[Public]

-----Original Message----- From: Greg KH gregkh@linuxfoundation.org Sent: Tuesday, December 3, 2024 3:50 AM To: Zhang Zekun zhangzekun11@huawei.com Cc: cve@kernel.org; linux-cve-announce@vger.kernel.org; stable@vger.kernel.org; Wang, Yang(Kevin) KevinYang.Wang@amd.com; Deucher, Alexander Alexander.Deucher@amd.com; liuyongqiang13@huawei.com Subject: Re: Possible wrong fix patch for some stable branches

On Tue, Dec 03, 2024 at 10:06:51AM +0800, Zhang Zekun wrote:

...

Hi, All

The mainline patch to fix CVE-2024-50282 add a check to fix a potential buffer

overflow issue in amdgpu_debugfs_gprwave_read() which is introduced in commit 553f973a0d7b ("drm/amd/amdgpu: Update debugfs for XCC support (v3)"), but some linux-stable fix patches add the check in some other funcitons, is something wrong here?

...

Stable version which contain the suspicious patches: Fixed in 4.19.324 with commit 673bdb4200c0: Fixed in amdgpu_debugfs_regs_smc_read() Fixed in 5.4.286 with commit 7ccd781794d2: Fixed in amdgpu_debugfs_regs_smc_read() Fixed in 5.10.230 with commit 17f5f18085ac: Fixed in amdgpu_debugfs_regs_pcie_write() Fixed in 5.15.172 with commit aaf6160a4b7f: Fixed in amdgpu_debugfs_regs_didt_write() Fixed in 6.1.117 with commit 25d7e84343e1: Fixed in amdgpu_debugfs_regs_pcie_write()

Link to mainline fix patch: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/com mit/?id=4d75b9468021c73108b4439794d69e892b1d24e3

If this is incorrect, can you send patches fixing this up?

All of these should be reverted: 4.19.324 with commit 673bdb4200c0 5.4.286 with commit 7ccd781794d2 5.10.230 with commit 17f5f18085ac 5.15.172 with commit aaf6160a4b7f 6.1.117 with commit 25d7e84343e1

The function which was patched didn't exist in kernel 6.1 and older kernels and the patches ended up patching a different function in the same file instead.

Thanks,

Alex

thanks,

greg k-h