linux.git/kernel/liveupdate, branch v6.19.12

liveupdate: luo_file: remember retrieve() status

2026-03-19T15:15:12+00:00

commit f85b1c6af5bc3872f994df0a5688c1162de07a62 upstream.

LUO keeps track of successful retrieve attempts on a LUO file.  It does so
to avoid multiple retrievals of the same file.  Multiple retrievals cause
problems because once the file is retrieved, the serialized data
structures are likely freed and the file is likely in a very different
state from what the code expects.

The retrieve boolean in struct luo_file keeps track of this, and is passed
to the finish callback so it knows what work was already done and what it
has left to do.

All this works well when retrieve succeeds.  When it fails,
luo_retrieve_file() returns the error immediately, without ever storing
anywhere that a retrieve was attempted or what its error code was.  This
results in an errored LIVEUPDATE_SESSION_RETRIEVE_FD ioctl to userspace,
but nothing prevents it from trying this again.

The retry is problematic for much of the same reasons listed above.  The
file is likely in a very different state than what the retrieve logic
normally expects, and it might even have freed some serialization data
structures.  Attempting to access them or free them again is going to
break things.

For example, if memfd managed to restore 8 of its 10 folios, but fails on
the 9th, a subsequent retrieve attempt will try to call
kho_restore_folio() on the first folio again, and that will fail with a
warning since it is an invalid operation.

Apart from the retry, finish() also breaks.  Since on failure the
retrieved bool in luo_file is never touched, the finish() call on session
close will tell the file handler that retrieve was never attempted, and it
will try to access or free the data structures that might not exist, much
in the same way as the retry attempt.

There is no sane way of attempting the retrieve again.  Remember the error
retrieve returned and directly return it on a retry.  Also pass this
status code to finish() so it can make the right decision on the work it
needs to do.

This is done by changing the bool to an integer.  A value of 0 means
retrieve was never attempted, a positive value means it succeeded, and a
negative value means it failed and the error code is the value.

Link: https://lkml.kernel.org/r/20260216132221.987987-1-pratyush@kernel.org
Fixes: 7c722a7f44e0 ("liveupdate: luo_file: implement file systems callbacks")
Signed-off-by: Pratyush Yadav (Google) 
Reviewed-by: Mike Rapoport (Microsoft) 
Cc: Pasha Tatashin 
Cc: 
Signed-off-by: Andrew Morton 
Signed-off-by: Greg Kroah-Hartman

kho: skip memoryless NUMA nodes when reserving scratch areas

2026-03-04T12:21:30+00:00

[ Upstream commit 427b2535f51342de3156babc6bdc3f3b7dd2c707 ]

kho_reserve_scratch() iterates over all online NUMA nodes to allocate
per-node scratch memory.  On systems with memoryless NUMA nodes (nodes
that have CPUs but no memory), memblock_alloc_range_nid() fails because
there is no memory available on that node.  This causes KHO initialization
to fail and kho_enable to be set to false.

Some ARM64 systems have NUMA topologies where certain nodes contain only
CPUs without any associated memory.  These configurations are valid and
should not prevent KHO from functioning.

Fix this by only counting nodes that have memory (N_MEMORY state) and skip
memoryless nodes in the per-node scratch allocation loop.

Link: https://lkml.kernel.org/r/20260120175913.34368-1-epetron@amazon.de
Fixes: 3dc92c311498 ("kexec: add Kexec HandOver (KHO) generation helpers").
Signed-off-by: Evangelos Petrongonas 
Reviewed-by: Pratyush Yadav 
Reviewed-by: Mike Rapoport (Microsoft) 
Reviewed-by: Pasha Tatashin 
Cc: Alexander Graf 
Cc: 
Signed-off-by: Andrew Morton 
Signed-off-by: Sasha Levin

liveupdate: luo_file: do not clear serialized_data on unfreeze

2026-02-03T02:43:55+00:00

Patch series "liveupdate: fixes in error handling".

This series contains some fixes in LUO's error handling paths.

The first patch deals with failed freeze() attempts.  The cleanup path
calls unfreeze, and that clears some data needed by later unpreserve
calls.

The second patch is a bit more involved.  It deals with failed retrieve()
attempts.  To do so properly, it reworks some of the error handling logic
in luo_file core.

Both these fixes are "theoretical" -- in the sense that I have not been
able to reproduce either of them in normal operation.  The only supported
file type right now is memfd, and there is nothing userspace can do right
now to make it fail its retrieve or freeze.  I need to make the retrieve
or freeze fail by artificially injecting errors.  The injected errors
trigger a use-after-free and a double-free.

That said, once more complex file handlers are added or memfd preservation
is used in ways not currently expected or covered by the tests, we will be
able to see them on real systems.


This patch (of 2):

The unfreeze operation is supposed to undo the effects of the freeze
operation.  serialized_data is not set by freeze, but by preserve. 
Consequently, the unpreserve operation needs to access serialized_data to
undo the effects of the preserve operation.  This includes freeing the
serialized data structures for example.

If a freeze callback fails, unfreeze is called for all frozen files.  This
would clear serialized_data for them.  Since live update has failed, it
can be expected that userspace aborts, releasing all sessions.  When the
sessions are released, unpreserve will be called for all files.  The
unfrozen files will see 0 in their serialized_data.  This is not expected
by file handlers, and they might either fail, leaking data and state, or
might even crash or cause invalid memory access.

Do not clear serialized_data on unfreeze so it gets passed on to
unpreserve.  There is no need to clear it on unpreserve since luo_file
will be freed immediately after.

Link: https://lkml.kernel.org/r/20260126230302.2936817-1-pratyush@kernel.org
Link: https://lkml.kernel.org/r/20260126230302.2936817-2-pratyush@kernel.org
Fixes: 7c722a7f44e0 ("liveupdate: luo_file: implement file systems callbacks")
Signed-off-by: Pratyush Yadav (Google) 
Reviewed-by: Mike Rapoport (Microsoft) 
Cc: Pasha Tatashin 
Signed-off-by: Andrew Morton

kho: kho_preserve_vmalloc(): don't return 0 when ENOMEM

2026-01-27T03:03:48+00:00

kho_preserve_vmalloc() should return -ENOMEM when new_vmalloc_chunk()
fails.

Reported-by: kernel test robot 
Reported-by: Dan Carpenter 
Closes: https://lore.kernel.org/r/202601211636.IRaejjdw-lkp@intel.com/
Reviewed-by: Pasha Tatashin 
Reviewed-by: Pratyush Yadav 
Reviewed-by: Mike Rapoport (Microsoft) 
Cc: Alexander Graf 
Cc: Jason Gunthorpe 
Cc: Pasha Tatashin 
Signed-off-by: Andrew Morton

kho: init alloc tags when restoring pages from reserved memory

2026-01-27T03:03:47+00:00

Memblock pages (including reserved memory) should have their allocation
tags initialized to CODETAG_EMPTY via clear_page_tag_ref() before being
released to the page allocator.  When kho restores pages through
kho_restore_page(), missing this call causes mismatched
allocation/deallocation tracking and below warning message:

alloc_tag was not set
WARNING: include/linux/alloc_tag.h:164 at ___free_pages+0xb8/0x260, CPU#1: swapper/0/1
RIP: 0010:___free_pages+0xb8/0x260
 kho_restore_vmalloc+0x187/0x2e0
 kho_test_init+0x3c4/0xa30
 do_one_initcall+0x62/0x2b0
 kernel_init_freeable+0x25b/0x480
 kernel_init+0x1a/0x1c0
 ret_from_fork+0x2d1/0x360

Add missing clear_page_tag_ref() annotation in kho_restore_page() to
fix this.

Link: https://lkml.kernel.org/r/20260122132740.176468-1-ranxiaokai627@163.com
Fixes: fc33e4b44b27 ("kexec: enable KHO support for memory preservation")
Signed-off-by: Ran Xiaokai 
Reviewed-by: Pratyush Yadav 
Reviewed-by: Pasha Tatashin 
Reviewed-by: Mike Rapoport (Microsoft) 
Cc: Alexander Graf 
Cc: Kent Overstreet 
Cc: Suren Baghdasaryan 
Cc: 
Signed-off-by: Andrew Morton

kho: validate preserved memory map during population

2026-01-15T06:16:21+00:00

If the previous kernel enabled KHO but did not call kho_finalize() (e.g.,
CONFIG_LIVEUPDATE=n or userspace skipped the finalization step), the
'preserved-memory-map' property in the FDT remains empty/zero.

Previously, kho_populate() would succeed regardless of the memory map's
state, reserving the incoming scratch regions in memblock.  However,
kho_memory_init() would later fail to deserialize the empty map.  By that
time, the scratch regions were already registered, leading to partial
initialization and subsequent list corruption (freeing scratch area twice)
during kho_init().

Move the validation of the preserved memory map earlier into
kho_populate(). If the memory map is empty/NULL:
1. Abort kho_populate() immediately with -ENOENT.
2. Do not register or reserve the incoming scratch memory, allowing the new
   kernel to reclaim those pages as standard free memory.
3. Leave the global 'kho_in' state uninitialized.

Consequently, kho_memory_init() sees no active KHO context
(kho_in.mem_chunks_phys is 0) and falls back to kho_reserve_scratch(),
allocating fresh scratch memory as if it were a standard cold boot.

Link: https://lkml.kernel.org/r/20251223140140.2090337-1-pasha.tatashin@soleen.com
Fixes: de51999e687c ("kho: allow memory preservation state updates after finalization")
Signed-off-by: Pasha Tatashin 
Reported-by: Ricardo Neri 
Closes: https://lore.kernel.org/all/20251218215613.GA17304@ranerica-svr.sc.intel.com
Reviewed-by: Mike Rapoport (Microsoft) 
Tested-by: Ricardo Neri 
Reviewed-by: Pratyush Yadav 
Cc: Alexander Graf 
Signed-off-by: Andrew Morton

mm: memfd_luo: add CONFIG_SHMEM dependency

2025-12-11T00:07:44+00:00

The new memfd code fails to link without SHMEM:

aarch64-linux-ld: mm/memfd_luo.o: in function `memfd_luo_retrieve_folios':
memfd_luo.c:(.text.memfd_luo_retrieve_folios+0xdc): undefined reference to `shmem_add_to_page_cache'
memfd_luo.c:(.text.memfd_luo_retrieve_folios+0x11c): undefined reference to `shmem_inode_acct_blocks'
memfd_luo.c:(.text.memfd_luo_retrieve_folios+0x134): undefined reference to `shmem_recalc_inode'

Add a Kconfig dependency to disallow that configuration.

Link: https://lkml.kernel.org/r/20251204100203.1034394-1-arnd@kernel.org
Fixes: b3749f174d68 ("mm: memfd_luo: allow preserving memfd")
Signed-off-by: Arnd Bergmann 
Reviewed-by: Pratyush Yadav 
Acked-by: Mike Rapoport (Microsoft) 
Reviewed-by: Pasha Tatashin 
Signed-off-by: Andrew Morton

liveupdate: luo_core: fix redundant bound check in luo_ioctl()

2025-12-11T00:07:42+00:00

The kernel test robot reported a Smatch warning:
kernel/liveupdate/luo_core.c:402 luo_ioctl() warn: unsigned 'nr' is
never less than zero.

This occurs because 'nr' is unsigned and LIVEUPDATE_CMD_BASE is currently
defined as 0, making the check (nr < LIVEUPDATE_CMD_BASE) always false.

Remove the explicit lower bound check.  The logic remains correct because
'nr' is unsigned; if nr is less than LIVEUPDATE_CMD_BASE, the expression
(nr - LIVEUPDATE_CMD_BASE) will wrap around to a large positive value. 
This will inevitably be larger than ARRAY_SIZE(luo_ioctl_ops) and be
caught by the upper bound check.

Link: https://lkml.kernel.org/r/20251130010919.1488230-1-pasha.tatashin@soleen.com
Signed-off-by: Pasha Tatashin 
Reported-by: kernel test robot 
Closes: https://lore.kernel.org/oe-kbuild-all/202511280300.6pvBmXUS-lkp@intel.com/
Reviewed-by: Mike Rapoport (Microsoft) 
Cc: David Matlack 
Cc: Pratyush Yadav 
Signed-off-by: Andrew Morton

liveupdate: luo_file: don't use invalid list iterator

2025-12-11T00:07:41+00:00

If we exit a list_for_each_entry() without hitting a break then the list
iterator points to an offset from the list_head.  It's a non-NULL but
invalid pointer and dereferencing it isn't allowed.

Introduce a new "found" variable to test instead.

Link: https://lkml.kernel.org/r/aSlMc4SS09Re4_xn@stanley.mountain
Fixes: 3ee1d673194e ("liveupdate: luo_file: implement file systems callbacks")
Signed-off-by: Dan Carpenter 
Reported-by: kernel test robot 
Closes: https://lore.kernel.org/r/202511280420.y9O4fyhX-lkp@intel.com/
Reviewed-by: Mike Rapoport (Microsoft) 
Reviewed-by: Pasha Tatashin 
Cc: Pratyush Yadav 
Signed-off-by: Andrew Morton

kho: fix restoring of contiguous ranges of order-0 pages

2025-11-27T22:24:44+00:00

When contiguous ranges of order-0 pages are restored, kho_restore_page()
calls prep_compound_page() with the first page in the range and order as
parameters and then kho_restore_pages() calls split_page() to make sure
all pages in the range are order-0.

However, since split_page() is not intended to split compound pages and
with VM_DEBUG enabled it will trigger a VM_BUG_ON_PAGE().

Update kho_restore_page() so that it will use prep_compound_page() when it
restores a folio and make sure it properly sets page count for both large
folios and ranges of order-0 pages.

Link: https://lkml.kernel.org/r/20251125110917.843744-3-rppt@kernel.org
Fixes: a667300bd53f ("kho: add support for preserving vmalloc allocations")
Signed-off-by: Mike Rapoport (Microsoft) 
Reported-by: Pratyush Yadav 
Cc: Alexander Graf 
Cc: Pasha Tatashin 
Signed-off-by: Andrew Morton