Skip to content

Update patches (November 22, 2024) #790

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 3 commits into from
Nov 23, 2024
Merged

Update patches (November 22, 2024) #790

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
576f1ae
patches: mainline: Update Hexagon patch's metadata
nathanchance Nov 23, 2024
f942c8b
patches: stable: Apply Hexagon fix
nathanchance Nov 23, 2024
c13391f
patches: arm64-fixes: Apply fix for "HYP init too big" error
nathanchance Nov 23, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
141 changes: 141 additions & 0 deletions patches/arm64-fixes/afa9b48f327c9ef36bfba4c643a29385a633252b.patch
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,141 @@
From afa9b48f327c9ef36bfba4c643a29385a633252b Mon Sep 17 00:00:00 2001
From: Marc Zyngier <[email protected]>
Date: Thu, 10 Oct 2024 16:13:26 +0100
Subject: KVM: arm64: Shave a few bytes from the EL2 idmap code

Our idmap is becoming too big, to the point where it doesn't fit in
a 4kB page anymore.

There are some low-hanging fruits though, such as the el2_init_state
horror that is expanded 3 times in the kernel. Let's at least limit
ourselves to two copies, which makes the kernel link again.

At some point, we'll have to have a better way of doing this.

Reported-by: Nathan Chancellor <[email protected]>
Signed-off-by: Marc Zyngier <[email protected]>
Link: https://lore.kernel.org/r/20241009204903.GA3353168@thelio-3990X
---
Link: https://git.kernel.org/linus/afa9b48f327c9ef36bfba4c643a29385a633252b
---
arch/arm64/include/asm/kvm_asm.h | 1 +
arch/arm64/kernel/asm-offsets.c | 1 +
arch/arm64/kvm/hyp/nvhe/hyp-init.S | 52 +++++++++++++++++++++-----------------
3 files changed, 31 insertions(+), 23 deletions(-)

diff --git a/arch/arm64/include/asm/kvm_asm.h b/arch/arm64/include/asm/kvm_asm.h
index b36a3b6cc01169..67afac659231ed 100644
--- a/arch/arm64/include/asm/kvm_asm.h
+++ b/arch/arm64/include/asm/kvm_asm.h
@@ -178,6 +178,7 @@ struct kvm_nvhe_init_params {
unsigned long hcr_el2;
unsigned long vttbr;
unsigned long vtcr;
+ unsigned long tmp;
};

/*
diff --git a/arch/arm64/kernel/asm-offsets.c b/arch/arm64/kernel/asm-offsets.c
index 27de1dddb0abee..b21dd24b8efc3b 100644
--- a/arch/arm64/kernel/asm-offsets.c
+++ b/arch/arm64/kernel/asm-offsets.c
@@ -146,6 +146,7 @@ int main(void)
DEFINE(NVHE_INIT_HCR_EL2, offsetof(struct kvm_nvhe_init_params, hcr_el2));
DEFINE(NVHE_INIT_VTTBR, offsetof(struct kvm_nvhe_init_params, vttbr));
DEFINE(NVHE_INIT_VTCR, offsetof(struct kvm_nvhe_init_params, vtcr));
+ DEFINE(NVHE_INIT_TMP, offsetof(struct kvm_nvhe_init_params, tmp));
#endif
#ifdef CONFIG_CPU_PM
DEFINE(CPU_CTX_SP, offsetof(struct cpu_suspend_ctx, sp));
diff --git a/arch/arm64/kvm/hyp/nvhe/hyp-init.S b/arch/arm64/kvm/hyp/nvhe/hyp-init.S
index 401af1835be6b7..fc186622606767 100644
--- a/arch/arm64/kvm/hyp/nvhe/hyp-init.S
+++ b/arch/arm64/kvm/hyp/nvhe/hyp-init.S
@@ -24,28 +24,25 @@
.align 11

SYM_CODE_START(__kvm_hyp_init)
- ventry __invalid // Synchronous EL2t
- ventry __invalid // IRQ EL2t
- ventry __invalid // FIQ EL2t
- ventry __invalid // Error EL2t
+ ventry . // Synchronous EL2t
+ ventry . // IRQ EL2t
+ ventry . // FIQ EL2t
+ ventry . // Error EL2t

- ventry __invalid // Synchronous EL2h
- ventry __invalid // IRQ EL2h
- ventry __invalid // FIQ EL2h
- ventry __invalid // Error EL2h
+ ventry . // Synchronous EL2h
+ ventry . // IRQ EL2h
+ ventry . // FIQ EL2h
+ ventry . // Error EL2h

ventry __do_hyp_init // Synchronous 64-bit EL1
- ventry __invalid // IRQ 64-bit EL1
- ventry __invalid // FIQ 64-bit EL1
- ventry __invalid // Error 64-bit EL1
+ ventry . // IRQ 64-bit EL1
+ ventry . // FIQ 64-bit EL1
+ ventry . // Error 64-bit EL1

- ventry __invalid // Synchronous 32-bit EL1
- ventry __invalid // IRQ 32-bit EL1
- ventry __invalid // FIQ 32-bit EL1
- ventry __invalid // Error 32-bit EL1
-
-__invalid:
- b .
+ ventry . // Synchronous 32-bit EL1
+ ventry . // IRQ 32-bit EL1
+ ventry . // FIQ 32-bit EL1
+ ventry . // Error 32-bit EL1

/*
* Only uses x0..x3 so as to not clobber callee-saved SMCCC registers.
@@ -76,6 +73,13 @@ __do_hyp_init:
eret
SYM_CODE_END(__kvm_hyp_init)

+SYM_CODE_START_LOCAL(__kvm_init_el2_state)
+ /* Initialize EL2 CPU state to sane values. */
+ init_el2_state // Clobbers x0..x2
+ finalise_el2_state
+ ret
+SYM_CODE_END(__kvm_init_el2_state)
+
/*
* Initialize the hypervisor in EL2.
*
@@ -102,9 +106,12 @@ SYM_CODE_START_LOCAL(___kvm_hyp_init)
// TPIDR_EL2 is used to preserve x0 across the macro maze...
isb
msr tpidr_el2, x0
- init_el2_state
- finalise_el2_state
+ str lr, [x0, #NVHE_INIT_TMP]
+
+ bl __kvm_init_el2_state
+
mrs x0, tpidr_el2
+ ldr lr, [x0, #NVHE_INIT_TMP]

1:
ldr x1, [x0, #NVHE_INIT_TPIDR_EL2]
@@ -199,9 +206,8 @@ SYM_CODE_START_LOCAL(__kvm_hyp_init_cpu)

2: msr SPsel, #1 // We want to use SP_EL{1,2}

- /* Initialize EL2 CPU state to sane values. */
- init_el2_state // Clobbers x0..x2
- finalise_el2_state
+ bl __kvm_init_el2_state
+
__init_el2_nvhe_prepare_eret

/* Enable MMU, set vectors and stack. */
--
cgit 1.2.3-korg

1 change: 1 addition & 0 deletions patches/arm64-fixes/series
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,2 +1,3 @@
506a1ac4c4464a61e4336e135841067dbc040aaa.patch
20241018_arnd_media_mediatek_vcodec_mark_vdec_vp9_slice_map_counts_eob_coef_noinline.patch
afa9b48f327c9ef36bfba4c643a29385a633252b.patch
2 changes: 1 addition & 1 deletion patches/mainline/series
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,2 +1,2 @@
20240819_nathan_hexagon_disable_constant_extender_optimization_for_llvm_prior_to_19_1_0.patch
20241018_arnd_media_mediatek_vcodec_mark_vdec_vp9_slice_map_counts_eob_coef_noinline.patch
v2_20241121_nathan_hexagon_disable_constant_extender_optimization_for_llvm_prior_to_19_1_0.patch
28 changes: 20 additions & 8 deletions ...timization_for_llvm_prior_to_19_1_0.patch → ...timization_for_llvm_prior_to_19_1_0.patch
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,15 +1,16 @@
From git@z Thu Jan 1 00:00:00 1970
Subject: [PATCH] hexagon: Disable constant extender optimization for LLVM
prior to 19.1.0
Subject: [PATCH v2] hexagon: Disable constant extender optimization for
LLVM prior to 19.1.0
From: Nathan Chancellor <[email protected]>
Date: Mon, 19 Aug 2024 11:16:10 -0700
Message-Id: <20240819-hexagon-disable-constant-expander-pass-v1-1-36a734e9527d@kernel.org>
Date: Thu, 21 Nov 2024 11:22:18 -0700
Message-Id: <20241121-hexagon-disable-constant-expander-pass-v2-1-1a92e9afb0f4@kernel.org>
MIME-Version: 1.0
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: 7bit

The Hexagon-specific constant extender optimization in LLVM may crash on
Linux kernel code [1] (such as with a bcachefs change in -next):
Linux kernel code [1], such as fs/bcache/btree_io.c after
commit 32ed4a620c54 ("bcachefs: Btree path tracepoints") in 6.12:

clang: llvm/lib/Target/Hexagon/HexagonConstExtenders.cpp:745: bool (anonymous namespace)::HexagonConstExtenders::ExtRoot::operator<(const HCE::ExtRoot &) const: Assertion `ThisB->getParent() == OtherB->getParent()' failed.
Stack dump:
Expand All @@ -26,18 +27,29 @@ This has been resolved in LLVM main (20.0.0) [2] and backported to LLVM
constant expander optimization using the '-mllvm' option when using a
toolchain that is not fixed.

Cc: [email protected]
Link: https://github.com/llvm/llvm-project/issues/99714 [1]
Link: https://github.com/llvm/llvm-project/commit/68df06a0b2998765cb0a41353fcf0919bbf57ddb [2]
Link: https://github.com/llvm/llvm-project/commit/2ab8d93061581edad3501561722ebd5632d73892 [3]
Reviewed-by: Brian Cain <[email protected]>
Link: https://lore.kernel.org/r/20240819-hexagon-disable-constant-expander-pass-v1-1-36a734e9527d@kernel.org
Link: https://lore.kernel.org/r/20241121-hexagon-disable-constant-expander-pass-v2-1-1a92e9afb0f4@kernel.org
Signed-off-by: Nathan Chancellor <[email protected]>
---
Andrew, can you please take this for 6.13? Our CI continues to hit this.

Changes in v2:
- Rebase on 6.12 to make sure it is still applicable
- Name exact bcachefs commit that introduces crash now that it is
merged
- Add 'Cc: stable' as this is now visible in a stable release
- Carry forward Brian's reviewed-by
- Link to v1: https://lore.kernel.org/r/20240819-hexagon-disable-constant-expander-pass-v1-1-36a734e9527d@kernel.org
---
arch/hexagon/Makefile | 6 ++++++
1 file changed, 6 insertions(+)

diff --git a/arch/hexagon/Makefile b/arch/hexagon/Makefile
index 92d005958dfb..ff172cbe5881 100644
index 92d005958dfb232d48a4ca843b46262a84a08eb4..ff172cbe5881a074f9d9430c37071992a4c8beac 100644
--- a/arch/hexagon/Makefile
+++ b/arch/hexagon/Makefile
@@ -32,3 +32,9 @@ KBUILD_LDFLAGS += $(ldflags-y)
Expand All @@ -52,7 +64,7 @@ index 92d005958dfb..ff172cbe5881 100644
+endif

---
base-commit: 47ac09b91befbb6a235ab620c32af719f8208399
base-commit: adc218676eef25575469234709c2d87185ca223a
change-id: 20240802-hexagon-disable-constant-expander-pass-8b6b61db6afc

Best regards,
Expand Down
1 change: 1 addition & 0 deletions patches/stable/series
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1 +1,2 @@
20241018_arnd_media_mediatek_vcodec_mark_vdec_vp9_slice_map_counts_eob_coef_noinline.patch
v2_20241121_nathan_hexagon_disable_constant_extender_optimization_for_llvm_prior_to_19_1_0.patch
73 changes: 73 additions & 0 deletions ...1121_nathan_hexagon_disable_constant_extender_optimization_for_llvm_prior_to_19_1_0.patch
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,73 @@
From git@z Thu Jan 1 00:00:00 1970
Subject: [PATCH v2] hexagon: Disable constant extender optimization for
LLVM prior to 19.1.0
From: Nathan Chancellor <[email protected]>
Date: Thu, 21 Nov 2024 11:22:18 -0700
Message-Id: <20241121-hexagon-disable-constant-expander-pass-v2-1-1a92e9afb0f4@kernel.org>
MIME-Version: 1.0
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: 7bit

The Hexagon-specific constant extender optimization in LLVM may crash on
Linux kernel code [1], such as fs/bcache/btree_io.c after
commit 32ed4a620c54 ("bcachefs: Btree path tracepoints") in 6.12:

clang: llvm/lib/Target/Hexagon/HexagonConstExtenders.cpp:745: bool (anonymous namespace)::HexagonConstExtenders::ExtRoot::operator<(const HCE::ExtRoot &) const: Assertion `ThisB->getParent() == OtherB->getParent()' failed.
Stack dump:
0. Program arguments: clang --target=hexagon-linux-musl ... fs/bcachefs/btree_io.c
1. <eof> parser at end of file
2. Code generation
3. Running pass 'Function Pass Manager' on module 'fs/bcachefs/btree_io.c'.
4. Running pass 'Hexagon constant-extender optimization' on function '@__btree_node_lock_nopath'

Without assertions enabled, there is just a hang during compilation.

This has been resolved in LLVM main (20.0.0) [2] and backported to LLVM
19.1.0 but the kernel supports LLVM 13.0.1 and newer, so disable the
constant expander optimization using the '-mllvm' option when using a
toolchain that is not fixed.

Cc: [email protected]
Link: https://github.com/llvm/llvm-project/issues/99714 [1]
Link: https://github.com/llvm/llvm-project/commit/68df06a0b2998765cb0a41353fcf0919bbf57ddb [2]
Link: https://github.com/llvm/llvm-project/commit/2ab8d93061581edad3501561722ebd5632d73892 [3]
Reviewed-by: Brian Cain <[email protected]>
Link: https://lore.kernel.org/r/20241121-hexagon-disable-constant-expander-pass-v2-1-1a92e9afb0f4@kernel.org
Signed-off-by: Nathan Chancellor <[email protected]>
---
Andrew, can you please take this for 6.13? Our CI continues to hit this.

Changes in v2:
- Rebase on 6.12 to make sure it is still applicable
- Name exact bcachefs commit that introduces crash now that it is
merged
- Add 'Cc: stable' as this is now visible in a stable release
- Carry forward Brian's reviewed-by
- Link to v1: https://lore.kernel.org/r/20240819-hexagon-disable-constant-expander-pass-v1-1-36a734e9527d@kernel.org
---
arch/hexagon/Makefile | 6 ++++++
1 file changed, 6 insertions(+)

diff --git a/arch/hexagon/Makefile b/arch/hexagon/Makefile
index 92d005958dfb232d48a4ca843b46262a84a08eb4..ff172cbe5881a074f9d9430c37071992a4c8beac 100644
--- a/arch/hexagon/Makefile
+++ b/arch/hexagon/Makefile
@@ -32,3 +32,9 @@ KBUILD_LDFLAGS += $(ldflags-y)
TIR_NAME := r19
KBUILD_CFLAGS += -ffixed-$(TIR_NAME) -DTHREADINFO_REG=$(TIR_NAME) -D__linux__
KBUILD_AFLAGS += -DTHREADINFO_REG=$(TIR_NAME)
+
+# Disable HexagonConstExtenders pass for LLVM versions prior to 19.1.0
+# https://github.com/llvm/llvm-project/issues/99714
+ifneq ($(call clang-min-version, 190100),y)
+KBUILD_CFLAGS += -mllvm -hexagon-cext=false
+endif

---
base-commit: adc218676eef25575469234709c2d87185ca223a
change-id: 20240802-hexagon-disable-constant-expander-pass-8b6b61db6afc

Best regards,
--
Nathan Chancellor <[email protected]>