Skip to contentSkip to conversation

x86/cleanups for v6.14

GIT PULL Merged Ingo Molnar 6 participants 16 messages Jan 21, 2025 – Jan 28, 2025 x86

lore
Share & tools
lore
#0 lore raw

Linus,

Please pull the latest x86/cleanups Git tree from:

git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git x86-cleanups-2025-01-21

# HEAD: 0094014be0cd75273ef7f2934c17fb8cffd4db6e x86/ioapic: Remove a stray tab in the IO-APIC type string

Miscellaneous x86 cleanups and typo fixes, and also the removal of the "disablelapic" boot parameter.

Thanks,

Ingo

------------------>
Alan Song (1):
x86/ioapic: Remove a stray tab in the IO-APIC type string

Baoquan He (2):
x86/ioremap: Simplify setup_data mapping variants x86/ioremap: Remove unused size parameter in remapping functions

Borislav Petkov (AMD) (5):
x86/boot/compressed: Remove unused header includes from kaslr.c
Documentation: Merge x86-specific boot options doc into kernel-parameters.txt x86/apic: Remove "disablelapic" cmdline option Documentation/kernel-parameters: Fix a typo in kvm.enable_virt_at_load text x86/cpufeatures: Remove "AMD" from the comments to the AMD-specific leaf

Raag Jadav (1):

      x86/cpu: Fix typo in x86_match_cpu()'s doc
 Documentation/admin-guide/kernel-parameters.rst    |   3 -
 Documentation/admin-guide/kernel-parameters.txt    | 239 +++++++++++++++-
 Documentation/arch/x86/x86_64/boot-options.rst     | 312 ---------------------
 .../arch/x86/x86_64/fake-numa-for-cpusets.rst      |   2 +-
 Documentation/arch/x86/x86_64/index.rst            |   1 -
 arch/x86/Kconfig.debug                             |   2 +-
 arch/x86/boot/compressed/kaslr.c                   |   4 -
 arch/x86/include/asm/cpufeatures.h                 |  12 +-
 arch/x86/kernel/apic/apic.c                        |   9 +-
 arch/x86/kernel/apic/io_apic.c                     |   2 +-
 arch/x86/kernel/cpu/match.c                        |   2 +-
 arch/x86/kernel/cpu/topology.c                     |   2 +-
 arch/x86/kernel/pci-dma.c                          |   4 -
 arch/x86/mm/ioremap.c                              | 117 +++-----
 14 files changed, 277 insertions(+), 434 deletions(-)
 delete mode 100644 Documentation/arch/x86/x86_64/boot-options.rst
#1 lore raw
Quoted 10 lines · Ingo Molnar

Linus,

Please pull the latest x86/cleanups Git tree from:

git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git x86-cleanups-2025-01-21

# HEAD: 0094014be0cd75273ef7f2934c17fb8cffd4db6e x86/ioapic: Remove a stray tab in the IO-APIC type string

Miscellaneous x86 cleanups and typo fixes, and also the removal of the "disablelapic" boot parameter.

Merge note: if you've pulled x86/sev from Boris already, then there will be a new conflict in arch/x86/include/asm/cpufeatures.h due to overlapping (but compatible) changes to the same lines of code.

My conflict resolution is below, for reference.

Thanks,

Ingo

=======================>

Merge branch 'x86/cleanups' into tmp.tmp

Conflicts:
arch/x86/include/asm/cpufeatures.h

arch/x86/include/asm/cpufeatures.h +0 -0
diff --cc arch/x86/include/asm/cpufeatures.h
index 8b55685255d6,09e1e54676f4..9746e75a1866
@@@ -443,18 -443,15 +443,18 @@@
  #define X86_FEATURE_SPEC_CTRL_SSBD	(18*32+31) /* Speculative Store Bypass Disable */
  
  /* AMD-defined memory encryption features, CPUID level 0x8000001f (EAX), word 19 */
- #define X86_FEATURE_SME			(19*32+ 0) /* "sme" AMD Secure Memory Encryption */
- #define X86_FEATURE_SEV			(19*32+ 1) /* "sev" AMD Secure Encrypted Virtualization */
+ #define X86_FEATURE_SME			(19*32+ 0) /* "sme" Secure Memory Encryption */
+ #define X86_FEATURE_SEV			(19*32+ 1) /* "sev" Secure Encrypted Virtualization */
  #define X86_FEATURE_VM_PAGE_FLUSH	(19*32+ 2) /* VM Page Flush MSR is supported */
- #define X86_FEATURE_SEV_ES		(19*32+ 3) /* "sev_es" AMD Secure Encrypted Virtualization - Encrypted State */
- #define X86_FEATURE_SEV_SNP		(19*32+ 4) /* "sev_snp" AMD Secure Encrypted Virtualization - Secure Nested Paging */
+ #define X86_FEATURE_SEV_ES		(19*32+ 3) /* "sev_es" Secure Encrypted Virtualization - Encrypted State */
+ #define X86_FEATURE_SEV_SNP		(19*32+ 4) /* "sev_snp" Secure Encrypted Virtualization - Secure Nested Paging */
  #define X86_FEATURE_V_TSC_AUX		(19*32+ 9) /* Virtual TSC_AUX */
- #define X86_FEATURE_SME_COHERENT	(19*32+10) /* AMD hardware-enforced cache coherency */
- #define X86_FEATURE_DEBUG_SWAP		(19*32+14) /* "debug_swap" AMD SEV-ES full debug state swap support */
+ #define X86_FEATURE_SME_COHERENT	(19*32+10) /* hardware-enforced cache coherency */
+ #define X86_FEATURE_DEBUG_SWAP		(19*32+14) /* "debug_swap" SEV-ES full debug state swap support */
 +#define X86_FEATURE_RMPREAD		(19*32+21) /* RMPREAD instruction */
 +#define X86_FEATURE_SEGMENTED_RMP	(19*32+23) /* Segmented RMP support */
  #define X86_FEATURE_SVSM		(19*32+28) /* "svsm" SVSM present */
 +#define X86_FEATURE_HV_INUSE_WR_ALLOWED	(19*32+30) /* Allow Write to in-use hypervisor-owned pages */
  
  /* AMD-defined Extended Feature 2 EAX, CPUID level 0x80000021 (EAX), word 20 */
  #define X86_FEATURE_NO_NESTED_DATA_BP	(20*32+ 0) /* No Nested Data Breakpoints */
maintainer
#2 lore raw
Quoted 1 line · Ingo Molnar

Merge note: if you've pulled x86/sev from Boris already, then there

Right, FWIW, I try to merge the branches locally here in the same order we send them to Linus so that I can anticipate any merge conflicts.

Quoted 4 lines

will be a new conflict in arch/x86/include/asm/cpufeatures.h due to overlapping (but compatible) changes to the same lines of code.

My conflict resolution is below, for reference.

And yes, I can confirm the same resolution I've been doing here too.

Thx.

Signature · 4 lines
Regards/Gruss,
    Boris.

https://people.kernel.org/tglx/notes-about-netiquette
merge bot Merged
#3 lore raw
In reply to Ingo Molnar
The pull request you sent on Tue, 21 Jan 2025 08:28:27 +0100:

The pull request you sent on Tue, 21 Jan 2025 08:28:27 +0100:

Quoted 1 line

git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git x86-cleanups-2025-01-21

has been merged into torvalds/linux.git:
https://git.kernel.org/torvalds/c/858df1de2158bd7ab88d31b557592e990bacc0b5

Thank you!

Signature · 2 lines
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/prtracker.html
#4 lore raw

[GIT PULL] x86/boot enhancements for v6.14

In reply to Ingo Molnar

Linus,

Please pull the latest x86/boot Git tree from:

git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git x86-boot-2025-01-21

# HEAD: cf4ca80650908628bf1c0c29e3fd236b1915d789 x86/sev: Disable ftrace branch profiling in SEV startup code

MERGE NOTE: there will be a new conflict when pulling this tree, due to overlapping additions to arch/x86/include/asm/sev-common.h. My resolution was to advance the Linux-internal enumeration of reason codes, in merge order:

     #define GHCB_TERM_SECURE_TSC		10	/* Secure TSC initialization failed */
     #define GHCB_TERM_SVSM_CA_REMAP_FAIL	11	/* SVSM is present but CA could not be remapped */

x86/boot changes for v6.14:

- A large and involved preparatory series to pave the way to add exception handling for relocate_kernel - which will be a debugging facility that has aided in the field to debug an exceptionally hard to debug early boot bug. Plus assorted cleanups and fixes that were discovered along the way, by David Woodhouse:

- Clean up and document register use in relocate_kernel_64.S
- Use named labels in swap_pages in relocate_kernel_64.S
- Only swap pages for ::preserve_context mode
- Allocate PGD for x86_64 transition page tables separately
- Copy control page into place in machine_kexec_prepare()
- Invoke copy of relocate_kernel() instead of the original
- Move relocate_kernel to kernel .data section
- Add data section to relocate_kernel
- Drop page_list argument from relocate_kernel()
- Eliminate writes through kernel mapping of relocate_kernel page
- Clean up register usage in relocate_kernel()
- Mark relocate_kernel page as ROX instead of RWX
- Disable global pages before writing to control page
- Ensure preserve_context flag is set on return to kernel
- Use correct swap page in swap_pages function
- Fix stack and handling of re-entry point for ::preserve_context
- Mark machine_kexec() with __nocfi
- Cope with relocate_kernel() not being at the start of the page
- Use typedef for relocate_kernel_fn function prototype
- Fix location of relocate_kernel with -ffunction-sections (fix by Nathan Chancellor)

- A series to remove the last remaining absolute symbol references from .head.text, and enforce this at build time, by Ard Biesheuvel:

- Avoid WARN()s and panic()s in early boot code
- Don't hang but terminate on failure to remap SVSM CA
- Determine VA/PA offset before entering C code
- Avoid intentional absolute symbol references in .head.text
- Disable UBSAN in early boot code
- Move ENTRY_TEXT to the start of the image
- Move .head.text into its own output section
- Reject absolute references in .head.text

- Which build-time enforcement uncovered a handful of bugs of essentially non-working code, and a wrokaround for a toolchain bug, fixed by Ard Biesheuvel as well:

- Fix spurious undefined reference when CONFIG_X86_5LEVEL=n, on GCC-12
- Disable UBSAN on SEV code that may execute very early
- Disable ftrace branch profiling in SEV startup code

- And miscellaneous cleanups:

- kexec_core: Add and update comments regarding the KEXEC_JUMP flow (Rafael J. Wysocki)
- x86/sysfs: Constify 'struct bin_attribute' (Thomas Weißschuh)

Ard Biesheuvel (11):
x86/sev: Avoid WARN()s and panic()s in early boot code
x86/boot/64: Determine VA/PA offset before entering C code x86/boot/64: Avoid intentional absolute symbol references in .head.text x86/boot: Disable UBSAN in early boot code
x86/kernel: Move ENTRY_TEXT to the start of the image
x86/boot: Move .head.text into its own output section
x86/boot: Reject absolute references in .head.text
x86/boot/64: Fix spurious undefined reference when CONFIG_X86_5LEVEL=n, on GCC-12 x86/sev: Disable UBSAN on SEV code that may execute very early x86/sev: Don't hang but terminate on failure to remap SVSM CA x86/sev: Disable ftrace branch profiling in SEV startup code

David Woodhouse (19):
x86/kexec: Clean up and document register use in relocate_kernel_64.S x86/kexec: Use named labels in swap_pages in relocate_kernel_64.S x86/kexec: Only swap pages for ::preserve_context mode x86/kexec: Allocate PGD for x86_64 transition page tables separately x86/kexec: Copy control page into place in machine_kexec_prepare() x86/kexec: Invoke copy of relocate_kernel() instead of the original x86/kexec: Move relocate_kernel to kernel .data section x86/kexec: Add data section to relocate_kernel
x86/kexec: Drop page_list argument from relocate_kernel() x86/kexec: Eliminate writes through kernel mapping of relocate_kernel page x86/kexec: Clean up register usage in relocate_kernel() x86/kexec: Mark relocate_kernel page as ROX instead of RWX x86/kexec: Disable global pages before writing to control page x86/kexec: Ensure preserve_context flag is set on return to kernel x86/kexec: Use correct swap page in swap_pages function x86/kexec: Fix stack and handling of re-entry point for ::preserve_context x86/kexec: Mark machine_kexec() with __nocfi
x86/kexec: Cope with relocate_kernel() not being at the start of the page x86/kexec: Use typedef for relocate_kernel_fn function prototype

Ingo Molnar (1):
Merge branch 'x86/urgent' into x86/boot, to pick up dependent fixes

Nathan Chancellor (1):
x86/kexec: Fix location of relocate_kernel with -ffunction-sections

Rafael J. Wysocki (1):
kexec_core: Add and update comments regarding the KEXEC_JUMP flow

Thomas Weißschuh (1):
x86/sysfs: Constify 'struct bin_attribute'

 arch/x86/coco/sev/Makefile           |   3 +
 arch/x86/coco/sev/core.c             |  15 ++-
 arch/x86/coco/sev/shared.c           |  16 +--
 arch/x86/include/asm/init.h          |   2 +-
 arch/x86/include/asm/kexec.h         |  54 +++++-----
 arch/x86/include/asm/sections.h      |   1 +
 arch/x86/include/asm/setup.h         |   2 +-
 arch/x86/include/asm/sev-common.h    |   1 +
 arch/x86/kernel/callthunks.c         |   6 ++
 arch/x86/kernel/head64.c             |  40 +++++---
 arch/x86/kernel/head_64.S            |  12 ++-
 arch/x86/kernel/ksysfs.c             |  18 ++--
 arch/x86/kernel/machine_kexec_32.c   |   7 +-
 arch/x86/kernel/machine_kexec_64.c   |  95 ++++++++++-------
 arch/x86/kernel/relocate_kernel_64.S | 193 +++++++++++++++++++----------------
 arch/x86/kernel/vmlinux.lds.S        |  45 +++++---
 arch/x86/tools/relocs.c              |   8 +-
 kernel/kexec_core.c                  |  23 +++--
 18 files changed, 317 insertions(+), 224 deletions(-)
merge bot Merged
#5 lore raw

Re: [GIT PULL] x86/boot enhancements for v6.14

The pull request you sent on Tue, 21 Jan 2025 22:29:29 +0100:

The pull request you sent on Tue, 21 Jan 2025 22:29:29 +0100:

Quoted 1 line

git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git x86-boot-2025-01-21

has been merged into torvalds/linux.git:
https://git.kernel.org/torvalds/c/5b7f7234ff18a5f48e0d36c417a0d07c7fdfc61a

Thank you!

Signature · 2 lines
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/prtracker.html
#6 lore raw

Re: [GIT PULL] x86/boot enhancements for v6.14

In reply to Ingo Molnar
Quoted 11 lines · Ingo Molnar

- A series to remove the last remaining absolute symbol references from .head.text, and enforce this at build time, by Ard Biesheuvel:
[...]
- Which build-time enforcement uncovered a handful of bugs of essentially non-working code, and a wrokaround for a toolchain bug, fixed by Ard Biesheuvel as well:

- Fix spurious undefined reference when CONFIG_X86_5LEVEL=n, on GCC-12
- Disable UBSAN on SEV code that may execute very early
- Disable ftrace branch profiling in SEV startup code

Bah. I only noticed this today, because I was on the road part of the week and didn't do my usual "build with clang".

But this is broken with my normal clang config, and I get a very unhelpful error message:

Absolute reference to symbol '.rodata' not permitted in .head.text

and I have no idea why, or where it comes from. And that error message doesn't specify enough information for me to even *guess* at what's going on.

And bisecting it obviously just points at faf0ed487415 ("x86/boot:
Reject absolute references in .head.text"), since it's just a random new check for an old issue.

New random rule, with new random error as a result, and totally opaque to anybody else than Ard.

Useless crap, in other words.

Why isn't the fix a revert? Because that error message is really bad. It needs to tell me where things went wrong, not just a "You're effed. Ha! Ha!".

And to add insult to injury, all of this is done in-place on the vmlinux file, so when it all fails, make does

make[2]: *** Deleting file 'vmlinux'

and doesn't even leave behind anything to look at.

So I really think this was all horribly badly done.

Anyway, not know what the right thing to do is, I hacked up the makefiles to squirrel off a copy of the vmlinux file, and did

        objdump --no-addresses --no-show-raw-insn \
                -j .head.text --disassemble \
                -rR ORIGINAL  | less -S

and sure enough, it shows things like this:

<snp_cpuid>:

        push   %rbp
        push   %r15
        push   %r14
        push   %rbx
        sub    $0x18,%rsp
        lea    0x5ac8fb(%rip),%r8        # <cpuid_table_copy>
                        R_X86_64_PC32   .rodata+0x4f1758
        mov    (%r8),%eax
        test   %eax,%eax
    ..
        lea    0x5ac841(%rip),%rax        # <cpuid_std_range_max>
                        R_X86_64_PC32   .rodata+0x4f174c
    ..
        jmp    *-0x7dfffe90(,%r9,8)
                        R_X86_64_32S    .rodata+0x170
    ..

and

<snp_init>:

        push   %rbp
        push   %r15
        push   %r14
        push   %r12
    ..
        mov    %rax,0x5ac3bb(%rip)        # <secrets_pa>
                        R_X86_64_PC32   .rodata+0x4f173c
        mov    0x18(%r15),%rsi
        test   %rsi,%rsi
    ..

and

<pvalidate_4k_page>:

        push   %rbx
        sub    $0x60,%rsp
        lea    0x5ac0c4(%rip),%rax        # <snp_vmpl>
   ..
        test   $0x4,%cl
        jne    <pvalidate_4k_page+0xad>
        lea    0x5ac068(%rip),%rax        # <boot_svsm_caa>
                        R_X86_64_PC32   .rodata+0x4f1744
        mov    (%rax),%rax
        jmp    <pvalidate_4k_page+0xb8>
    ..

so those relocs to rodata most definitely exist.

Anyway, that check needs to either

(a) die a painful death very quickly

(b) be made to actually print out useful information of WHERE the
relocation comes from and WHERE it points to

because the current implementation of that check is not acceptable.

The next time I have to play makefile games and then do objdump by hand to figure out what the %^$% the build is complaining about, I'm just reverting it outright and not writing this long explanation of the problem.

Linus

#7 lore raw

PATCH 1/1 Revert "x86/boot: Reject absolute references in .head.text

Quoted 38 lines · Linus Torvalds
Quoted 11 lines · Ingo Molnar

- A series to remove the last remaining absolute symbol references from .head.text, and enforce this at build time, by Ard Biesheuvel:
[...]
- Which build-time enforcement uncovered a handful of bugs of essentially non-working code, and a wrokaround for a toolchain bug, fixed by Ard Biesheuvel as well:

- Fix spurious undefined reference when CONFIG_X86_5LEVEL=n, on GCC-12
- Disable UBSAN on SEV code that may execute very early
- Disable ftrace branch profiling in SEV startup code

Bah. I only noticed this today, because I was on the road part of the week and didn't do my usual "build with clang".

But this is broken with my normal clang config, and I get a very unhelpful error message:

Absolute reference to symbol '.rodata' not permitted in .head.text

and I have no idea why, or where it comes from. And that error message doesn't specify enough information for me to even *guess* at what's going on.

And bisecting it obviously just points at faf0ed487415 ("x86/boot:
Reject absolute references in .head.text"), since it's just a random new check for an old issue.

New random rule, with new random error as a result, and totally opaque to anybody else than Ard.

Useless crap, in other words.

Why isn't the fix a revert? Because that error message is really bad. It needs to tell me where things went wrong, not just a "You're effed. Ha! Ha!".

Quoted 6 lines

And to add insult to injury, all of this is done in-place on the vmlinux file, so when it all fails, make does

make[2]: *** Deleting file 'vmlinux'

and doesn't even leave behind anything to look at.

Sorry about that - I was really hoping we'd never see this arguably maximally passive-aggressive message and build failure, as the patch has lived in -next for 1.5 months, but here we go ...

Quoted 6 lines

Anyway, that check needs to either

(a) die a painful death very quickly

(b) be made to actually print out useful information of WHERE the
relocation comes from and WHERE it points to

I'd go for (a)+(b): I've queued up a revert which I'll send to you in a few hours, so (b) can be done properly. Also feel free to queue up the revert directly:

I'd even argue that (b) should first be a non-fatal and 100% helpful message in a standard build warning format so that CI bots pick it up, and then in a kernel release or two we can do (c) that does the hard build failure ...

Thanks,

Ingo

===========================>
From: Ingo Molnar <[email protected]>
Date: Mon, 27 Jan 2025 11:08:14 +0100
Subject: [PATCH] Revert "x86/boot: Reject absolute references in .head.text"

This reverts commit faf0ed487415f76fe4acf7980ce360901f5e1698.

As Linus reported, the hard build failure is entirely unhelpful in tracking down the bug:

Absolute reference to symbol '.rodata' not permitted in .head.text

... and to add insult to injury, the offending vmlinux gets deleted, making it hard to figure out what's going on ...

So revert this until a (much) more developer-friendly version is merged.

---
 arch/x86/tools/relocs.c | 8 +-------
 1 file changed, 1 insertion(+), 7 deletions(-)
arch/x86/tools/relocs.c +1 -7
index e937be979ec8..27441e5863b2 100644
@@ -841,10 +841,10 @@ static int is_percpu_sym(ElfW(Sym) *sym, const char *symname)
 static int do_reloc64(struct section *sec, Elf_Rel *rel, ElfW(Sym) *sym,
 		      const char *symname)
 {
-	int headtext = !strcmp(sec_name(sec->shdr.sh_info), ".head.text");
 	unsigned r_type = ELF64_R_TYPE(rel->r_info);
 	ElfW(Addr) offset = rel->r_offset;
 	int shn_abs = (sym->st_shndx == SHN_ABS) && !is_reloc(S_REL, symname);
+
 	if (sym->st_shndx == SHN_UNDEF)
 		return 0;
 
@@ -900,12 +900,6 @@ static int do_reloc64(struct section *sec, Elf_Rel *rel, ElfW(Sym) *sym,
 			break;
 		}
 
-		if (headtext) {
-			die("Absolute reference to symbol '%s' not permitted in .head.text\n",
-			    symname);
-			break;
-		}
-
 		/*
 		 * Relocation offsets for 64 bit kernels are output
 		 * as 32 bits and sign extended back to 64 bits when
#8 lore raw

[GIT PULL] x86 fix

Linus,

Please pull the latest x86/urgent Git tree from:

git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git x86-urgent-2025-01-27

# HEAD: fb1102a09a3b6710ac1e033690f0283696dc94dc Revert "x86/boot: Reject absolute references in .head.text"

Fix a build regression on certain config and build environment combinations that generate absolute references to symbols.

Thanks,

Ingo

------------------>
Ingo Molnar (1):
Revert "x86/boot: Reject absolute references in .head.text"

 arch/x86/tools/relocs.c | 8 +-------
 1 file changed, 1 insertion(+), 7 deletions(-)
arch/x86/tools/relocs.c +1 -7
index e937be979ec8..27441e5863b2 100644
@@ -841,10 +841,10 @@ static int is_percpu_sym(ElfW(Sym) *sym, const char *symname)
 static int do_reloc64(struct section *sec, Elf_Rel *rel, ElfW(Sym) *sym,
 		      const char *symname)
 {
-	int headtext = !strcmp(sec_name(sec->shdr.sh_info), ".head.text");
 	unsigned r_type = ELF64_R_TYPE(rel->r_info);
 	ElfW(Addr) offset = rel->r_offset;
 	int shn_abs = (sym->st_shndx == SHN_ABS) && !is_reloc(S_REL, symname);
+
 	if (sym->st_shndx == SHN_UNDEF)
 		return 0;
 
@@ -900,12 +900,6 @@ static int do_reloc64(struct section *sec, Elf_Rel *rel, ElfW(Sym) *sym,
 			break;
 		}
 
-		if (headtext) {
-			die("Absolute reference to symbol '%s' not permitted in .head.text\n",
-			    symname);
-			break;
-		}
-
 		/*
 		 * Relocation offsets for 64 bit kernels are output
 		 * as 32 bits and sign extended back to 64 bits when
#9 lore raw

Re: [GIT PULL] x86 fix

Quoted 3 lines · Ingo Molnar

Fix a build regression on certain config and build environment combinations that generate absolute references to symbols.

I'll hold off on this, because it looks like Ard is very close to have this fixed (and also made it non-fatal in the process).

Linus

#10 lore raw

[GIT PULL] x86 fix

Quoted 7 lines · Linus Torvalds
Quoted 3 lines · Ingo Molnar

Fix a build regression on certain config and build environment combinations that generate absolute references to symbols.

I'll hold off on this, because it looks like Ard is very close to have this fixed (and also made it non-fatal in the process).

Makes sense - meanwhile while Ard is iterating his build-warning patch here's the SEV fix that should address the build failure you've reported.

Please pull the latest x86/urgent Git tree from:

git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git x86-urgent-2025-01-28

# HEAD: 1105ab42a84bc11c62597005f78ccad2434fbd66 x86/sev: Disable jump tables in SEV startup code

Fix a build regression on certain config and build environment combinations that generate absolute references to symbols.

Thanks,

Ingo

------------------>
Ard Biesheuvel (1):
x86/sev: Disable jump tables in SEV startup code

 arch/x86/coco/sev/Makefile | 4 ++++
 1 file changed, 4 insertions(+)
arch/x86/coco/sev/Makefile +4 -0
index 08de37559307..dcb06dc8b5ae 100644
@@ -2,6 +2,10 @@
 
 obj-y += core.o
 
+# jump tables are emitted using absolute references in non-PIC code
+# so they cannot be used in the early SEV startup code
+CFLAGS_core.o += -fno-jump-tables
+
 ifdef CONFIG_FUNCTION_TRACER
 CFLAGS_REMOVE_core.o = -pg
 endif
#11 lore raw

Re: [GIT PULL] x86 fix

Quoted 15 lines · Ingo Molnar
Quoted 7 lines · Linus Torvalds
Quoted 3 lines · Ingo Molnar

Fix a build regression on certain config and build environment combinations that generate absolute references to symbols.

I'll hold off on this, because it looks like Ard is very close to have this fixed (and also made it non-fatal in the process).

Makes sense - meanwhile while Ard is iterating his build-warning patch here's the SEV fix that should address the build failure you've reported.

Quoted 8 lines

Please pull the latest x86/urgent Git tree from:

git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git x86-urgent-2025-01-28

# HEAD: 1105ab42a84bc11c62597005f78ccad2434fbd66 x86/sev: Disable jump tables in SEV startup code

Fix a build regression on certain config and build environment combinations that generate absolute references to symbols.

Erm, I phrased this poorly: the primary fix is that of a potential early boot crash in SEV-SNP guests - the wider build error was the canary that highlighted the SEV boot code bug that was benign in your specific environment.

Thanks,

Ingo

merge bot Merged
#12 lore raw

Re: [GIT PULL] x86 fix

In reply to Ingo Molnar
The pull request you sent on Tue, 28 Jan 2025 23:19:45 +0100:

The pull request you sent on Tue, 28 Jan 2025 23:19:45 +0100:

Quoted 1 line

git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git x86-urgent-2025-01-28

has been merged into torvalds/linux.git:
https://git.kernel.org/torvalds/c/05dbaf8dd8bf537d4b4eb3115ab42a5fb40ff1f5

Thank you!

Signature · 2 lines
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/prtracker.html
merge bot Applied
#13 lore raw

[tip: x86/urgent] Revert "x86/boot: Reject absolute references in .head.text"

In reply to Linus Torvalds
The following commit has been merged into the x86/urgent branch of tip:

The following commit has been merged into the x86/urgent branch of tip:

Commit-ID:     fb1102a09a3b6710ac1e033690f0283696dc94dc
Gitweb:        https://git.kernel.org/tip/fb1102a09a3b6710ac1e033690f0283696dc94dc
Author:        Ingo Molnar <[email protected]>
AuthorDate:    Mon, 27 Jan 2025 11:08:14 +01:00
Committer:     Ingo Molnar <[email protected]>
CommitterDate: Mon, 27 Jan 2025 11:08:14 +01:00

Revert "x86/boot: Reject absolute references in .head.text"

This reverts commit faf0ed487415f76fe4acf7980ce360901f5e1698.

As Linus reported, the hard build failure is entirely unhelpful in tracking down the bug:

Absolute reference to symbol '.rodata' not permitted in .head.text

... and to add insult to injury, the offending vmlinux gets deleted, making it hard to figure out what's going on ...

So revert this until a (much) more developer-friendly version is merged.

---
 arch/x86/tools/relocs.c | 8 +-------
 1 file changed, 1 insertion(+), 7 deletions(-)
arch/x86/tools/relocs.c +1 -7
index e937be9..27441e5 100644
@@ -841,10 +841,10 @@ static int is_percpu_sym(ElfW(Sym) *sym, const char *symname)
 static int do_reloc64(struct section *sec, Elf_Rel *rel, ElfW(Sym) *sym,
 		      const char *symname)
 {
-	int headtext = !strcmp(sec_name(sec->shdr.sh_info), ".head.text");
 	unsigned r_type = ELF64_R_TYPE(rel->r_info);
 	ElfW(Addr) offset = rel->r_offset;
 	int shn_abs = (sym->st_shndx == SHN_ABS) && !is_reloc(S_REL, symname);
+
 	if (sym->st_shndx == SHN_UNDEF)
 		return 0;
 
@@ -900,12 +900,6 @@ static int do_reloc64(struct section *sec, Elf_Rel *rel, ElfW(Sym) *sym,
 			break;
 		}
 
-		if (headtext) {
-			die("Absolute reference to symbol '%s' not permitted in .head.text\n",
-			    symname);
-			break;
-		}
-
 		/*
 		 * Relocation offsets for 64 bit kernels are output
 		 * as 32 bits and sign extended back to 64 bits when
#14 lore raw

Re: [GIT PULL] x86/boot enhancements for v6.14

In reply to Linus Torvalds

(cc Nathan)

Quoted 14 lines · Linus Torvalds
Quoted 11 lines · Ingo Molnar

- A series to remove the last remaining absolute symbol references from .head.text, and enforce this at build time, by Ard Biesheuvel:
[...]
- Which build-time enforcement uncovered a handful of bugs of essentially non-working code, and a wrokaround for a toolchain bug, fixed by Ard Biesheuvel as well:

- Fix spurious undefined reference when CONFIG_X86_5LEVEL=n, on GCC-12
- Disable UBSAN on SEV code that may execute very early
- Disable ftrace branch profiling in SEV startup code

Quoted 8 lines

Bah. I only noticed this today, because I was on the road part of the week and didn't do my usual "build with clang".

But this is broken with my normal clang config, and I get a very unhelpful error message:

Absolute reference to symbol '.rodata' not permitted in .head.text

Agreed. I'll change this to section+addend if the symbol is STT_SECTION, and provide the offset into .head.text as well.

...

Quoted 10 lines

Anyway, not know what the right thing to do is, I hacked up the makefiles to squirrel off a copy of the vmlinux file, and did

        objdump --no-addresses --no-show-raw-insn \
                -j .head.text --disassemble \
                -rR ORIGINAL  | less -S

and sure enough, it shows things like this:

<snp_cpuid>:

Quoted 15 lines
        push   %rbp
        push   %r15
        push   %r14
        push   %rbx
        sub    $0x18,%rsp
        lea    0x5ac8fb(%rip),%r8        # <cpuid_table_copy>
                        R_X86_64_PC32   .rodata+0x4f1758
        mov    (%r8),%eax
        test   %eax,%eax
    ..
        lea    0x5ac841(%rip),%rax        # <cpuid_std_range_max>
                        R_X86_64_PC32   .rodata+0x4f174c
    ..
        jmp    *-0x7dfffe90(,%r9,8)
                        R_X86_64_32S    .rodata+0x170

So this should be the culprit - all the other references are RIP-relative so those are fine.

I guess you are disabling retpolines and IBT in your Clang config? This looks like the switch() in snp_cpuid_postprocess() being emitted as a jump table.

This is another example of a pattern that is simply broken and guaranteed to fail when booting this kernel as a SEV-SNP guest. So I hope we agree that these issues should be detected at build time, and it is only the quality of the diagnostic message that you are objecting to?

Quoted 7 lines

Anyway, that check needs to either

(a) die a painful death very quickly

(b) be made to actually print out useful information of WHERE the
relocation comes from and WHERE it points to

Quoted 7 lines

because the current implementation of that check is not acceptable.

The next time I have to play makefile games and then do objdump by hand to figure out what the %^$% the build is complaining about, I'm just reverting it outright and not writing this long explanation of the problem.

Understood.

#15 lore raw

Re: [GIT PULL] x86/boot enhancements for v6.14

Quoted 5 lines · Ard Biesheuvel

This is another example of a pattern that is simply broken and guaranteed to fail when booting this kernel as a SEV-SNP guest. So I hope we agree that these issues should be detected at build time, and it is only the quality of the diagnostic message that you are objecting to?

And I believe the fact that they are fatal messages is a problem - we should initially just emit (informative) warnings.

Note that in the above case 'SEV-SNP guest' is basically support for a niche usecase built into the kernel image, and in the vast majority of cases the kernel will boot and work just fine. We should not fail the build for breakage in a niche usecase, at least initially.

Thanks,

Ingo