linux.git/init/initramfs.c, branch v6.1.168

initramfs: avoid filename buffer overrun

2024-12-14T18:53:14+00:00

[ Upstream commit e017671f534dd3f568db9e47b0583e853d2da9b5 ]

The initramfs filename field is defined in
Documentation/driver-api/early-userspace/buffer-format.rst as:

 37 cpio_file := ALGN(4) + cpio_header + filename + "\0" + ALGN(4) + data
...
 55 ============= ================== =========================
 56 Field name    Field size         Meaning
 57 ============= ================== =========================
...
 70 c_namesize    8 bytes            Length of filename, including final \0

When extracting an initramfs cpio archive, the kernel's do_name() path
handler assumes a zero-terminated path at @collected, passing it
directly to filp_open() / init_mkdir() / init_mknod().

If a specially crafted cpio entry carries a non-zero-terminated filename
and is followed by uninitialized memory, then a file may be created with
trailing characters that represent the uninitialized memory. The ability
to create an initramfs entry would imply already having full control of
the system, so the buffer overrun shouldn't be considered a security
vulnerability.

Append the output of the following bash script to an existing initramfs
and observe any created /initramfs_test_fname_overrunAA* path. E.g.
  ./reproducer.sh | gzip >> /myinitramfs

It's easiest to observe non-zero uninitialized memory when the output is
gzipped, as it'll overflow the heap allocated @out_buf in __gunzip(),
rather than the initrd_start+initrd_size block.

---- reproducer.sh ----
nilchar="A"	# change to "\0" to properly zero terminate / pad
magic="070701"
ino=1
mode=$(( 0100777 ))
uid=0
gid=0
nlink=1
mtime=1
filesize=0
devmajor=0
devminor=1
rdevmajor=0
rdevminor=0
csum=0
fname="initramfs_test_fname_overrun"
namelen=$(( ${#fname} + 1 ))	# plus one to account for terminator

printf "%s%08x%08x%08x%08x%08x%08x%08x%08x%08x%08x%08x%08x%08x%s" \
	$magic $ino $mode $uid $gid $nlink $mtime $filesize \
	$devmajor $devminor $rdevmajor $rdevminor $namelen $csum $fname

termpadlen=$(( 1 + ((4 - ((110 + $namelen) & 3)) % 4) ))
printf "%.s${nilchar}" $(seq 1 $termpadlen)
---- reproducer.sh ----

Symlink filename fields handled in do_symlink() won't overrun past the
data segment, due to the explicit zero-termination of the symlink
target.

Fix filename buffer overrun by aborting the initramfs FSM if any cpio
entry doesn't carry a zero-terminator at the expected (name_len - 1)
offset.

Fixes: 1da177e4c3f41 ("Linux-2.6.12-rc2")
Signed-off-by: David Disseldorp 
Link: https://lore.kernel.org/r/20241030035509.20194-2-ddiss@suse.de
Signed-off-by: Christian Brauner 
Signed-off-by: Sasha Levin

init: open /initrd.image with O_LARGEFILE

2024-04-03T13:19:47+00:00

commit 4624b346cf67400ef46a31771011fb798dd2f999 upstream.

If initrd data is larger than 2Gb, we'll eventually fail to write to the
/initrd.image file when we hit that limit, unless O_LARGEFILE is set.

Link: https://lkml.kernel.org/r/20240317221522.896040-1-jsperbeck@google.com
Signed-off-by: John Sperbeck 
Cc: Jens Axboe 
Cc: Nick Desaulniers 
Cc: Peter Zijlstra 
Cc: Thomas Gleixner 
Cc: 
Signed-off-by: Andrew Morton 
Signed-off-by: Greg Kroah-Hartman

initramfs: mark my_inptr as __initdata

2022-09-12T04:55:11+00:00

As my_inptr is only used in __init function unpack_to_rootfs(), mark it as
__initdata to allow it be freed after boot.

Link: https://lkml.kernel.org/r/20220827071116.83078-1-wuchi.zero@gmail.com
Signed-off-by: wuchi 
Reviewed-by: David Disseldorp 
Cc: Alexander Viro 
Cc: Martin Wilck 
Cc: Matthew Wilcox 
Signed-off-by: Andrew Morton

Merge tag 'kthread-cleanups-for-v5.19' of git://git.kernel.org/pub/scm/linux/kernel/git/ebiederm/user-namespace

2022-06-03T23:03:05+00:00

Pull kthread updates from Eric Biederman:
 "This updates init and user mode helper tasks to be ordinary user mode
  tasks.

  Commit 40966e316f86 ("kthread: Ensure struct kthread is present for
  all kthreads") caused init and the user mode helper threads that call
  kernel_execve to have struct kthread allocated for them. This struct
  kthread going away during execve in turned made a use after free of
  struct kthread possible.

  Here, commit 343f4c49f243 ("kthread: Don't allocate kthread_struct for
  init and umh") is enough to fix the use after free and is simple
  enough to be backportable.

  The rest of the changes pass struct kernel_clone_args to clean things
  up and cause the code to make sense.

  In making init and the user mode helpers tasks purely user mode tasks
  I ran into two complications. The function task_tick_numa was
  detecting tasks without an mm by testing for the presence of
  PF_KTHREAD. The initramfs code in populate_initrd_image was using
  flush_delayed_fput to ensuere the closing of all it's file descriptors
  was complete, and flush_delayed_fput does not work in a userspace
  thread.

  I have looked and looked and more complications and in my code review
  I have not found any, and neither has anyone else with the code
  sitting in linux-next"

* tag 'kthread-cleanups-for-v5.19' of git://git.kernel.org/pub/scm/linux/kernel/git/ebiederm/user-namespace:
  sched: Update task_tick_numa to ignore tasks without an mm
  fork: Stop allowing kthreads to call execve
  fork: Explicitly set PF_KTHREAD
  init: Deal with the init process being a user mode process
  fork: Generalize PF_IO_WORKER handling
  fork: Explicity test for idle tasks in copy_thread
  fork: Pass struct kernel_clone_args into copy_thread
  kthread: Don't allocate kthread_struct for init and umh

initramfs: support cpio extraction with file checksums

2022-05-10T01:29:20+00:00

Add support for extraction of checksum-enabled "070702" cpio archives,
specified in Documentation/driver-api/early-userspace/buffer-format.rst. 
Fail extraction if the calculated file data checksum doesn't match the
value carried in the header.

Link: https://lkml.kernel.org/r/20220404093429.27570-7-ddiss@suse.de
Signed-off-by: David Disseldorp 
Suggested-by: Matthew Wilcox (Oracle) 
Cc: Al Viro 
Cc: Christian Brauner 
Cc: Martin Wilck 
Signed-off-by: Andrew Morton

initramfs: add INITRAMFS_PRESERVE_MTIME Kconfig option

2022-05-10T01:29:19+00:00

initramfs cpio mtime preservation, as implemented in commit 889d51a10712
("initramfs: add option to preserve mtime from initramfs cpio images"),
uses a linked list to defer directory mtime processing until after all
other items in the cpio archive have been processed.  This is done to
ensure that parent directory mtimes aren't overwritten via subsequent
child creation.

The lkml link below indicates that the mtime retention use case was for
embedded devices with applications running exclusively out of initramfs,
where the 32-bit mtime value provided a rough file version identifier. 
Linux distributions which discard an extracted initramfs immediately after
the root filesystem has been mounted may want to avoid the unnecessary
overhead.

This change adds a new INITRAMFS_PRESERVE_MTIME Kconfig option, which can
be used to disable on-by-default mtime retention and in turn speed up
initramfs extraction, particularly for cpio archives with large directory
counts.

Benchmarks with a one million directory cpio archive extracted 20 times
demonstrated:
				mean extraction time (s)	std dev
INITRAMFS_PRESERVE_MTIME=y		3.808			 0.006
INITRAMFS_PRESERVE_MTIME unset		3.056			 0.004

The above extraction times were measured using ftrace (initcall_finish -
initcall_start) values for populate_rootfs() with initramfs_async
disabled.

[ddiss@suse.de: rebase atop dir_entry.name flexible array member and drop separate initramfs_mtime.h header]
Link: https://lkml.org/lkml/2008/9/3/424
Link: https://lkml.kernel.org/r/20220404093429.27570-4-ddiss@suse.de
Signed-off-by: David Disseldorp 
Reviewed-by: Martin Wilck 
Cc: Al Viro 
Cc: Christian Brauner 
Cc: Matthew Wilcox (Oracle) 
Signed-off-by: Andrew Morton

initramfs: make dir_entry.name a flexible array member

2022-05-10T01:29:19+00:00

dir_entry.name is currently allocated via a separate kstrdup().  Change it
to a flexible array member and allocate it along with struct dir_entry.

Link: https://lkml.kernel.org/r/20220404093429.27570-3-ddiss@suse.de
Signed-off-by: David Disseldorp 
Acked-by: Christian Brauner 
Cc: Al Viro 
Cc: Martin Wilck 
Cc: Matthew Wilcox (Oracle) 
Signed-off-by: Andrew Morton

initramfs: refactor do_header() cpio magic checks

2022-05-10T01:29:19+00:00

Patch series "initramfs: "crc" cpio format and INITRAMFS_PRESERVE_MTIME", v7.

This patchset does some minor initramfs refactoring and allows cpio entry
mtime preservation to be disabled via a new Kconfig
INITRAMFS_PRESERVE_MTIME option.

Patches 4/6 to 6/6 implement support for creation and extraction of "crc"
cpio archives, which carry file data checksums.  Basic tests for this
functionality can be found at https://github.com/rapido-linux/rapido/pull/163


This patch (of 6):

do_header() is called for each cpio entry and fails if the first six bytes
don't match "newc" magic.  The magic check includes a special case error
message if POSIX.1 ASCII (cpio -H odc) magic is detected.  This special
case POSIX.1 check can be nested under the "newc" mismatch code path to
avoid calling memcmp() twice in a non-error case.

Link: https://lkml.kernel.org/r/20220404093429.27570-1-ddiss@suse.de
Link: https://lkml.kernel.org/r/20220404093429.27570-2-ddiss@suse.de
Signed-off-by: David Disseldorp 
Reviewed-by: Martin Wilck 
Acked-by: Christian Brauner 
Cc: Al Viro 
Cc: Matthew Wilcox 
Signed-off-by: Andrew Morton

init: Deal with the init process being a user mode process

2022-05-07T14:01:59+00:00

It is silly for user_mode_thread to leave PF_KTHREAD set
on the resulting task.  Update the init process so that
it does not care if PF_KTHREAD is set or not.

Ensure do_populate_rootfs flushes all delayed fput work by calling
task_work_run.  In the rare instance that async_schedule_domain calls
do_populate_rootfs synchronously it is possible do_populate_rootfs
will be called directly from the init process.  At which point fput
will call "task_work_add(current, ..., TWA_RESUME)".  The files on the
initramfs need to be completely put before we attempt to exec them
(which is before the code enters userspace).  So call task_work_run
just in case there are any pending fput operations.

Link: https://lkml.kernel.org/r/20220506141512.516114-5-ebiederm@xmission.com
Signed-off-by: "Eric W. Biederman"

memblock: use memblock_free for freeing virtual pointers

2021-11-06T20:30:41+00:00

Rename memblock_free_ptr() to memblock_free() and use memblock_free()
when freeing a virtual pointer so that memblock_free() will be a
counterpart of memblock_alloc()

The callers are updated with the below semantic patch and manual
addition of (void *) casting to pointers that are represented by
unsigned long variables.

    @@
    identifier vaddr;
    expression size;
    @@
    (
    - memblock_phys_free(__pa(vaddr), size);
    + memblock_free(vaddr, size);
    |
    - memblock_free_ptr(vaddr, size);
    + memblock_free(vaddr, size);
    )

[sfr@canb.auug.org.au: fixup]
  Link: https://lkml.kernel.org/r/20211018192940.3d1d532f@canb.auug.org.au

Link: https://lkml.kernel.org/r/20210930185031.18648-7-rppt@kernel.org
Signed-off-by: Mike Rapoport 
Signed-off-by: Stephen Rothwell 
Cc: Christophe Leroy 
Cc: Juergen Gross 
Cc: Shahab Vahedi 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds