Skip to content

Commit

Permalink
update the README a bit
Browse files Browse the repository at this point in the history
  • Loading branch information
PoroCYon committed Aug 24, 2020
1 parent 405d8eb commit 8d85d3d
Showing 1 changed file with 42 additions and 35 deletions.
77 changes: 42 additions & 35 deletions README.md
Original file line number Diff line number Diff line change
Expand Up @@ -18,15 +18,13 @@ PoC by Shiz, bugfixing and 64-bit version by PoroCYon.
the smol startup/symbol resolving code will jump to an undefined location.

```sh
./smold.py --use_interp --align_stack [--opts...] -lfoo -lbar input.o... output.elf
./smold.py -fuse-interp -falign-stack [--opts...] -lfoo -lbar input.o... output.elf
```

```
usage: smold.py [-h] [-m TARGET] [-l LIB] [-L DIR] [-s] [-n] [-d] [-fuse-interp] [-falign-stack] [-fuse-nx]
[-fuse-dnload-loader] [-fskip-zero-value] [-fuse-dt-debug] [-fuse-dl-fini] [-fskip-entries]
[-fno-start-arg] [-funsafe-dynamic] [--nasm NASM] [--cc CC] [--readelf READELF]
[--cflags CFLAGS] [--asflags ASFLAGS] [--ldflags LDFLAGS] [--smolrt SMOLRT] [--smolld SMOLLD]
[--verbose] [--keeptmp]
usage: smold.py [-h] [-m TARGET] [-l LIB] [-L DIR] [-s] [-n] [-d] [-fuse-interp] [-falign-stack] [-fuse-nx] [-fuse-dnload-loader] [-fskip-zero-value] [-fuse-dt-debug] [-fuse-dl-fini] [-fskip-entries] [-fno-start-arg] [-funsafe-dynamic]
[-fifunc-support] [-fifunc-strict-cconv] [--nasm NASM] [--cc CC] [--readelf READELF] [-Wc CFLAGS] [-Wa ASFLAGS] [-Wl LDFLAGS] [--smolrt SMOLRT] [--smolld SMOLLD] [--gen-rt-only] [--verbose] [--keeptmp]
[--debugout DEBUGOUT]
input [input ...] output
positional arguments:
Expand All @@ -41,43 +39,39 @@ optional arguments:
libraries to link against
-L DIR, --libdir DIR directories to search libraries in
-s, --hash16 Use 16-bit (BSD) hashes instead of 32-bit djb2 hashes. Implies -fuse-dnload-loader
-n, --nx Use NX (i.e. don't use RWE pages). Costs the size of one phdr, plus some extra bytes on
i386.
-n, --nx Use NX (i.e. don't use RWE pages). Costs the size of one phdr, plus some extra bytes on i386.
-d, --det Make the order of imports deterministic (default: just use whatever binutils throws at us)
-fuse-interp Include a program interpreter header (PT_INTERP). If not enabled, ld.so has to be invoked
manually by the end user.
-falign-stack Align the stack before running user code (_start). If not enabled, this has to be done
manually. Costs 1 byte.
-fuse-nx Don't use one big RWE segment, but use separate RW and RE ones. Use this to keep strict
kernels (PaX/grsec) happy. Costs at least the size of one program header entry.
-fuse-dnload-loader Use a dnload-style loader for resolving symbols, which doesn't depend on
nonstandard/undocumented ELF and ld.so features, but is slightly larger. If not enabled, a
smaller custom loader is used which assumes glibc.
-fskip-zero-value Skip an ELF symbol with a zero address (a weak symbol) when parsing libraries at runtime.
Try enabling this if you're experiencing sudden breakage. However, many libraries don't use
weak symbols, so this doesn't often pose a problem. Costs ~5 bytes.
-fuse-dt-debug Use the DT_DEBUG Dyn header to access the link_map, which doesn't depend on
nonstandard/undocumented ELF and ld.so features. If not enabled, the link_map is accessed
using data leaked to the entrypoint by ld.so, which assumes glibc. Costs ~10 bytes.
-fuse-dl-fini Pass _dl_fini to the user entrypoint, which should be done to properly comply with all
standards, but is very often not needed at all. Costs 2 bytes.
-fskip-entries Skip the first two entries in the link map (resp. ld.so and the vDSO). Speeds up symbol
resolving, but costs ~5 bytes.
-fno-start-arg Don't pass a pointer to argc/argv/envp to the entrypoint using the standard calling
convention. This means you need to read these yourself in assembly if you want to use them!
(envp is a preprequisite for X11, because it needs $DISPLAY.) Frees 3 bytes.
-funsafe-dynamic Don't end the ELF Dyn table with a DT_NULL entry. This might cause ld.so to interpret the
entire binary as the Dyn table, so only enable this if you're sure this won't break things!
-fuse-interp Include a program interpreter header (PT_INTERP). If not enabled, ld.so has to be invoked manually by the end user.
-falign-stack Align the stack before running user code (_start). If not enabled, this has to be done manually. Costs 1 byte.
-fuse-nx Don't use one big RWE segment, but use separate RW and RE ones. Use this to keep strict kernels (PaX/grsec) happy. Costs at least the size of one program header entry.
-fuse-dnload-loader Use a dnload-style loader for resolving symbols, which doesn't depend on nonstandard/undocumented ELF and ld.so features, but is slightly larger. If not enabled, a smaller custom loader is used which assumes
glibc.
-fskip-zero-value Skip an ELF symbol with a zero address (a weak symbol) when parsing libraries at runtime. Try enabling this if you're experiencing sudden breakage. However, many libraries don't use weak symbols, so this doesn't
often pose a problem. Costs ~5 bytes.
-fuse-dt-debug Use the DT_DEBUG Dyn header to access the link_map, which doesn't depend on nonstandard/undocumented ELF and ld.so features. If not enabled, the link_map is accessed using data leaked to the entrypoint by ld.so,
which assumes glibc. Costs ~10 bytes.
-fuse-dl-fini Pass _dl_fini to the user entrypoint, which should be done to properly comply with all standards, but is very often not needed at all. Costs 2 bytes.
-fskip-entries Skip the first two entries in the link map (resp. ld.so and the vDSO). Speeds up symbol resolving, but costs ~5 bytes.
-fno-start-arg Don't pass a pointer to argc/argv/envp to the entrypoint using the standard calling convention. This means you need to read these yourself in assembly if you want to use them! (envp is a preprequisite for X11,
because it needs $DISPLAY.) Frees 3 bytes.
-funsafe-dynamic Don't end the ELF Dyn table with a DT_NULL entry. This might cause ld.so to interpret the entire binary as the Dyn table, so only enable this if you're sure this won't break things!
-fifunc-support Support linking to IFUNCs. Probably needed on x86_64, but costs ~16 bytes. Ignored on platforms without IFUNC support.
-fifunc-strict-cconv On i386, if -fifunc-support is specified, strictly follow the calling convention rules. Probably not needed, but you never know.
--nasm NASM which nasm binary to use
--cc CC which cc binary to use (MUST BE GCC!)
--readelf READELF which readelf binary to use
--cflags CFLAGS Flags to pass to the C compiler for the relinking step
--asflags ASFLAGS Flags to pass to the assembler when creating the ELF header and runtime startup code
--ldflags LDFLAGS Flags to pass to the linker for the final linking step
-Wc CFLAGS, --cflags CFLAGS
Flags to pass to the C compiler for the relinking step
-Wa ASFLAGS, --asflags ASFLAGS
Flags to pass to the assembler when creating the ELF header and runtime startup code
-Wl LDFLAGS, --ldflags LDFLAGS
Flags to pass to the linker for the final linking step
--smolrt SMOLRT Directory containing the smol runtime sources
--smolld SMOLLD Directory containing the smol linker scripts
--gen-rt-only Only generate the headers/runtime assembly source file, instead of doing a full link. (I.e. fall back to pre-release behavior.)
--verbose Be verbose about what happens and which subcommands are invoked
--keeptmp Keep temp files (only useful for debugging)
--debugout DEBUGOUT Write out an additional, unrunnable debug ELF file with symbol information. (Useful for debugging with gdb, cannot be ran due to broken relocations.)
```

A minimal crt (and `_start` funcion) are provided in case you want to use `main`.
Expand All @@ -89,6 +83,19 @@ imported by a `smol`-ified binary. This can thus be used to detect user mistakes
during dynamic linking. (Think of it as an equivalent of `ldd`, except that it
also checks whether the imported functions are present as well.)

```
usage: smoldd.py [-h] [--cc CC] [--readelf READELF] [--map MAP] input
positional arguments:
input input file
optional arguments:
-h, --help show this help message and exit
--cc CC C compiler binary
--readelf READELF readelf binary
--map MAP Get the address of the symbol hash table from the linker map output instead of attempting to parse the binary.
```

## Internal workings

`smol.py` inspects the input object files for needed library files and symbols.
Expand Down

0 comments on commit 8d85d3d

Please sign in to comment.