Re: [GIT PULL v1 01/23] s390/vfio_ccw: Free all memory if cp_init() fails

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:

Pre-existing issues:
- [High] A malicious guest can trigger unbounded kernel stack recursion and host panic by providing a long chain of non-looping Transfer in Channel (TIC) CCWs.
--

--- Patch [1]: [GIT PULL v1 01/23] s390/vfio_ccw: Free all memory if cp_init() fails ---
commit 2ed68383152f2b2060167f3fc63cdb973d979950
Author: Eric Farman <farman@linux.ibm.com>

s390/vfio_ccw: Free all memory if cp_init() fails

This commit moves the ccwchain cleanup out of the recursive
ccwchain_handle_ccw() and ccwchain_loop_tic() logic into a new wrapper
function ccwchain_build_ccws(). This ensures that all segments allocated for
a channel program are freed if initialization fails.

> diff --git a/drivers/s390/cio/vfio_ccw_cp.c b/drivers/s390/cio/vfio_ccw_cp.c
> index 7561aa7d3e017..086d1b54bdb05 100644
> --- a/drivers/s390/cio/vfio_ccw_cp.c
> +++ b/drivers/s390/cio/vfio_ccw_cp.c
[ ... ]
> @@ -455,9 +455,6 @@ static int ccwchain_handle_ccw(dma32_t cda, struct channel_program *cp)
>  	/* Loop for tics on this new chain. */
>  	ret = ccwchain_loop_tic(chain, cp);
>  

[Severity: High]
This is a pre-existing issue, but can this recursive call chain lead to
unbounded kernel stack consumption?

Looking at ccwchain_handle_ccw() and ccwchain_loop_tic(), they recursively
call each other:

ccwchain_handle_ccw()
   ccwchain_loop_tic()
      ccwchain_handle_ccw()

If a guest constructs a channel program with a long sequence of Transfer
in Channel (TIC) CCWs, where each points to a distinct guest physical
address, tic_target_chain_exists() will return false.

Does this allow an unprivileged guest to trigger unbounded recursion and
overflow the host kernel stack, resulting in a host panic?

> -	if (ret)
> -		ccwchain_free(chain);
> -
>  	return ret;
>  }

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260803160924.236807-1-imbrenda@linux.ibm.com?part=1