mirror of
https://github.com/torvalds/linux.git
synced 2024-11-25 05:32:00 +00:00
d2a793dae2
Instead of making increasingly complicated ALTERNATIVE_n() implementations, use a nested alternative expression. The only difference between: ALTERNATIVE_2(oldinst, newinst1, flag1, newinst2, flag2) and ALTERNATIVE(ALTERNATIVE(oldinst, newinst1, flag1), newinst2, flag2) is that the outer alternative can add additional padding when the inner alternative is the shorter one, which then results in alt_instr::instrlen being inconsistent. However, this is easily remedied since the alt_instr entries will be consecutive and it is trivial to compute the max(alt_instr::instrlen) at runtime while patching. Specifically, after this the ALTERNATIVE_2 macro, after CPP expansion (and manual layout), looks like this: .macro ALTERNATIVE_2 oldinstr, newinstr1, ft_flags1, newinstr2, ft_flags2 740: 740: \oldinstr ; 741: .skip -(((744f-743f)-(741b-740b)) > 0) * ((744f-743f)-(741b-740b)),0x90 ; 742: .pushsection .altinstructions,"a" ; altinstr_entry 740b,743f,\ft_flags1,742b-740b,744f-743f ; .popsection ; .pushsection .altinstr_replacement,"ax" ; 743: \newinstr1 ; 744: .popsection ; ; 741: .skip -(((744f-743f)-(741b-740b)) > 0) * ((744f-743f)-(741b-740b)),0x90 ; 742: .pushsection .altinstructions,"a" ; altinstr_entry 740b,743f,\ft_flags2,742b-740b,744f-743f ; .popsection ; .pushsection .altinstr_replacement,"ax" ; 743: \newinstr2 ; 744: .popsection ; .endm The only label that is ambiguous is 740, however they all reference the same spot, so that doesn't matter. NOTE: obviously only @oldinstr may be an alternative; making @newinstr an alternative would mean patching .altinstr_replacement which very likely isn't what is intended, also the labels will be confused in that case. [ bp: Debug an issue where it would match the wrong two insns and and consider them nested due to the same signed offsets in the .alternative section and use instr_va() to compare the full virtual addresses instead. - Use new labels to denote that the new, nested alternatives are being used when staring at preprocessed output. - Use the %c constraint everywhere instead of %P and document the difference for future reference. ] Signed-off-by: Peter Zijlstra <peterz@infradead.org> Co-developed-by: Borislav Petkov (AMD) <bp@alien8.de> Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de> Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org> Link: https://lkml.kernel.org/r/20230628104952.GA2439977@hirez.programming.kicks-ass.net
163 lines
4.9 KiB
C
163 lines
4.9 KiB
C
// SPDX-License-Identifier: GPL-2.0-or-later
|
|
#include <string.h>
|
|
|
|
#include <objtool/special.h>
|
|
#include <objtool/builtin.h>
|
|
|
|
#define X86_FEATURE_POPCNT (4 * 32 + 23)
|
|
#define X86_FEATURE_SMAP (9 * 32 + 20)
|
|
|
|
void arch_handle_alternative(unsigned short feature, struct special_alt *alt)
|
|
{
|
|
static struct special_alt *group, *prev;
|
|
|
|
/*
|
|
* Recompute orig_len for nested ALTERNATIVE()s.
|
|
*/
|
|
if (group && group->orig_sec == alt->orig_sec &&
|
|
group->orig_off == alt->orig_off) {
|
|
|
|
struct special_alt *iter = group;
|
|
for (;;) {
|
|
unsigned int len = max(iter->orig_len, alt->orig_len);
|
|
iter->orig_len = alt->orig_len = len;
|
|
|
|
if (iter == prev)
|
|
break;
|
|
|
|
iter = list_next_entry(iter, list);
|
|
}
|
|
|
|
} else group = alt;
|
|
|
|
prev = alt;
|
|
|
|
switch (feature) {
|
|
case X86_FEATURE_SMAP:
|
|
/*
|
|
* If UACCESS validation is enabled; force that alternative;
|
|
* otherwise force it the other way.
|
|
*
|
|
* What we want to avoid is having both the original and the
|
|
* alternative code flow at the same time, in that case we can
|
|
* find paths that see the STAC but take the NOP instead of
|
|
* CLAC and the other way around.
|
|
*/
|
|
if (opts.uaccess)
|
|
alt->skip_orig = true;
|
|
else
|
|
alt->skip_alt = true;
|
|
break;
|
|
case X86_FEATURE_POPCNT:
|
|
/*
|
|
* It has been requested that we don't validate the !POPCNT
|
|
* feature path which is a "very very small percentage of
|
|
* machines".
|
|
*/
|
|
alt->skip_orig = true;
|
|
break;
|
|
default:
|
|
break;
|
|
}
|
|
}
|
|
|
|
bool arch_support_alt_relocation(struct special_alt *special_alt,
|
|
struct instruction *insn,
|
|
struct reloc *reloc)
|
|
{
|
|
return true;
|
|
}
|
|
|
|
/*
|
|
* There are 3 basic jump table patterns:
|
|
*
|
|
* 1. jmpq *[rodata addr](,%reg,8)
|
|
*
|
|
* This is the most common case by far. It jumps to an address in a simple
|
|
* jump table which is stored in .rodata.
|
|
*
|
|
* 2. jmpq *[rodata addr](%rip)
|
|
*
|
|
* This is caused by a rare GCC quirk, currently only seen in three driver
|
|
* functions in the kernel, only with certain obscure non-distro configs.
|
|
*
|
|
* As part of an optimization, GCC makes a copy of an existing switch jump
|
|
* table, modifies it, and then hard-codes the jump (albeit with an indirect
|
|
* jump) to use a single entry in the table. The rest of the jump table and
|
|
* some of its jump targets remain as dead code.
|
|
*
|
|
* In such a case we can just crudely ignore all unreachable instruction
|
|
* warnings for the entire object file. Ideally we would just ignore them
|
|
* for the function, but that would require redesigning the code quite a
|
|
* bit. And honestly that's just not worth doing: unreachable instruction
|
|
* warnings are of questionable value anyway, and this is such a rare issue.
|
|
*
|
|
* 3. mov [rodata addr],%reg1
|
|
* ... some instructions ...
|
|
* jmpq *(%reg1,%reg2,8)
|
|
*
|
|
* This is a fairly uncommon pattern which is new for GCC 6. As of this
|
|
* writing, there are 11 occurrences of it in the allmodconfig kernel.
|
|
*
|
|
* As of GCC 7 there are quite a few more of these and the 'in between' code
|
|
* is significant. Esp. with KASAN enabled some of the code between the mov
|
|
* and jmpq uses .rodata itself, which can confuse things.
|
|
*
|
|
* TODO: Once we have DWARF CFI and smarter instruction decoding logic,
|
|
* ensure the same register is used in the mov and jump instructions.
|
|
*
|
|
* NOTE: MITIGATION_RETPOLINE made it harder still to decode dynamic jumps.
|
|
*/
|
|
struct reloc *arch_find_switch_table(struct objtool_file *file,
|
|
struct instruction *insn)
|
|
{
|
|
struct reloc *text_reloc, *rodata_reloc;
|
|
struct section *table_sec;
|
|
unsigned long table_offset;
|
|
|
|
/* look for a relocation which references .rodata */
|
|
text_reloc = find_reloc_by_dest_range(file->elf, insn->sec,
|
|
insn->offset, insn->len);
|
|
if (!text_reloc || text_reloc->sym->type != STT_SECTION ||
|
|
!text_reloc->sym->sec->rodata)
|
|
return NULL;
|
|
|
|
table_offset = reloc_addend(text_reloc);
|
|
table_sec = text_reloc->sym->sec;
|
|
|
|
if (reloc_type(text_reloc) == R_X86_64_PC32)
|
|
table_offset += 4;
|
|
|
|
/*
|
|
* Make sure the .rodata address isn't associated with a
|
|
* symbol. GCC jump tables are anonymous data.
|
|
*
|
|
* Also support C jump tables which are in the same format as
|
|
* switch jump tables. For objtool to recognize them, they
|
|
* need to be placed in the C_JUMP_TABLE_SECTION section. They
|
|
* have symbols associated with them.
|
|
*/
|
|
if (find_symbol_containing(table_sec, table_offset) &&
|
|
strcmp(table_sec->name, C_JUMP_TABLE_SECTION))
|
|
return NULL;
|
|
|
|
/*
|
|
* Each table entry has a rela associated with it. The rela
|
|
* should reference text in the same function as the original
|
|
* instruction.
|
|
*/
|
|
rodata_reloc = find_reloc_by_dest(file->elf, table_sec, table_offset);
|
|
if (!rodata_reloc)
|
|
return NULL;
|
|
|
|
/*
|
|
* Use of RIP-relative switch jumps is quite rare, and
|
|
* indicates a rare GCC quirk/bug which can leave dead
|
|
* code behind.
|
|
*/
|
|
if (reloc_type(text_reloc) == R_X86_64_PC32)
|
|
file->ignore_unreachables = true;
|
|
|
|
return rodata_reloc;
|
|
}
|