PT-2026-90267 · Linux · Linux
CVE-2026-89551
·
Published
2026-09-11
·
Updated
2026-09-11
None
No severity ratings or metrics are available. When they are, we'll update the corresponding info on the page.
In the Linux kernel, the following vulnerability has been resolved:
SUNRPC: xdr buf trim: clamp buf->len to avoid underflow
xdr buf trim() trims
len bytes from the tail of an xdr buf by
walking the tail, pages, and head iovecs. Each per-section step
uses min t() so it never removes more bytes than that section
holds, but the final accounting at the fix len label subtracts the
total bytes actually consumed from buf->len without any clamp:fix len:
buf->len -= (len - trim);
When the caller has set buf->len to a value smaller than the sum
of the iov lens, (len - trim) can exceed buf->len and the unsigned
subtraction wraps to near UINT MAX. gss krb5 unwrap v2() reaches
xdr buf trim() in exactly that state:
buf->head[0].iov len -= GSS KRB5 TOK HDR LEN + headskip;
buf->len = len - (GSS KRB5 TOK HDR LEN + headskip);
xdr buf trim(buf, ec + GSS KRB5 TOK HDR LEN + tailskip);
buf->len is a small wire-derived value while the iov lens are at
page scale, so the per-section loops legitimately consume far more
bytes than buf->len records. The wrapped buf->len then propagates
as the authoritative stream bound into every downstream XDR
decoder.
Fix by clamping the decrement so buf->len bottoms out at zero:
buf->len -= min t(unsigned int, buf->len, len - trim);
On the normal path where the iov lens sum to buf->len, (len - trim)
is always <= buf->len and the result is identical to before. No
callers change behavior outside the underflow case.
Found an issue in the description? Have something to add? Feel free to write us 👾
Related Identifiers
Affected Products
Linux