linux-kernel - Re: [PATCH] ath11k: Fix potential RCU dereference issue in ath11k_debugfs_htt_ext_stats

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <29bf2e9c-ca24-46d1-93e4-3b3f3f812705@quicinc.com>
Date: Fri, 30 Aug 2024 09:01:36 -0700
From: Jeff Johnson <quic_jjohnson@...cinc.com>
To: Jiawei Ye <jiawei.ye@...mail.com>, <kvalo@...nel.org>,
        <jjohnson@...nel.org>, <corbet@....net>
CC: <linux-wireless@...r.kernel.org>, <ath11k@...ts.infradead.org>,
        <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] ath11k: Fix potential RCU dereference issue in
 ath11k_debugfs_htt_ext_stats_handler

On 8/30/2024 5:02 AM, Jiawei Ye wrote:
> In the `ath11k_debugfs_htt_ext_stats_handler` function, the `ar` pointer
> obtained via RCU lock is accessed after the RCU read-side critical
> section might be unlocked. According to RCU usage rules, this is illegal.
> Reusing this pointer can lead to unpredictable behavior, including
> accessing memory that has been updated or causing use-after-free issues.
> The `ath12k_debugfs_htt_ext_stats_handler` function in the
> `drivers/net/wireless/ath/ath12k/debugfs_htt_stats.c` file provides a good
> example to follow for addressing this issue.
> 
> This possible bug was identified using a static analysis tool developed
> by myself, specifically designed to detect RCU-related issues.
> 
> To address this issue, the RCU read lock is now kept until all accesses
> to the `ar` pointer are completed. A `goto exit` statement is introduced
> to ensure that the RCU read unlock is called appropriately, regardless of
> the function's exit path.

This analysis is incorrect since ar is not an RCU-protected structure

The rcu_read_lock() is required internally within
ath11k_mac_get_ar_by_pdev_id() when retrieving the RCU-protected pdev pointer.

So NAK this patch.