All of lore.kernel.org
 help / color / mirror / Atom feed
From: Andrew Morton <akpm@linux-foundation.org>
To: akpm@linux-foundation.org, anthony.yznaga@oracle.com,
	avagin@openvz.org, chenjingwen6@huawei.com,
	ebiederm@xmission.com, keescook@chromium.org,
	khalid.aziz@oracle.com, linux-mm@kvack.org,
	linux@armlinux.org.uk, mhocko@suse.com,
	mm-commits@vger.kernel.org, mpe@ellerman.id.au,
	torvalds@linux-foundation.org, viro@zeniv.linux.org.uk
Subject: [patch 42/87] binfmt_elf: reintroduce using MAP_FIXED_NOREPLACE
Date: Mon, 08 Nov 2021 18:33:37 -0800	[thread overview]
Message-ID: <20211109023337.byL1_5rp_%akpm@linux-foundation.org> (raw)
In-Reply-To: <20211108183057.809e428e841088b657a975ec@linux-foundation.org>

From: Kees Cook <keescook@chromium.org>
Subject: binfmt_elf: reintroduce using MAP_FIXED_NOREPLACE

Commit b212921b13bd ("elf: don't use MAP_FIXED_NOREPLACE for elf
executable mappings") reverted back to using MAP_FIXED to map ELF LOAD
segments because it was found that the segments in some binaries overlap
and can cause MAP_FIXED_NOREPLACE to fail.

The original intent of MAP_FIXED_NOREPLACE in the ELF loader was to
prevent the silent clobbering of an existing mapping (e.g.  stack) by the
ELF image, which could lead to exploitable conditions.  Quoting commit
4ed28639519c ("fs, elf: drop MAP_FIXED usage from elf_map"), which
originally introduced the use of MAP_FIXED_NOREPLACE in the loader:

    Both load_elf_interp and load_elf_binary rely on elf_map to map
    segments [to a specific] address and they use MAP_FIXED to enforce
    that. This is however [a] dangerous thing prone to silent data
    corruption which can be even exploitable.
    ...
    Let's take CVE-2017-1000253 as an example ... we could end up mapping
    [the executable] over the existing stack ... The [stack layout] issue
    has been fixed since then ... So we should be safe and any [similar]
    attack should be impractical. On the other hand this is just too
    subtle [an] assumption ... it can break quite easily and [be] hard to
    spot.
    ...
    Address this [weakness] by changing MAP_FIXED to the newly added
    MAP_FIXED_NOREPLACE. This will mean that mmap will fail if there is
    an existing mapping clashing with the requested one [instead of
    silently] clobbering it.

Then processing ET_DYN binaries the loader already calculates a total size
for the image when the first segment is mapped, maps the entire image, and
then unmaps the remainder before the remaining segments are then
individually mapped.  To avoid the earlier problems (legitimate
overlapping LOAD segments specified in the ELF), apply the same logic to
ET_EXEC binaries as well.  For both ET_EXEC and ET_DYN+INTERP use
MAP_FIXED_NOREPLACE for the initial total size mapping and then use
MAP_FIXED to build the final (possibly legitimately overlapping) mappings.
For ET_DYN w/out INTERP, continue to map at a system-selected address in
the mmap region.

Link: https://lkml.kernel.org/r/20210916215947.3993776-1-keescook@chromium.org
Link: https://lore.kernel.org/lkml/1595869887-23307-2-git-send-email-anthony.yznaga@oracle.com
Co-developed-by: Anthony Yznaga <anthony.yznaga@oracle.com>
Signed-off-by: Anthony Yznaga <anthony.yznaga@oracle.com>
Signed-off-by: Kees Cook <keescook@chromium.org>
Cc: Russell King <linux@armlinux.org.uk>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Eric Biederman <ebiederm@xmission.com>
Cc: Chen Jingwen <chenjingwen6@huawei.com>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: Andrei Vagin <avagin@openvz.org>
Cc: Khalid Aziz <khalid.aziz@oracle.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
---

 fs/binfmt_elf.c |   31 ++++++++++++++++++++++---------
 1 file changed, 22 insertions(+), 9 deletions(-)

--- a/fs/binfmt_elf.c~binfmt_elf-reintroduce-using-map_fixed_noreplace
+++ a/fs/binfmt_elf.c
@@ -1074,20 +1074,26 @@ out_free_interp:
 
 		vaddr = elf_ppnt->p_vaddr;
 		/*
-		 * If we are loading ET_EXEC or we have already performed
-		 * the ET_DYN load_addr calculations, proceed normally.
+		 * The first time through the loop, load_addr_set is false:
+		 * layout will be calculated. Once set, use MAP_FIXED since
+		 * we know we've already safely mapped the entire region with
+		 * MAP_FIXED_NOREPLACE in the once-per-binary logic following.
 		 */
-		if (elf_ex->e_type == ET_EXEC || load_addr_set) {
+		if (load_addr_set) {
 			elf_flags |= MAP_FIXED;
+		} else if (elf_ex->e_type == ET_EXEC) {
+			/*
+			 * This logic is run once for the first LOAD Program
+			 * Header for ET_EXEC binaries. No special handling
+			 * is needed.
+			 */
+			elf_flags |= MAP_FIXED_NOREPLACE;
 		} else if (elf_ex->e_type == ET_DYN) {
 			/*
 			 * This logic is run once for the first LOAD Program
 			 * Header for ET_DYN binaries to calculate the
 			 * randomization (load_bias) for all the LOAD
-			 * Program Headers, and to calculate the entire
-			 * size of the ELF mapping (total_size). (Note that
-			 * load_addr_set is set to true later once the
-			 * initial mapping is performed.)
+			 * Program Headers.
 			 *
 			 * There are effectively two types of ET_DYN
 			 * binaries: programs (i.e. PIE: ET_DYN with INTERP)
@@ -1108,7 +1114,7 @@ out_free_interp:
 			 * Therefore, programs are loaded offset from
 			 * ELF_ET_DYN_BASE and loaders are loaded into the
 			 * independently randomized mmap region (0 load_bias
-			 * without MAP_FIXED).
+			 * without MAP_FIXED nor MAP_FIXED_NOREPLACE).
 			 */
 			if (interpreter) {
 				load_bias = ELF_ET_DYN_BASE;
@@ -1117,7 +1123,7 @@ out_free_interp:
 				alignment = maximum_alignment(elf_phdata, elf_ex->e_phnum);
 				if (alignment)
 					load_bias &= ~(alignment - 1);
-				elf_flags |= MAP_FIXED;
+				elf_flags |= MAP_FIXED_NOREPLACE;
 			} else
 				load_bias = 0;
 
@@ -1129,7 +1135,14 @@ out_free_interp:
 			 * is then page aligned.
 			 */
 			load_bias = ELF_PAGESTART(load_bias - vaddr);
+		}
 
+		/*
+		 * Calculate the entire size of the ELF mapping (total_size).
+		 * (Note that load_addr_set is set to true later once the
+		 * initial mapping is performed.)
+		 */
+		if (!load_addr_set) {
 			total_size = total_mapping_size(elf_phdata,
 							elf_ex->e_phnum);
 			if (!total_size) {
_

  parent reply	other threads:[~2021-11-09  2:33 UTC|newest]

Thread overview: 98+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-11-09  2:30 incoming Andrew Morton
2021-11-09  2:31 ` [patch 01/87] vfs: keep inodes with page cache off the inode shrinker LRU Andrew Morton
2021-11-09  2:31 ` [patch 02/87] mm,hugetlb: remove mlock ulimit for SHM_HUGETLB Andrew Morton
2021-11-09  2:31 ` [patch 03/87] procfs: do not list TID 0 in /proc/<pid>/task Andrew Morton
2021-11-09  2:31 ` [patch 04/87] x86/xen: update xen_oldmem_pfn_is_ram() documentation Andrew Morton
2021-11-09  2:31 ` [patch 05/87] x86/xen: simplify xen_oldmem_pfn_is_ram() Andrew Morton
2021-11-09  2:31 ` [patch 06/87] x86/xen: print a warning when HVMOP_get_mem_type fails Andrew Morton
2021-11-09  2:31 ` [patch 07/87] proc/vmcore: let pfn_is_ram() return a bool Andrew Morton
2021-11-09  2:31 ` [patch 08/87] proc/vmcore: convert oldmem_pfn_is_ram callback to more generic vmcore callbacks Andrew Morton
2021-11-09  3:59   ` Dave Young
2021-11-09  6:40     ` David Hildenbrand
2021-11-09 10:30       ` Dave Young
2021-11-10  7:22   ` Baoquan He
2021-11-10  8:10     ` David Hildenbrand
2021-11-10 11:11       ` Dave Young
2021-11-10 11:21         ` David Hildenbrand
2021-11-10 11:28           ` Dave Young
2021-11-10 12:05             ` David Hildenbrand
2021-11-10 13:11               ` Dave Young
2021-11-09  2:31 ` [patch 09/87] virtio-mem: factor out hotplug specifics from virtio_mem_init() into virtio_mem_init_hotplug() Andrew Morton
2021-11-09  2:31 ` [patch 10/87] virtio-mem: factor out hotplug specifics from virtio_mem_probe() " Andrew Morton
2021-11-09  2:31 ` [patch 11/87] virtio-mem: factor out hotplug specifics from virtio_mem_remove() into virtio_mem_deinit_hotplug() Andrew Morton
2021-11-09  2:32 ` [patch 12/87] virtio-mem: kdump mode to sanitize /proc/vmcore access Andrew Morton
2021-11-09  2:32 ` [patch 13/87] proc: allow pid_revalidate() during LOOKUP_RCU Andrew Morton
2021-11-09  2:32 ` [patch 14/87] kernel.h: drop unneeded <linux/kernel.h> inclusion from other headers Andrew Morton
2021-11-09  2:32 ` [patch 15/87] kernel.h: split out container_of() and typeof_member() macros Andrew Morton
2021-11-09  2:32 ` [patch 16/87] include/kunit/test.h: replace kernel.h with the necessary inclusions Andrew Morton
2021-11-09  2:32 ` [patch 17/87] include/linux/list.h: " Andrew Morton
2021-11-09  2:32 ` [patch 18/87] include/linux/llist.h: " Andrew Morton
2021-11-09  2:32 ` [patch 19/87] include/linux/plist.h: " Andrew Morton
2021-11-09  2:32 ` [patch 20/87] include/media/media-entity.h: " Andrew Morton
2021-11-09  2:32 ` [patch 21/87] include/linux/delay.h: " Andrew Morton
2021-11-09  2:32 ` [patch 22/87] include/linux/sbitmap.h: " Andrew Morton
2021-11-09  2:32 ` [patch 23/87] include/linux/radix-tree.h: " Andrew Morton
2021-11-09  2:32 ` [patch 24/87] include/linux/generic-radix-tree.h: " Andrew Morton
2021-11-09  2:32 ` [patch 25/87] kernel.h: split out instruction pointer accessors Andrew Morton
2021-11-09  2:32 ` [patch 26/87] linux/container_of.h: switch to static_assert Andrew Morton
2021-11-09  2:32 ` [patch 27/87] mailmap: update email address for Colin King Andrew Morton
2021-11-09  2:32 ` [patch 28/87] MAINTAINERS: add "exec & binfmt" section with myself and Eric Andrew Morton
2021-11-09  2:32 ` [patch 29/87] MAINTAINERS: rectify entry for ARM/TOSHIBA VISCONTI ARCHITECTURE Andrew Morton
2021-11-09  2:32 ` [patch 30/87] MAINTAINERS: rectify entry for HIKEY960 ONBOARD USB GPIO HUB DRIVER Andrew Morton
2021-11-09  2:33 ` [patch 31/87] MAINTAINERS: rectify entry for INTEL KEEM BAY DRM DRIVER Andrew Morton
2021-11-09  2:33 ` [patch 32/87] MAINTAINERS: rectify entry for ALLWINNER HARDWARE SPINLOCK SUPPORT Andrew Morton
2021-11-09  2:33 ` [patch 33/87] lib, stackdepot: check stackdepot handle before accessing slabs Andrew Morton
2021-11-09  2:33 ` [patch 34/87] lib, stackdepot: add helper to print stack entries Andrew Morton
2021-11-09  2:33 ` [patch 35/87] lib, stackdepot: add helper to print stack entries into buffer Andrew Morton
2021-11-09  2:33 ` [patch 36/87] include/linux/string_helpers.h: add linux/string.h for strlen() Andrew Morton
2021-11-09  2:33 ` [patch 37/87] lib: uninline simple_strntoull() as well Andrew Morton
2021-11-09  2:33 ` [patch 38/87] mm/scatterlist: replace the !preemptible warning in sg_miter_stop() Andrew Morton
2021-11-09  2:33 ` [patch 39/87] const_structs.checkpatch: add a few sound ops structs Andrew Morton
2021-11-09  2:33 ` [patch 40/87] checkpatch: improve EXPORT_SYMBOL test for EXPORT_SYMBOL_NS uses Andrew Morton
2021-11-09  2:33 ` [patch 41/87] checkpatch: get default codespell dictionary path from package location Andrew Morton
2021-11-09  2:33 ` Andrew Morton [this message]
2021-11-09  2:33 ` [patch 43/87] ELF: simplify STACK_ALLOC macro Andrew Morton
2021-11-09  2:33 ` [patch 44/87] kallsyms: remove arch specific text and data check Andrew Morton
2021-11-09  2:33 ` [patch 45/87] kallsyms: fix address-checks for kernel related range Andrew Morton
2021-11-09  2:33 ` [patch 46/87] sections: move and rename core_kernel_data() to is_kernel_core_data() Andrew Morton
2021-11-09  2:33 ` [patch 47/87] sections: move is_kernel_inittext() into sections.h Andrew Morton
2021-11-09  2:33 ` [patch 48/87] x86: mm: rename __is_kernel_text() to is_x86_32_kernel_text() Andrew Morton
2021-11-09  2:34 ` [patch 49/87] sections: provide internal __is_kernel() and __is_kernel_text() helper Andrew Morton
2021-11-09  2:34 ` [patch 50/87] mm: kasan: use is_kernel() helper Andrew Morton
2021-11-09  2:34 ` [patch 51/87] extable: use is_kernel_text() helper Andrew Morton
2021-11-09  2:34 ` [patch 52/87] powerpc/mm: use core_kernel_text() helper Andrew Morton
2021-11-09  2:34 ` [patch 53/87] microblaze: use is_kernel_text() helper Andrew Morton
2021-11-09  2:34 ` [patch 54/87] alpha: " Andrew Morton
2021-11-09  2:34 ` [patch 55/87] ramfs: fix mount source show for ramfs Andrew Morton
2021-11-09  2:34 ` [patch 56/87] init: make unknown command line param message clearer Andrew Morton
2021-11-09  2:34 ` [patch 57/87] coda: avoid NULL pointer dereference from a bad inode Andrew Morton
2021-11-09  2:34 ` [patch 58/87] coda: check for async upcall request using local state Andrew Morton
2021-11-09  2:34 ` [patch 59/87] coda: remove err which no one care Andrew Morton
2021-11-09  2:34 ` [patch 60/87] coda: avoid flagging NULL inodes Andrew Morton
2021-11-09  2:34 ` [patch 61/87] coda: avoid hidden code duplication in rename Andrew Morton
2021-11-09  2:34 ` [patch 62/87] coda: avoid doing bad things on inode type changes during revalidation Andrew Morton
2021-11-09  2:34 ` [patch 63/87] coda: convert from atomic_t to refcount_t on coda_vm_ops->refcnt Andrew Morton
2021-11-09  2:34 ` [patch 64/87] coda: use vmemdup_user to replace the open code Andrew Morton
2021-11-09  2:34 ` [patch 65/87] coda: bump module version to 7.2 Andrew Morton
2021-11-09  2:34 ` [patch 66/87] nilfs2: replace snprintf in show functions with sysfs_emit Andrew Morton
2021-11-09  2:35 ` [patch 67/87] nilfs2: remove filenames from file comments Andrew Morton
2021-11-09  2:35 ` [patch 68/87] hfs/hfsplus: use WARN_ON for sanity check Andrew Morton
2021-11-09  2:35 ` [patch 69/87] crash_dump: fix boolreturn.cocci warning Andrew Morton
2021-11-09  2:35 ` [patch 70/87] crash_dump: remove duplicate include in crash_dump.h Andrew Morton
2021-11-09  2:35 ` [patch 71/87] signal: remove duplicate include in signal.h Andrew Morton
2021-11-09  2:35 ` [patch 72/87] seq_file: move seq_escape() to a header Andrew Morton
2021-11-09  2:35 ` [patch 73/87] seq_file: fix passing wrong private data Andrew Morton
2021-11-09  2:35 ` [patch 74/87] kernel/fork.c: unshare(): use swap() to make code cleaner Andrew Morton
2021-11-09  2:35 ` [patch 75/87] sysv: use BUILD_BUG_ON instead of runtime check Andrew Morton
2021-11-09  2:35 ` [patch 76/87] Documentation/kcov: include types.h in the example Andrew Morton
2021-11-09  2:35 ` [patch 77/87] Documentation/kcov: define `ip' " Andrew Morton
2021-11-09  2:35 ` [patch 78/87] kcov: allocate per-CPU memory on the relevant node Andrew Morton
2021-11-09  2:35 ` [patch 79/87] kcov: avoid enable+disable interrupts if !in_task() Andrew Morton
2021-11-09  2:35 ` [patch 80/87] kcov: replace local_irq_save() with a local_lock_t Andrew Morton
2021-11-09  2:35 ` [patch 81/87] scripts/gdb: handle split debug for vmlinux Andrew Morton
2021-11-09  2:35 ` [patch 82/87] kernel/resource: clean up and optimize iomem_is_exclusive() Andrew Morton
2021-11-09  2:35 ` [patch 83/87] kernel/resource: disallow access to exclusive system RAM regions Andrew Morton
2021-11-09  2:35 ` [patch 84/87] virtio-mem: disallow mapping virtio-mem memory via /dev/mem Andrew Morton
2021-11-09  2:35 ` [patch 85/87] selftests/kselftest/runner/run_one(): allow running non-executable files Andrew Morton
2021-11-09  2:35 ` [patch 86/87] ipc: check checkpoint_restore_ns_capable() to modify C/R proc files Andrew Morton
2021-11-09  2:36 ` [patch 87/87] ipc/ipc_sysctl.c: remove fallback for !CONFIG_PROC_SYSCTL Andrew Morton

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=20211109023337.byL1_5rp_%akpm@linux-foundation.org \
    --to=akpm@linux-foundation.org \
    --cc=anthony.yznaga@oracle.com \
    --cc=avagin@openvz.org \
    --cc=chenjingwen6@huawei.com \
    --cc=ebiederm@xmission.com \
    --cc=keescook@chromium.org \
    --cc=khalid.aziz@oracle.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=linux@armlinux.org.uk \
    --cc=mhocko@suse.com \
    --cc=mm-commits@vger.kernel.org \
    --cc=mpe@ellerman.id.au \
    --cc=torvalds@linux-foundation.org \
    --cc=viro@zeniv.linux.org.uk \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.