[ELF] Error if a section address is smaller than image base

When using `-no-pie` without a `SECTIONS` command, the linker uses the
target's default image base. If `-Ttext=` or `--section-start` specifies
an output section address below this base, the result is likely
unintended.

- With `--no-rosegment`, the PT_LOAD segment covering the ELF header cannot include `.text` if `.text`'s address is too low, causing an `error: output file too large`.
- With default `--rosegment`:
  - If a read-only section (e.g., `.rodata`) exists, a similar `error: output file too large` occurs.
  - Without read-only sections, the PT_LOAD segment covering the ELF header and program headers includes no sections, which is unusual and likely undesired. This also causes non-ascending PT_LOAD `p_vaddr` values related to the PT_LOAD that overlaps with PT_PHDR (#138584).

To prevent these issues, report an error if a section address is below
the image base and suggest `--image-base`. This check also applies when
`--image-base` is explicitly set but is skipped when a `SECTIONS`
command is used.

Pull Request: https://github.com/llvm/llvm-project/pull/140187
This commit is contained in:
Fangrui Song 2025-05-21 09:19:47 -07:00 committed by GitHub
parent 81b81354f8
commit 0edc8b59ab
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
6 changed files with 65 additions and 19 deletions

View File

@ -1614,18 +1614,33 @@ template <class ELFT> void Writer<ELFT>::finalizeAddressDependentContent() {
for (OutputSection *sec : ctx.outputSections)
sec->addr = 0;
// If addrExpr is set, the address may not be a multiple of the alignment.
// Warn because this is error-prone.
for (SectionCommand *cmd : ctx.script->sectionCommands)
if (auto *osd = dyn_cast<OutputDesc>(cmd)) {
OutputSection *osec = &osd->osec;
if (osec->addr % osec->addralign != 0)
Warn(ctx) << "address (0x" << Twine::utohexstr(osec->addr)
<< ") of section " << osec->name
<< " is not a multiple of alignment (" << osec->addralign
<< ")";
uint64_t imageBase = ctx.script->hasSectionsCommand || ctx.arg.relocatable
? 0
: ctx.target->getImageBase();
for (SectionCommand *cmd : ctx.script->sectionCommands) {
auto *osd = dyn_cast<OutputDesc>(cmd);
if (!osd)
continue;
OutputSection *osec = &osd->osec;
// Error if the address is below the image base when SECTIONS is absent
// (e.g. when -Ttext is specified and smaller than the default target image
// base for no-pie).
if (osec->addr < imageBase && (osec->flags & SHF_ALLOC)) {
Err(ctx) << "section '" << osec->name << "' address (0x"
<< Twine::utohexstr(osec->addr)
<< ") is smaller than image base (0x"
<< Twine::utohexstr(imageBase) << "); specify --image-base";
}
// If addrExpr is set, the address may not be a multiple of the alignment.
// Warn because this is error-prone.
if (osec->addr % osec->addralign != 0)
Warn(ctx) << "address (0x" << Twine::utohexstr(osec->addr)
<< ") of section " << osec->name
<< " is not a multiple of alignment (" << osec->addralign
<< ")";
}
// Sizes are no longer allowed to grow, so all allowable spills have been
// taken. Remove any leftover potential spills.
ctx.script->erasePotentialSpillSections();

View File

@ -64,6 +64,12 @@ Breaking changes
in the same segment by default. Pass ``--xosegment`` to lld in order to get
the old behavior back.
* When using ``--no-pie`` without a ``SECTIONS`` command, the linker uses the
target's default image base. If ``-Ttext=`` or ``--section-start`` specifies
an output section address below this base, there will now be an error.
``--image-base`` can be set at a lower address to fix the error.
(`#140187 <https://github.com/llvm/llvm-project/pull/140187>`_)
COFF Improvements
-----------------

View File

@ -31,6 +31,18 @@
# CHECK-NEXT: 5 .hash 00000010 000000000000201c
# CHECK-NEXT: 6 .text 00000008 000000000000202c
# RUN: ld.lld -e 0 -o %t --script %t.script %t.o --fatal-warnings
# RUN: llvm-readelf -Sl %t | FileCheck %s --check-prefix=CHECK1
# CHECK1: Name Type Address Off Size ES Flg Lk Inf Al
# CHECK1-NEXT: NULL 0000000000000000 000000 000000 00 0 0 0
# CHECK1-NEXT: .text PROGBITS 0000000000000000 001000 000008 00 AX 0 0 4
# CHECK1-NEXT: .data PROGBITS 0000000000004000 002000 000008 00 WA 0 0 1
# CHECK1: Program Headers:
# CHECK1-NEXT: Type Offset VirtAddr PhysAddr FileSiz MemSiz Flg Align
# CHECK1-NEXT: LOAD 0x001000 0x0000000000000000 0x0000000000000000 0x000008 0x000008 R E 0x1000
# CHECK1-NEXT: LOAD 0x002000 0x0000000000004000 0x0000000000004000 0x000008 0x000008 RW 0x1000
.quad 0
.data
.quad 0

View File

@ -10,7 +10,7 @@
# RUN: llvm-readelf -S %t | FileCheck %s
## Check we don't warn in the absence of SECTIONS.
# RUN: ld.lld --fatal-warnings -Ttext=0x10000 %t.o -o /dev/null
# RUN: ld.lld --fatal-warnings -Ttext=0x10000 --image-base=0x10000 %t.o -o /dev/null
# WARN: warning: address (0x10004) of section .data.rel.ro is not a multiple of alignment (16)
# WARN: warning: address (0x20001) of section .data2 is not a multiple of alignment (8)

View File

@ -1,9 +1,15 @@
# REQUIRES: x86
# RUN: llvm-mc -filetype=obj -triple=x86_64-pc-linux %s -o %t.o
# RUN: not ld.lld %t.o --section-start .text=0x100000 \
# RUN: --section-start=.data=0x110000 --section-start .bss=0x200000
# RUN: ld.lld %t.o --section-start .text=0x100000 \
# RUN: --section-start=.data=0x110000 --section-start .bss=0x200000 -o %t
# RUN: --section-start=.data=0x110000 --section-start .bss=0x200000 --noinhibit-exec -o %t 2>&1 | \
# RUN: FileCheck %s --check-prefix=LINK --implicit-check-not=warning:
# RUN: llvm-objdump --section-headers %t | FileCheck %s
# LINK: warning: section '.text' address (0x100000) is smaller than image base (0x200000); specify --image-base
# LINK-NEXT: warning: section '.data' address (0x110000) is smaller than image base (0x200000); specify --image-base
# CHECK: Sections:
# CHECK-NEXT: Idx Name Size VMA Type
# CHECK-NEXT: 0 00000000 0000000000000000
@ -11,21 +17,26 @@
# CHECK-NEXT: 2 .data 00000004 0000000000110000 DATA
# CHECK-NEXT: 3 .bss 00000004 0000000000200000 BSS
## The same, but dropped "0x" prefix.
# RUN: ld.lld %t.o --section-start .text=100000 \
# RUN: --section-start .data=110000 --section-start .bss=0x200000 -o %t1
## The errors go away when the image base is 0.
# RUN: ld.lld %t.o -pie --section-start .text=0x100000 \
# RUN: --section-start=.data=0x110000 --section-start .bss=0x200000 -o %t 2>&1 | count 0
# RUN: llvm-objdump --section-headers %t | FileCheck %s
## The same, but dropped "0x" prefix. Specify a smaller --image-base to suppress warnings.
# RUN: ld.lld %t.o --image-base=0x90000 --section-start .text=100000 \
# RUN: --section-start .data=110000 --section-start .bss=0x200000 -o %t1 2>&1 | count 0
# RUN: llvm-objdump --section-headers %t1 | FileCheck %s
## Use -Ttext, -Tdata, -Tbss as replacement for --section-start:
# RUN: ld.lld %t.o -Ttext=0x100000 -Tdata=0x110000 -Tbss=0x200000 -o %t4
# RUN: ld.lld %t.o --image-base=0x90000 -Ttext=0x100000 -Tdata=0x110000 -Tbss=0x200000 -o %t4
# RUN: llvm-objdump --section-headers %t4 | FileCheck %s
## The same, but dropped "0x" prefix.
# RUN: ld.lld %t.o -Ttext=100000 -Tdata=110000 -Tbss=200000 -o %t5
# RUN: ld.lld %t.o --image-base=0x90000 -Ttext=100000 -Tdata=110000 -Tbss=200000 -o %t5
# RUN: llvm-objdump --section-headers %t5 | FileCheck %s
## Check form without assignment:
# RUN: ld.lld %t.o -Ttext 0x100000 -Tdata 0x110000 -Tbss 0x200000 -o %t4
# RUN: ld.lld %t.o --image-base=0x90000 -Ttext 0x100000 -Tdata 0x110000 -Tbss 0x200000 -o %t4
# RUN: llvm-objdump --section-headers %t4 | FileCheck %s
## Errors:

View File

@ -16,8 +16,10 @@
## If -Ttext is smaller than the image base (which defaults to 0x200000 for -no-pie),
## the headers will still be allocated, but mapped at a higher address,
## which may look strange.
# RUN: ld.lld -Ttext 0x0 -Tdata 0x4000 -Tbss 0x8000 %t.o -o %t2
# RUN: ld.lld -Ttext 0x0 -Tdata 0x4000 -Tbss 0x8000 %t.o --noinhibit-exec -o %t2 2>&1 | FileCheck %s --check-prefix=LINK1
# RUN: llvm-readelf -S -l %t2 | FileCheck %s --check-prefix=USER1
# LINK1: warning: section '.text' address (0x0) is smaller than image base (0x200000); specify --image-base
# LINK1-NEXT: warning: section '.data' address (0x4000) is smaller than image base (0x200000); specify --image-base
# USER1: .text PROGBITS 0000000000000000 001000 000001
# USER1-NEXT: .data PROGBITS 0000000000004000 002000 000008
# USER1-NEXT: .bss NOBITS 0000000000008000 002008 000008