From: Ard Biesheuvel <ardb+git@google.com>
To: linux-kernel@vger.kernel.org
Cc: Ard Biesheuvel <ardb@kernel.org>,
x86@kernel.org, "H. Peter Anvin" <hpa@zytor.com>,
Andy Lutomirski <luto@kernel.org>,
Peter Zijlstra <peterz@infradead.org>,
Uros Bizjak <ubizjak@gmail.com>, Dennis Zhou <dennis@kernel.org>,
Tejun Heo <tj@kernel.org>, Christoph Lameter <cl@linux.com>,
Mathieu Desnoyers <mathieu.desnoyers@efficios.com>,
Paolo Bonzini <pbonzini@redhat.com>,
Vitaly Kuznetsov <vkuznets@redhat.com>,
Juergen Gross <jgross@suse.com>,
Boris Ostrovsky <boris.ostrovsky@oracle.com>,
Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
Arnd Bergmann <arnd@arndb.de>,
Masahiro Yamada <masahiroy@kernel.org>,
Kees Cook <kees@kernel.org>,
Nathan Chancellor <nathan@kernel.org>,
Keith Packard <keithp@keithp.com>,
Justin Stitt <justinstitt@google.com>,
Josh Poimboeuf <jpoimboe@kernel.org>,
Arnaldo Carvalho de Melo <acme@kernel.org>,
Namhyung Kim <namhyung@kernel.org>, Jiri Olsa <jolsa@kernel.org>,
Ian Rogers <irogers@google.com>,
Adrian Hunter <adrian.hunter@intel.com>,
Kan Liang <kan.liang@linux.intel.com>,
linux-doc@vger.kernel.org, linux-pm@vger.kernel.org,
kvm@vger.kernel.org, xen-devel@lists.xenproject.org,
linux-efi@vger.kernel.org, linux-arch@vger.kernel.org,
linux-sparse@vger.kernel.org, linux-kbuild@vger.kernel.org,
linux-perf-users@vger.kernel.org,
rust-for-linux@vger.kernel.org, llvm@lists.linux.dev
Subject: [RFC PATCH 09/28] x86/tools: Remove special relocation handling for per-CPU variables
Date: Wed, 25 Sep 2024 17:01:09 +0200 [thread overview]
Message-ID: <20240925150059.3955569-39-ardb+git@google.com> (raw)
In-Reply-To: <20240925150059.3955569-30-ardb+git@google.com>
From: Ard Biesheuvel <ardb@kernel.org>
Due to the placement of per-CPU variables in a special, 0x0 based
disjoint memory segment in the ELF binary, the KASLR relocation tool
needed to perform special processing for references to such variables,
as they were not affected by KASLR displacement.
This meant that absolute references could be ignored, and RIP-relative
references had to be compensated for KASLR, by applying the same offset
but negated.
None of this is necessary any longer, so remove this handling from the
relocation host tool.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
---
arch/x86/boot/compressed/misc.c | 14 +--
arch/x86/tools/relocs.c | 130 +-------------------
2 files changed, 2 insertions(+), 142 deletions(-)
diff --git a/arch/x86/boot/compressed/misc.c b/arch/x86/boot/compressed/misc.c
index 04a35b2c26e9..89f01375cdb7 100644
--- a/arch/x86/boot/compressed/misc.c
+++ b/arch/x86/boot/compressed/misc.c
@@ -235,7 +235,7 @@ static void handle_relocations(void *output, unsigned long output_len,
/*
* Process relocations: 32 bit relocations first then 64 bit after.
- * Three sets of binary relocations are added to the end of the kernel
+ * Two sets of binary relocations are added to the end of the kernel
* before compression. Each relocation table entry is the kernel
* address of the location which needs to be updated stored as a
* 32-bit value which is sign extended to 64 bits.
@@ -245,8 +245,6 @@ static void handle_relocations(void *output, unsigned long output_len,
* kernel bits...
* 0 - zero terminator for 64 bit relocations
* 64 bit relocation repeated
- * 0 - zero terminator for inverse 32 bit relocations
- * 32 bit inverse relocation repeated
* 0 - zero terminator for 32 bit relocations
* 32 bit relocation repeated
*
@@ -267,16 +265,6 @@ static void handle_relocations(void *output, unsigned long output_len,
long extended = *reloc;
extended += map;
- ptr = (unsigned long)extended;
- if (ptr < min_addr || ptr > max_addr)
- error("inverse 32-bit relocation outside of kernel!\n");
-
- *(int32_t *)ptr -= delta;
- }
- for (reloc--; *reloc; reloc--) {
- long extended = *reloc;
- extended += map;
-
ptr = (unsigned long)extended;
if (ptr < min_addr || ptr > max_addr)
error("64-bit relocation outside of kernel!\n");
diff --git a/arch/x86/tools/relocs.c b/arch/x86/tools/relocs.c
index 10add45b99f1..942c029a5067 100644
--- a/arch/x86/tools/relocs.c
+++ b/arch/x86/tools/relocs.c
@@ -29,7 +29,6 @@ static struct relocs relocs16;
static struct relocs relocs32;
#if ELF_BITS == 64
-static struct relocs relocs32neg;
static struct relocs relocs64;
# define FMT PRIu64
#else
@@ -287,34 +286,6 @@ static const char *sym_name(const char *sym_strtab, Elf_Sym *sym)
return name;
}
-static Elf_Sym *sym_lookup(const char *symname)
-{
- int i;
-
- for (i = 0; i < shnum; i++) {
- struct section *sec = &secs[i];
- long nsyms;
- const char *strtab;
- Elf_Sym *symtab;
- Elf_Sym *sym;
-
- if (sec->shdr.sh_type != SHT_SYMTAB)
- continue;
-
- nsyms = sec->shdr.sh_size/sizeof(Elf_Sym);
- symtab = sec->symtab;
- strtab = sec->link->strtab;
-
- for (sym = symtab; --nsyms >= 0; sym++) {
- if (!sym->st_name)
- continue;
- if (strcmp(symname, strtab + sym->st_name) == 0)
- return sym;
- }
- }
- return 0;
-}
-
#if BYTE_ORDER == LITTLE_ENDIAN
# define le16_to_cpu(val) (val)
# define le32_to_cpu(val) (val)
@@ -722,79 +693,8 @@ static void walk_relocs(int (*process)(struct section *sec, Elf_Rel *rel,
}
}
-/*
- * The .data..percpu section is a special case for x86_64 SMP kernels.
- * It is used to initialize the actual per_cpu areas and to provide
- * definitions for the per_cpu variables that correspond to their offsets
- * within the percpu area. Since the values of all of the symbols need
- * to be offsets from the start of the per_cpu area the virtual address
- * (sh_addr) of .data..percpu is 0 in SMP kernels.
- *
- * This means that:
- *
- * Relocations that reference symbols in the per_cpu area do not
- * need further relocation (since the value is an offset relative
- * to the start of the per_cpu area that does not change).
- *
- * Relocations that apply to the per_cpu area need to have their
- * offset adjusted by by the value of __per_cpu_load to make them
- * point to the correct place in the loaded image (because the
- * virtual address of .data..percpu is 0).
- *
- * For non SMP kernels .data..percpu is linked as part of the normal
- * kernel data and does not require special treatment.
- *
- */
-static int per_cpu_shndx = -1;
-static Elf_Addr per_cpu_load_addr;
-
-static void percpu_init(void)
-{
- int i;
-
- for (i = 0; i < shnum; i++) {
- ElfW(Sym) *sym;
-
- if (strcmp(sec_name(i), ".data..percpu"))
- continue;
-
- if (secs[i].shdr.sh_addr != 0) /* non SMP kernel */
- return;
-
- sym = sym_lookup("__per_cpu_load");
- if (!sym)
- die("can't find __per_cpu_load\n");
-
- per_cpu_shndx = i;
- per_cpu_load_addr = sym->st_value;
-
- return;
- }
-}
-
#if ELF_BITS == 64
-/*
- * Check to see if a symbol lies in the .data..percpu section.
- *
- * The linker incorrectly associates some symbols with the
- * .data..percpu section so we also need to check the symbol
- * name to make sure that we classify the symbol correctly.
- *
- * The GNU linker incorrectly associates:
- * __init_begin
- * __per_cpu_load
- */
-static int is_percpu_sym(ElfW(Sym) *sym, const char *symname)
-{
- int shndx = sym_index(sym);
-
- return (shndx == per_cpu_shndx) &&
- strcmp(symname, "__init_begin") &&
- strcmp(symname, "__per_cpu_load");
-}
-
-
static int do_reloc64(struct section *sec, Elf_Rel *rel, ElfW(Sym) *sym,
const char *symname)
{
@@ -805,12 +705,6 @@ static int do_reloc64(struct section *sec, Elf_Rel *rel, ElfW(Sym) *sym,
if (sym->st_shndx == SHN_UNDEF)
return 0;
- /*
- * Adjust the offset if this reloc applies to the percpu section.
- */
- if (sec->shdr.sh_info == per_cpu_shndx)
- offset += per_cpu_load_addr;
-
switch (r_type) {
case R_X86_64_NONE:
/* NONE can be ignored. */
@@ -819,33 +713,22 @@ static int do_reloc64(struct section *sec, Elf_Rel *rel, ElfW(Sym) *sym,
case R_X86_64_PC32:
case R_X86_64_PLT32:
/*
- * PC relative relocations don't need to be adjusted unless
- * referencing a percpu symbol.
+ * PC relative relocations don't need to be adjusted.
*
* NB: R_X86_64_PLT32 can be treated as R_X86_64_PC32.
*/
- if (is_percpu_sym(sym, symname))
- add_reloc(&relocs32neg, offset);
break;
case R_X86_64_PC64:
/*
* Only used by jump labels
*/
- if (is_percpu_sym(sym, symname))
- die("Invalid R_X86_64_PC64 relocation against per-CPU symbol %s\n", symname);
break;
case R_X86_64_32:
case R_X86_64_32S:
case R_X86_64_64:
case R_X86_64_GOTPCREL:
- /*
- * References to the percpu area don't need to be adjusted.
- */
- if (is_percpu_sym(sym, symname))
- break;
-
if (shn_abs) {
/*
* Whitelisted absolute symbols do not require
@@ -1076,7 +959,6 @@ static void emit_relocs(int as_text, int use_real_mode)
/* Order the relocations for more efficient processing */
sort_relocs(&relocs32);
#if ELF_BITS == 64
- sort_relocs(&relocs32neg);
sort_relocs(&relocs64);
#else
sort_relocs(&relocs16);
@@ -1109,13 +991,6 @@ static void emit_relocs(int as_text, int use_real_mode)
for (i = 0; i < relocs64.count; i++)
if (!i || relocs64.offset[i] != relocs64.offset[i - 1])
write_reloc(relocs64.offset[i], stdout);
-
- /* Print a stop */
- write_reloc(0, stdout);
-
- /* Now print each inverse 32-bit relocation */
- for (i = 0; i < relocs32neg.count; i++)
- write_reloc(relocs32neg.offset[i], stdout);
#endif
/* Print a stop */
@@ -1180,9 +1055,6 @@ void process(FILE *fp, int use_real_mode, int as_text,
read_symtabs();
read_relocs();
- if (ELF_BITS == 64)
- percpu_init();
-
if (show_absolute_syms) {
print_absolute_symbols();
return;
--
2.46.0.792.g87dc391469-goog
next prev parent reply other threads:[~2024-09-25 15:02 UTC|newest]
Thread overview: 73+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-09-25 15:01 [RFC PATCH 00/28] x86: Rely on toolchain for relocatable code Ard Biesheuvel
2024-09-25 15:01 ` [RFC PATCH 01/28] x86/pvh: Call C code via the kernel virtual mapping Ard Biesheuvel
2024-09-25 21:12 ` Jason Andryuk
2024-09-25 15:01 ` [RFC PATCH 02/28] Documentation: Bump minimum GCC version to 8.1 Ard Biesheuvel
2024-09-25 15:58 ` Arnd Bergmann
2024-12-19 11:53 ` Mark Rutland
2024-12-19 12:02 ` Arnd Bergmann
2024-09-26 21:35 ` Miguel Ojeda
2024-09-27 16:22 ` Mark Rutland
2024-09-25 15:01 ` [RFC PATCH 03/28] x86/tools: Use mmap() to simplify relocs host tool Ard Biesheuvel
2024-09-25 15:01 ` [RFC PATCH 04/28] x86/boot: Permit GOTPCREL relocations for x86_64 builds Ard Biesheuvel
2024-10-01 5:33 ` Josh Poimboeuf
2024-10-01 6:56 ` Ard Biesheuvel
2024-09-25 15:01 ` [RFC PATCH 05/28] x86: Define the stack protector guard symbol explicitly Ard Biesheuvel
2024-09-25 15:53 ` Ian Rogers
2024-09-25 17:43 ` Ard Biesheuvel
2024-09-25 17:48 ` Ian Rogers
2024-09-25 18:32 ` Uros Bizjak
2024-09-28 13:41 ` Brian Gerst
2024-10-04 13:15 ` Ard Biesheuvel
2024-10-08 14:36 ` Brian Gerst
2024-10-04 10:01 ` Uros Bizjak
2024-09-25 15:01 ` [RFC PATCH 06/28] x86/percpu: Get rid of absolute per-CPU variable placement Ard Biesheuvel
2024-09-25 17:56 ` Christoph Lameter (Ampere)
2024-09-25 15:01 ` [RFC PATCH 07/28] scripts/kallsyms: Avoid 0x0 as the relative base Ard Biesheuvel
2024-09-25 15:01 ` [RFC PATCH 08/28] scripts/kallsyms: Remove support for absolute per-CPU variables Ard Biesheuvel
2024-09-25 15:01 ` Ard Biesheuvel [this message]
2024-09-25 15:01 ` [RFC PATCH 10/28] x86/xen: Avoid relocatable quantities in Xen ELF notes Ard Biesheuvel
2024-09-25 15:01 ` [RFC PATCH 11/28] x86/pvh: Avoid absolute symbol references in .head.text Ard Biesheuvel
2024-09-25 21:10 ` Jason Andryuk
2024-09-25 21:50 ` Ard Biesheuvel
2024-09-25 22:40 ` Jason Andryuk
2024-09-25 15:01 ` [RFC PATCH 12/28] x86/pm-trace: Use RIP-relative accesses for .tracedata Ard Biesheuvel
2024-09-25 15:01 ` [RFC PATCH 13/28] x86/kvm: Use RIP-relative addressing Ard Biesheuvel
2024-09-25 15:01 ` [RFC PATCH 14/28] x86/rethook: Use RIP-relative reference for return address Ard Biesheuvel
2024-09-25 16:39 ` Linus Torvalds
2024-09-25 16:45 ` Ard Biesheuvel
2024-09-25 15:01 ` [RFC PATCH 15/28] x86/sync_core: Use RIP-relative addressing Ard Biesheuvel
2024-09-25 15:01 ` [RFC PATCH 16/28] x86/entry_64: " Ard Biesheuvel
2024-09-25 15:01 ` [RFC PATCH 17/28] x86/hibernate: Prefer RIP-relative accesses Ard Biesheuvel
2024-09-25 15:01 ` [RFC PATCH 18/28] x86/boot/64: Determine VA/PA offset before entering C code Ard Biesheuvel
2024-09-25 15:01 ` [RFC PATCH 19/28] x86/boot/64: Avoid intentional absolute symbol references in .head.text Ard Biesheuvel
2024-09-25 15:01 ` [RFC PATCH 20/28] x64/acpi: Use PIC-compatible references in wakeup_64.S Ard Biesheuvel
2024-09-25 15:01 ` [RFC PATCH 21/28] x86/head: Use PIC-compatible symbol references in startup code Ard Biesheuvel
2024-09-25 15:01 ` [RFC PATCH 22/28] asm-generic: Treat PIC .data.rel.ro sections as .rodata Ard Biesheuvel
2024-09-25 15:01 ` [RFC PATCH 23/28] tools/objtool: Mark generated sections as writable Ard Biesheuvel
2024-09-25 15:01 ` [RFC PATCH 24/28] tools/objtool: Treat indirect ftrace calls as direct calls Ard Biesheuvel
2024-10-01 7:18 ` Josh Poimboeuf
2024-10-01 7:39 ` Ard Biesheuvel
2024-09-25 15:01 ` [RFC PATCH 25/28] x86: Use PIE codegen for the core kernel Ard Biesheuvel
2024-10-01 21:13 ` H. Peter Anvin
2024-10-02 15:25 ` Ard Biesheuvel
2024-10-02 20:01 ` Linus Torvalds
2024-10-03 11:13 ` Ard Biesheuvel
2024-10-04 21:06 ` H. Peter Anvin
2024-10-05 8:31 ` Uros Bizjak
2024-10-05 23:36 ` H. Peter Anvin
2024-10-06 0:00 ` Linus Torvalds
2024-10-06 8:06 ` Uros Bizjak
2024-10-06 7:59 ` Uros Bizjak
2024-10-06 18:00 ` David Laight
2024-10-06 19:17 ` Uros Bizjak
2024-10-06 19:38 ` H. Peter Anvin
2024-09-25 15:01 ` [RFC PATCH 26/28] x86/boot: Implement support for ELF RELA/RELR relocations Ard Biesheuvel
2024-09-25 15:01 ` [RFC PATCH 27/28] x86/kernel: Switch to PIE linking for the core kernel Ard Biesheuvel
2024-09-25 18:54 ` Uros Bizjak
2024-09-25 19:14 ` Ard Biesheuvel
2024-09-25 19:39 ` Uros Bizjak
2024-09-25 20:01 ` Ard Biesheuvel
2024-09-25 20:22 ` Uros Bizjak
2024-09-25 20:24 ` Vegard Nossum
2024-09-26 13:38 ` Ard Biesheuvel
2024-09-25 15:01 ` [RFC PATCH 28/28] x86/tools: Drop x86_64 support from 'relocs' tool 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-all 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=20240925150059.3955569-39-ardb+git@google.com \
--to=ardb+git@google.com \
--cc=acme@kernel.org \
--cc=adrian.hunter@intel.com \
--cc=ardb@kernel.org \
--cc=arnd@arndb.de \
--cc=boris.ostrovsky@oracle.com \
--cc=cl@linux.com \
--cc=dennis@kernel.org \
--cc=gregkh@linuxfoundation.org \
--cc=hpa@zytor.com \
--cc=irogers@google.com \
--cc=jgross@suse.com \
--cc=jolsa@kernel.org \
--cc=jpoimboe@kernel.org \
--cc=justinstitt@google.com \
--cc=kan.liang@linux.intel.com \
--cc=kees@kernel.org \
--cc=keithp@keithp.com \
--cc=kvm@vger.kernel.org \
--cc=linux-arch@vger.kernel.org \
--cc=linux-doc@vger.kernel.org \
--cc=linux-efi@vger.kernel.org \
--cc=linux-kbuild@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-perf-users@vger.kernel.org \
--cc=linux-pm@vger.kernel.org \
--cc=linux-sparse@vger.kernel.org \
--cc=llvm@lists.linux.dev \
--cc=luto@kernel.org \
--cc=masahiroy@kernel.org \
--cc=mathieu.desnoyers@efficios.com \
--cc=namhyung@kernel.org \
--cc=nathan@kernel.org \
--cc=pbonzini@redhat.com \
--cc=peterz@infradead.org \
--cc=rust-for-linux@vger.kernel.org \
--cc=tj@kernel.org \
--cc=ubizjak@gmail.com \
--cc=vkuznets@redhat.com \
--cc=x86@kernel.org \
--cc=xen-devel@lists.xenproject.org \
/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;
as well as URLs for NNTP newsgroup(s).