Look up another CVE · open CVE-2024-53094 in the interactive tracker
CVE-2024-53094
RDMA/siw: Add sendpage_ok() check to disable MSG_SPLICE_PAGES
Published 2024-11-21 (23 months ago) by the Linux kernel CNA. Subsystem drivers/infiniband. Debian urgency: not yet assigned.
Debian does not build this code. The per-release build detail is under "Where it lives" below.
What to do, per Debian release
Debian's own status for CVE-2024-53094 in every release this site tracks, and what each one means in practice. Every status below is the one the Debian Security Team publishes for this CVE. source
| Release | Status | Version | What this means |
|---|---|---|---|
| Debian 12bookworm | vulnerable | 6.1.190-1 | No fix published anywhere yet, neither in Debian nor upstream for the series bookworm tracks. Watch the Debian tracker page for this CVE. Advisories for bookworm have historically landed a median of 23 days after publication. |
| Debian 13trixie | fixed | 6.11.9-1 | Update the kernel and reboot. Fixed in 6.11.9-1. A running kernel keeps the old code until the machine restarts. |
| Debian 14forky | fixed | 6.11.9-1 | Update the kernel and reboot. Fixed in 6.11.9-1. A running kernel keeps the old code until the machine restarts. |
| Debian unstablesid | fixed | 6.11.9-1 | Update the kernel and reboot. Fixed in 6.11.9-1. A running kernel keeps the old code until the machine restarts. |
The vulnerability
What the bug is
While running ISER over SIW, the initiator machine encounters a warning from skb_splice_from_iter() indicating that a slab page is being used in send_page. To address this, it is better to add a sendpage_ok() check within the driver itself, and if it returns 0, then MSG_SPLICE_PAGES flag should be disabled before entering the network stack. A similar issue has been discussed for NVMe in this thread: https://lore.kernel.org/all/20240530142417.146696-1-ofir.gal@volumez.com/ WARNING: CPU: 0 PID: 5342 at net/core/skbuff.c:7140 skb_splice_from_iter+0x173/0x320 Call Trace: tcp_sendmsg_locked+0x368/0xe40 siw_tx_hdt+0x695/0xa40 [siw] siw_qp_sq_process+0x102/0xb00 [siw] siw_sq_resume+0x39/0x110 [siw] siw_run_sq+0x74/0x160 [siw] kthread+0xd2/0x100 ret_from_fork+0x34/0x40 ret_from_fork_asm+0x1a/0x30
Where it lives
The fix touches these files. If the subsystem is one you do not use, the practical exposure is lower, though the package is still the vulnerable one.
drivers/infiniband/sw/siw/siw_qp_tx.c
- What this is
- SOFT-IWARP DRIVER (siw)MAINTAINERS section matched on drivers/infiniband/sw/siw/ · linux-rdma@vger.kernel.org · Supported
- Kernel module
- none to look for. The file would have gone into siw, and Debian does not compile it into that module, so finding siw in lsmod would say nothing about this CVE.
Check it on the machine
grep -wE 'CONFIG_INFINIBAND|CONFIG_RDMA_SIW' /boot/config-$(uname -r)Reads the booted kernel's own configuration, on the machine and nowhere else. =y is built in, =m is a module, and an is not set line means the code is not there.
Show whether Debian builds this, per release
Debian does not build this code. In 4 releases below, every kernel flavour Debian publishes has the switch off, so the vulnerable file is compiled into no kernel binary Debian ships.
- bookworm, trixie, forky, sidn
- n
- not enabled on any flavour - this release does not build the code
- Build switches
- CONFIG_INFINIBAND, CONFIG_RDMA_SIW
- Reads as
- all on means the file is compiled; any one off means it is not.
From the configuration Debian publishes for every kernel flavour it builds. It covers only the architectures Debian ships a linux-config package for, and nothing about a kernel you built yourself. Debian's own status above stays authoritative about the source package.MAINTAINERS kbuild Makefiles Debian linux-config packages
Signals
- Published
- 2024-11-21 (23 months ago)source
- CVSS
- 9.8 critical
CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:Hsource recompute it - Attack vector
- network (AV:N)source
- EPSS
- 0.59%, higher than 47.0% of all CVEssource
- CISA KEV
- not listedthe KEV catalogue
- Debian urgency
- not yet assignedsource
- Subsystem
- drivers/infiniband
Upstream fixes
The versions the CNA records as introducing and fixing this, and the commit for each. source
| Introduced | Fixed in | Commit |
|---|---|---|
| 5.3 | 6.6.62 | 3406bfc813a9 |
| 5.3 | 6.11.9 | bb5738957d92 |
| 5.3 | 6.12 | 4e1e3dd88a4c |
Why the CNA scored it this way
Quoted from the kernel CNA's own scoring file, word for word. A highlighted metric carries the most severe value CVSS v3.1 defines for it, which is what pushes the score up. source
- AV:NAttack vector, network
- siw is an RDMA-over-TCP transport, and the vulnerable transmit path is driven by remote input — `siw_init_rresp()` takes the read length, local address and stag directly from a peer's RDMA Read Request off the TCP stream, and any length ≥ PAGE_SIZE selects the `MSG_SPLICE_PAGES` zero-copy path. The reported failure also occurs in ordinary remote-peer-driven iSER-over-siw traffic.
- AC:LAttack complexity, low
- No race or unpredictable state is involved; slab-backed kernel-ULP buffers are the normal case (the reporter hit it in routine ISER-over-SIW operation), and the peer simply issues a read/transfer of at least PAGE_SIZE against an advertised region. `zcopy_tx` is hardcoded true and READ_RESPONSE WQEs are always unsignalled, so the zero-copy path is taken deterministically.
- PR:NPrivileges required, none
- iWARP/MPA has no in-kernel authentication, and the RX path processes an inbound RDMA Read Request and queues the READ_RESPONSE with `perms = 0` — no access-rights or credential check gates the vulnerable transmit path. A peer at the far end of the RDMA connection holds no privileges on the victim host.
- UI:NUser interaction, none
- Triggering requires only normal or attacker-generated RDMA traffic on an established connection; no action by a local user or administrator on the victim system is needed.
- S:UScope, unchanged
- The fault, the corrupted page references and the resulting crash all occur within the victim kernel's own security authority; no VM, IOMMU or sandbox boundary is crossed.
- C:HConfidentiality impact, high
- Handing non-`sendpage_ok()` pages to the network stack is exactly the condition `sendpage_ok()` exists to prevent — TCP retains references to slab pages that the allocator may recycle, so retransmitted segments can leak recycled kernel slab contents to the remote peer, and the resulting refcount corruption is a memory-corruption class defect leverageable for disclosure.
- I:HIntegrity impact, high
- `get_page()`/`put_page()` on slab pages corrupts the page refcount and can release a page that is still in use (the "delayed Oops" documented in include/linux/net.h), a use-after-free-class primitive; data transmitted from buffers the ULP has already freed or reused also corrupts on-wire data and MPA CRCs.
- A:HAvailability impact, high
- The bad page trips `WARN_ON_ONCE()` in `skb_splice_from_iter()`, which is a kernel panic on any system with `panic_on_warn=1`, and returns `-EIO`, which aborts the FPDU, drives the QP to error state and tears down the RDMA storage/NFS session — repeatably, at the remote peer's discretion.