mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: "Lorenzo Stoakes (ARM)" <ljs@kernel.org>
To: Brendan Jackman <jackmanb@google.com>
Cc: Andrew Morton <akpm@linux-foundation.org>,
	 David Hildenbrand <david@kernel.org>,
	"Liam R. Howlett" <liam@infradead.org>,
	 Vlastimil Babka <vbabka@kernel.org>,
	Mike Rapoport <rppt@kernel.org>,
	 Suren Baghdasaryan <surenb@google.com>,
	Michal Hocko <mhocko@suse.com>,
	linux-mm@kvack.org,  linux-kernel@vger.kernel.org,
	brendan.jackman@linux.dev
Subject: Re: [PATCH v2] mm/secretmem: don't allow highmem folios
Date: Fri, 17 Jul 2026 13:44:52 +0100	[thread overview]
Message-ID: <alojNp3Ow9YiA2js@lucifer> (raw)
In-Reply-To: <20260717-secretmem-highmem-v2-1-1f1a961ca91e@google.com>

On Fri, Jul 17, 2026 at 09:48:59AM +0000, Brendan Jackman wrote:
> secretmem_fault() allocates a folio with GFP_HIGHUSER and then calls
> set_direct_map_invalid_noflush() without checking folio_test_highmem().
> This causes a warning and process crash (vibe-coded reproducer in Link
> below):
>
> Su[   30.071284] ------------[ cut here ]------------
> ccessfully allocated and mapped 2097152000 bytes at 0x3a449000
> Populating memor[   30.074614] CPA: called for zero pte. vaddr = 0 cpa->vaddr = 0
> y...
> [   30.078636] WARNING: arch/x86/mm/pat/set_memory.c:1840 at __cpa_process_fault+0x34d/0x360, CPU#5: allocate_secret/570
> [   30.084789] CPU: 5 UID: 0 PID: 570 Comm: allocate_secret Not tainted 7.1.0-14063-g4edcdefd4083-dirty #10 PREEMPTLAZY
> [   30.090937] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.17.0-0-gb52ca86e094d-prebuilt.qemu.org 04/01/2014
> [   30.097543] EIP: __cpa_process_fault+0x34d/0x360
> [   30.100514] Code: ff ff 85 c0 0f 89 7d fe ff ff e9 3d fe ff ff 8b 03 8b 00 c7 04 24 c8 ff 64 c1 89 44 24 08 8b 45 e8 89 44 24 04 e8 53 7
> a 00 00 <0f> 0b c7 45 f0 f2 ff ff ff e9 fc fc ff ff 90 8d 74 26 00 55 25 00
> [   30.110829] EAX: 00000000 EBX: f64afe98 ECX: 00000000 EDX: 00000000
> [   30.114799] ESI: 00000000 EDI: f64afe98 EBP: f64afe04 ESP: f64afdcc
> [   30.118785] DS: 007b ES: 007b FS: 00d8 GS: 0033 SS: 0068 EFLAGS: 00010246
> [   30.123020] CR0: 80050033 CR2: 46c48ffc CR3: 038c8000 CR4: 00000690
> [   30.127010] Call Trace:
> [   30.129078]  __change_page_attr_set_clr+0x5e7/0x870
> [   30.132275]  ? console_unlock+0x99/0x130
> [   30.135069]  ? irq_work_queue+0x36/0x70
> [   30.137853]  ? page_address+0xd3/0xf0
> [   30.140421]  set_direct_map_invalid_noflush+0x52/0x60
> [   30.143782]  secretmem_fault+0x128/0x210
> [   30.146560]  __do_fault+0x25/0x90
> [   30.149053]  handle_mm_fault+0x6d1/0xcb0
> [   30.151759]  exc_page_fault+0x135/0x3b0
> [   30.154487]  ? doublefault_shim+0x150/0x150
> [   30.157416]  handle_exception+0x130/0x130
> [   30.160137] EIP: 0x804d29f
> [   30.162307] Code: 89 54 08 e1 89 54 08 e5 89 54 08 e9 89 54 08 ed c3 0f b6 44 24 08 89 7c 24 0c 69 c0 01 01 01 01 8b 7c 24 04 f7 c7 0f 0
> 0 00 00 <89> 44 0f fc 75 0e c1 e9 02 f3 ab 8b 44 24 04 8b 7c 24 0c c3 31 d2
> [   30.172936] EAX: 5a5a5a5a EBX: 00000000 ECX: 0c800000 EDX: 3a449000
> [   30.176927] ESI: 00000000 EDI: 3a449000 EBP: bfbbae18 ESP: bfbbadac
> [   30.180897] DS: 007b ES: 007b FS: 0000 GS: 0033 SS: 007b EFLAGS: 00010246
> [   30.185161]  ? doublefault_shim+0x150/0x150
> [   30.187979] ---[ end trace 0000000000000000 ]---
> Bus error                  (core dumped) ./allocate_secret_i686 2000M
>
> The equivalent bug was pointed out by a local Sashiko instance on
> https://lore.kernel.org/all/20260410151746.61150-3-kalyazin@amazon.com/
>
> This hasn't been reproduced it on older kernel versions but from code
> inspection the bug seems to go back to the original introduction in
> commit 1507f51255c9f ("mm: introduce memfd_secret system call to create
> "secret" memory areas"). If this configuration has always been broken,
> there's no need to worry too much about feature regression here.
>
> Nonetheless, instead of just completely disabling secretmem under
> !HIGHMEM, just drop __GFP_HIGHMEM. This means that now where you
> previously got a crash, instead you'll just see the secretmem process
> OOM.
>
> Could secretmem just support highmem by saying "this isn't in the direct
> map anyway" and bailing out before the set_direct_map_invalid_noflush()?
> Maybe. That depends on requirements that are not well-defined (e.g. is
> it OK that kmap_local_page() is not a NOP for those pages?), and would
> require some research and deep thinking. Let's "defer" that until
> an actual usecase arises.

I mean, I vote let's get rid of highmem altogether ;)

>
> Link: https://github.com/bjackman/limmat-kernel-nix/commit/7b2acba2d3a5ef01400d493a155beb1d135b6bb5
> Suggested-by: Andrew Morton <akpm@linux-foundation.org>
> Link: https://lore.kernel.org/all/20260704192603.40aa80cf9242b77aa75e8d8d@linux-foundation.org/
> Fixes: 1507f51255c9f ("mm: introduce memfd_secret system call to create "secret" memory areas")
> Signed-off-by: Brendan Jackman <jackmanb@google.com>

LGTM, so:

Reviewed-by: Lorenzo Stoakes (ARM) <ljs@kernel.org>

> ---
> Based on mm-new with `[PATCH] mm/secretmem: disable under HIGHMEM` [1]
> reverted.
>
> Note this was found by Sashiko but in a local instance, so I can't
> provide a link to the report and I guess it's technically inaccurate to
> mention sashiko-bot@kernel.org.
>
> I wrote in [0] that I needed to compile a 32bit OS but I was wrong, the
> build is already done and this is tested on QEMU, the vibe-coded repro
> linked in the patch now gets OOM-killed as appropriate:
>
> Populated 838860800/2097152000 bytes (40%)...
> [  202.930157] allocate_secret invoked oom-killer: gfp_mask=0x140dc0(GFP_USER|__GFP_ZERO|__GFP_COMP), order=0, oom_score_adj=0
> [  202.937102] CPU: 1 UID: 0 PID: 588 Comm: allocate_secret Not tainted 7.2.0-rc2-00579-g00ba42d1257f #56 PREEMPTLAZY
> [  202.937105] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.17.0-0-gb52ca86e094d-prebuilt.qemu.org 04/01/2014
> [  202.937106] Call Trace:
> [  202.937108]  dump_stack_lvl+0x44/0x64
> [  202.937112]  dump_stack+0xd/0x12
> [  202.937113]  dump_header+0x52/0x2f8
> [  202.937115]  ? ___ratelimit+0x169/0x220
> [  202.937118]  oom_kill_process.cold+0x5b/0xe4
> [  202.937119]  ? __percpu_counter_sum+0x96/0xb0
> [  202.937122]  ? try_to_free_pages+0x371/0x400
> [  202.937125]  ? oom_badness+0x13b/0x1c0
> [  202.937127]  out_of_memory+0xa8/0x3f0
> [  202.937128]  __alloc_frozen_pages_noprof+0x790/0xa60
> [  202.937131]  __folio_alloc_noprof+0x1f/0x40
> [  202.937132]  secretmem_fault+0x108/0x1c0
> [  202.937134]  __do_fault+0x30/0xa0
> [  202.937136]  handle_mm_fault+0x831/0xe10
> [  202.937138]  exc_page_fault+0x135/0x3b0
> [  202.937140]  ? doublefault_shim+0x150/0x150
> [  202.937141]  handle_exception+0x130/0x130
> [  202.937143] EIP: 0x804d2a8
> [  202.937145] Code: 54 08 e9 89 54 08 ed c3 0f b6 44 24 08 89 7c 24 0c 69 c0 01 01 01 01 8b 7c 24 04 f7 c7 0f 00 00 00 89 44 0f fc 75 0e c1 e9 02 <f3> ab 8b 44 24 04 8b 7c 24 0c c3 31 d2 29 fa 83 e2 0f 89 07 89 47
> [  202.937146] EAX: 5a5a5a5a EBX: 00000000 ECX: 02c49c00 EDX: 6106d000
> [  202.937148] ESI: 00000000 EDI: 62746000 EBP: bfc9b598 ESP: bfc9b52c
> [  202.937149] DS: 007b ES: 007b FS: 0000 GS: 0033 SS: 007b EFLAGS: 00010206
> [  202.937150]  ? doublefault_shim+0x150/0x150
> [  202.937152] Mem-Info:
> [  203.033004] active_anon:76 inactive_anon:5407 isolated_anon:0
> [  203.033004]  active_file:3587 inactive_file:1924 isolated_file:0
> [  203.033004]  unevictable:229432 dirty:19 writeback:0
> [  203.033004]  slab_reclaimable:1836 slab_unreclaimable:3150
> [  203.033004]  mapped:215094 shmem:2040 pagetables:680
> [  203.033004]  sec_pagetables:0 bounce:0
> [  203.033004]  kernel_misc_reclaimable:0
> [  203.033004]  free:7605 free_pcp:2490 free_cma:0
> [  203.055764] Node 0 active_anon:304kB inactive_anon:21628kB active_file:14308kB inactive_file:7692kB unevictable:917728kB isolated(anon):0kB isolated(file):0kB mapped:860376kB dirty:76kB writeback:0kB shmem:8160kB kernel_stack:992kB pagetables:2720kB sec_pagetables:0kB all_unreclaimable? yes Balloon:0kB gpu_active:0kB gpu_reclaim:0kB
> [  203.073182] DMA free:3444kB boost:0kB min:64kB low:80kB high:96kB reserved_highatomic:0KB free_highatomic:0KB active_anon:0kB inactive_anon:0kB active_file:0kB inactive_file:0kB unevictable:11732kB writepending:0kB zspages:0kB present:15992kB managed:15360kB mlocked:11732kB bounce:0kB free_pcp:124kB local_pcp:124kB free_cma:0kB
> [  203.090364] lowmem_reserve[]: 0 852 990 990
> [  203.093476] Normal free:3408kB boost:0kB min:3644kB low:4552kB high:5460kB reserved_highatomic:0KB free_highatomic:0KB active_anon:0kB inactive_anon:0kB active_file:308kB inactive_file:276kB unevictable:834060kB writepending:28kB zspages:0kB present:890872kB managed:873360kB mlocked:830868kB bounce:0kB free_pcp:9184kB local_pcp:3192kB free_cma:0kB
> [  203.111774] lowmem_reserve[]: 0 0 1102 1102
> [  203.114773] DMA: 1*4kB (U) 0*8kB 1*16kB (U) 1*32kB (U) 1*64kB (U) 0*128kB 1*256kB (U) 0*512kB 1*1024kB (U) 1*2048kB (U) 0*4096kB = 3444kB
> [  203.122523] Normal: 182*4kB (UME) 34*8kB (UE) 22*16kB (UE) 5*32kB (UE) 6*64kB (UE) 1*128kB (E) 0*256kB 1*512kB (E) 1*1024kB (M) 0*2048kB 0*4096kB = 3560kB
> [  203.131052] 237037 total pagecache pages
> [  203.134049] 0 pages in swap cache
> [  203.136716] Free swap  = 0kB
> [  203.139008] Total swap = 0kB
> [  203.141426] 262009 pages RAM
> [  203.143843] 35293 pages HighMem/MovableOnly
> [  203.146820] 4536 pages reserved
> [  203.149415] Out of memory: Killed process 588 (allocate_secret) total-vm:2048216kB, anon-rss:24kB, file-rss:842648kB, shmem-rss:0kB, UID:0 pgtables:1680kB oom_score_adj:0
> Killed                     ./allocate_secret_i686 2000M
>
> [0] https://lore.kernel.org/all/DK0QLG6P7LXQ.OUG1KJ75V4MW@linux.dev/
> [1] https://lore.kernel.org/all/20260703-secretmem-highmem-v1-1-30d5ff944664@google.com/
> ---
> Changes in v2:
> - Instead of disabling CONFIG_SECRETMEM, just drop __GFP_HIGHMEM
> - Link to v1: https://patch.msgid.link/20260703-secretmem-highmem-v1-1-30d5ff944664@google.com
> ---
>  mm/secretmem.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/mm/secretmem.c b/mm/secretmem.c
> index 4877c262cb1f6..d29865075b6ea 100644
> --- a/mm/secretmem.c
> +++ b/mm/secretmem.c
> @@ -202,7 +202,7 @@ static struct file *secretmem_file_create(unsigned long flags)
>  	if (IS_ERR(file))
>  		goto err_free_inode;
>
> -	mapping_set_gfp_mask(inode->i_mapping, GFP_HIGHUSER);
> +	mapping_set_gfp_mask(inode->i_mapping, GFP_USER);
>  	mapping_set_unevictable(inode->i_mapping);
>
>  	inode->i_op = &secretmem_iops;
>
> ---
> base-commit: 6efb6baa4a240b06bcef6b83cfca04cd2bad6852
> change-id: 20260703-secretmem-highmem-44eb42fda318
>
> Best regards,
> --
> Brendan Jackman <jackmanb@google.com>
>

Cheers, Lorenzo

      parent reply	other threads:[~2026-07-17 12:45 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-07-17  9:48 Brendan Jackman
2026-07-17 12:03 ` David Hildenbrand (Arm)
2026-07-17 12:44 ` Lorenzo Stoakes (ARM) [this message]

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=alojNp3Ow9YiA2js@lucifer \
    --to=ljs@kernel.org \
    --cc=akpm@linux-foundation.org \
    --cc=brendan.jackman@linux.dev \
    --cc=david@kernel.org \
    --cc=jackmanb@google.com \
    --cc=liam@infradead.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=mhocko@suse.com \
    --cc=rppt@kernel.org \
    --cc=surenb@google.com \
    --cc=vbabka@kernel.org \
    /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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox