From: Zhihao Cheng <chengzhihao1@huawei.com>
To: Jenny Guanni Qu <qguanni@gmail.com>, <richard@nod.at>
Cc: <linux-mtd@lists.infradead.org>, <klaudia@vidocsecurity.com>,
<dawid@vidocsecurity.com>
Subject: Re: [PATCH] ubifs: fix out-of-bounds write in LPT commit padding
Date: Sat, 14 Mar 2026 10:15:46 +0800 [thread overview]
Message-ID: <3fa4d791-4520-0e6d-d657-89d6fc4ba5cb@huawei.com> (raw)
In-Reply-To: <20260313222213.2181408-1-qguanni@gmail.com>
在 2026/3/14 6:22, Jenny Guanni Qu 写道:
> In write_cnodes(), when writing LPT data to a LEB, the code pads the
> write buffer to min_io_size alignment using:
>
> memset(buf + offs, 0xff, alen - wlen)
>
> where wlen = offs - from and alen = ALIGN(wlen, min_io_size). The
> buffer (c->lpt_buf) is allocated with size c->leb_size. When offs is
> near leb_size and from is non-zero, the alignment padding can push
> the memset past the end of the buffer.
>
> For example, with leb_size=4096, offs=4000, from=3584, min_io_size=1024:
> wlen=416, alen=1024, padding=608 bytes at offset 4000
> writes to offsets 4000..4607, 512 bytes past the 4096-byte buffer
Hi, Guanni
How does 'from' become '3584'(which is not aligned with min_io_size[1024])?
As far as I know, the 'from' comes from 'c->nhead_offs', which is always
aligned with min_io_size.
>
> Clamp the padding size to not exceed the buffer boundary in all 4
> instances of this pattern.
>
> The OOB write was confirmed with KASAN using a test module that
> reproduces the arithmetic with matching buffer and alignment values.
>
> Fixes: 1e51764a3c2a ("UBIFS: add new flash file system")
> Reported-by: Klaudia Kloc <klaudia@vidocsecurity.com>
> Reported-by: Dawid Moczadło <dawid@vidocsecurity.com>
> Tested-by: Jenny Guanni Qu <qguanni@gmail.com>
> Signed-off-by: Jenny Guanni Qu <qguanni@gmail.com>
> ---
> fs/ubifs/lpt_commit.c | 8 ++++----
> 1 file changed, 4 insertions(+), 4 deletions(-)
>
> diff --git a/fs/ubifs/lpt_commit.c b/fs/ubifs/lpt_commit.c
> index 07351fdce722..77a7a2cd418f 100644
> --- a/fs/ubifs/lpt_commit.c
> +++ b/fs/ubifs/lpt_commit.c
> @@ -402,7 +402,7 @@ static int write_cnodes(struct ubifs_info *c)
> wlen = offs - from;
> if (wlen) {
> alen = ALIGN(wlen, c->min_io_size);
> - memset(buf + offs, 0xff, alen - wlen);
> + memset(buf + offs, 0xff, min_t(int, alen - wlen, c->leb_size - offs));
> err = ubifs_leb_write(c, lnum, buf + from, from,
> alen);
> if (err)
> @@ -461,7 +461,7 @@ static int write_cnodes(struct ubifs_info *c)
> if (offs + c->lsave_sz > c->leb_size) {
> wlen = offs - from;
> alen = ALIGN(wlen, c->min_io_size);
> - memset(buf + offs, 0xff, alen - wlen);
> + memset(buf + offs, 0xff, min_t(int, alen - wlen, c->leb_size - offs));
> err = ubifs_leb_write(c, lnum, buf + from, from, alen);
> if (err)
> return err;
> @@ -487,7 +487,7 @@ static int write_cnodes(struct ubifs_info *c)
> if (offs + c->ltab_sz > c->leb_size) {
> wlen = offs - from;
> alen = ALIGN(wlen, c->min_io_size);
> - memset(buf + offs, 0xff, alen - wlen);
> + memset(buf + offs, 0xff, min_t(int, alen - wlen, c->leb_size - offs));
> err = ubifs_leb_write(c, lnum, buf + from, from, alen);
> if (err)
> return err;
> @@ -510,7 +510,7 @@ static int write_cnodes(struct ubifs_info *c)
> /* Write remaining data in buffer */
> wlen = offs - from;
> alen = ALIGN(wlen, c->min_io_size);
> - memset(buf + offs, 0xff, alen - wlen);
> + memset(buf + offs, 0xff, min_t(int, alen - wlen, c->leb_size - offs));
> err = ubifs_leb_write(c, lnum, buf + from, from, alen);
> if (err)
> return err;
>
______________________________________________________
Linux MTD discussion mailing list
http://lists.infradead.org/mailman/listinfo/linux-mtd/
next prev parent reply other threads:[~2026-03-14 2:18 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-03-13 22:22 [PATCH] ubifs: fix out-of-bounds write in LPT commit padding Jenny Guanni Qu
2026-03-14 2:15 ` Zhihao Cheng [this message]
2026-03-14 22:33 ` Guanni Qu
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=3fa4d791-4520-0e6d-d657-89d6fc4ba5cb@huawei.com \
--to=chengzhihao1@huawei.com \
--cc=dawid@vidocsecurity.com \
--cc=klaudia@vidocsecurity.com \
--cc=linux-mtd@lists.infradead.org \
--cc=qguanni@gmail.com \
--cc=richard@nod.at \
/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