(zen_untrain_ret, SYM_L_GLOBAL, SYM_A_NONE)
+	ANNOTATE_NOENDBR
 	/*
 	 * As executed from zen_untrain_ret, this is:
 	 *
maintainer
#1 lore raw
Quoted 17 lines · On Sat, May 13, 2023 at 05:17:13PM +0900, [email protected] wrote:

The patch below does not apply to the 5.15-stable tree.
If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <[email protected]>.

To reproduce the conflict and resubmit, you may use the following commands:

git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-5.15.y
git checkout FETCH_HEAD
git cherry-pick -x 9a48d604672220545d209e9996c2a1edbb5637f6
# <resolve conflicts, build, test, etc.>
git commit -s
git send-email --to '<[email protected]>' --in-reply-to '2023051313-wrangle-brick-b43d@gregkh' --subject-prefix 'PATCH 5.15.y' HEAD^..

Possible dependencies:

Quoted 8 lines

9a48d6046722 ("x86/retbleed: Fix return thunk alignment")
a149180fbcf3 ("x86: Add magic AMD return-thunk")
d9e9d2300681 ("x86,objtool: Create .return_sites")
15e67227c49a ("x86: Undo return-thunk damage")
0b53c374b9ef ("x86/retpoline: Use -mfunction-return")
369ae6ffc41a ("x86/retpoline: Cleanup some #ifdefery")
a883d624aed4 ("x86/cpufeatures: Move RETPOLINE flags to word 11")
22922deae13f ("Merge tag 'objtool-core-2022-05-23' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip")

Right, so that looks like this:

The patch which causes the misalignment is

c4691712b546 ("x86/linkage: Add ENDBR to SYM_FUNC_START*()")

which came in 5.18 (v5.17-rc8-21-gc4691712b546) and the one which adds the actual untraining sequence:

a149180fbcf3 ("x86: Add magic AMD return-thunk")

came in 5.19.

So adding a Fixes: tag pointing to a patch which goes before the actual patch didn't make any sense to me last night.

HOWEVER, Thadeu did backport it here as

a9c0926fc754 ("x86: Add magic AMD return-thunk")

but the other patch wasn't backported.

So the 5.15 build looks good:

ffffffff81c01f7f <zen_untrain_ret>:
ffffffff81c01f7f: f6 .byte 0xf6

ffffffff81c01f80 <__x86_return_thunk>:

ffffffff81c01f80:       c3                      ret
ffffffff81c01f81:       cc                      int3
ffffffff81c01f82:       0f ae e8                lfence
ffffffff81c01f85:       eb f9                   jmp    ffffffff81c01f80 <__x86_return_thunk>
ffffffff81c01f87:       cc                      int3

So 5.15 doesn't need it.

Now lemme look at 5.10.

Thx.

Signature · 4 lines
Regards/Gruss,
    Boris.

https://people.kernel.org/tglx/notes-about-netiquette
author maintainer
#2 lore raw
Quoted 54 lines · Borislav Petkov
Quoted 17 lines · On Sat, May 13, 2023 at 05:17:13PM +0900, [email protected] wrote:

The patch below does not apply to the 5.15-stable tree.
If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <[email protected]>.

To reproduce the conflict and resubmit, you may use the following commands:

git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-5.15.y
git checkout FETCH_HEAD
git cherry-pick -x 9a48d604672220545d209e9996c2a1edbb5637f6
# <resolve conflicts, build, test, etc.>
git commit -s
git send-email --to '<[email protected]>' --in-reply-to '2023051313-wrangle-brick-b43d@gregkh' --subject-prefix 'PATCH 5.15.y' HEAD^..

Possible dependencies:

Quoted 8 lines

9a48d6046722 ("x86/retbleed: Fix return thunk alignment")
a149180fbcf3 ("x86: Add magic AMD return-thunk")
d9e9d2300681 ("x86,objtool: Create .return_sites")
15e67227c49a ("x86: Undo return-thunk damage")
0b53c374b9ef ("x86/retpoline: Use -mfunction-return")
369ae6ffc41a ("x86/retpoline: Cleanup some #ifdefery")
a883d624aed4 ("x86/cpufeatures: Move RETPOLINE flags to word 11")
22922deae13f ("Merge tag 'objtool-core-2022-05-23' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip")

Right, so that looks like this:

The patch which causes the misalignment is

c4691712b546 ("x86/linkage: Add ENDBR to SYM_FUNC_START*()")

which came in 5.18 (v5.17-rc8-21-gc4691712b546) and the one which adds the actual untraining sequence:

a149180fbcf3 ("x86: Add magic AMD return-thunk")

came in 5.19.

So adding a Fixes: tag pointing to a patch which goes before the actual patch didn't make any sense to me last night.

HOWEVER, Thadeu did backport it here as

a9c0926fc754 ("x86: Add magic AMD return-thunk")

but the other patch wasn't backported.

So the 5.15 build looks good:

ffffffff81c01f7f <zen_untrain_ret>:
ffffffff81c01f7f: f6 .byte 0xf6

Quoted 8 lines

ffffffff81c01f80 <__x86_return_thunk>:

ffffffff81c01f80:       c3                      ret
ffffffff81c01f81:       cc                      int3
ffffffff81c01f82:       0f ae e8                lfence
ffffffff81c01f85:       eb f9                   jmp    ffffffff81c01f80 <__x86_return_thunk>
ffffffff81c01f87:       cc                      int3

So 5.15 doesn't need it.

Great, thanks for looking into this, much appreciated.

greg k-h

maintainer
#3 lore raw
Quoted 1 line · Greg KH

Great, thanks for looking into this, much appreciated.

Sure, np.

And I just checked the 6.x streams - yes the backport is needed there so you taking them there is good. :)

Thx.

Signature · 4 lines
Regards/Gruss,
    Boris.

https://people.kernel.org/tglx/notes-about-netiquette