linux-kernel - Re: qrtr/mhi: NULL-deref with in-kernel pd-mapper

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: <Zy3jeixo2uTooRGo@hovoldconsulting.com>
Date: Fri, 8 Nov 2024 11:10:02 +0100
From: Johan Hovold <johan@...nel.org>
To: Chris Lew <quic_clew@...cinc.com>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@...aro.org>,
	Bjorn Andersson <bjorn.andersson@....qualcomm.com>,
	Qiang Yu <quic_qianyu@...cinc.com>,
	Jeffrey Hugo <quic_jhugo@...cinc.com>,
	Krishna chaitanya chundru <quic_krichai@...cinc.com>,
	Konrad Dybcio <konradybcio@...nel.org>, mhi@...ts.linux.dev,
	linux-arm-msm@...r.kernel.org, linux-remoteproc@...r.kernel.org,
	linux-kernel@...r.kernel.org
Subject: Re: qrtr/mhi: NULL-deref with in-kernel pd-mapper

On Thu, Nov 07, 2024 at 05:19:19PM +0100, Johan Hovold wrote:
> On Tue, Nov 05, 2024 at 10:26:40AM -0800, Chris Lew wrote:
> > On 11/4/2024 9:08 PM, Johan Hovold wrote:

> > > I naively tried adding a sleep after registering the endpoint, but that
> > > is at least not sufficient to trigger the NULL-deref.

> No, neither tqftpserv or diag-router are used here, but after digging
> through the code it seems my hunch about this being related to the
> in-kernel pd-mapper was correct.
> 
> The qrtr worker, qrtr_ns_worker(), is called when the in-kernel
> pd-mapper adds the server, and processing the QRTR_TYPE_NEW_SERVER
> command eventually ends up in mhi_gen_tre() for the modem:
> 
> [    9.026694] qcom_pdm_start - adding server
> [    9.034684] ctrl_cmd_new_server - service = 0x40, instance = 0x101, node_id = 1, port = 0
> [    9.042155] mhi-pci-generic 0005:01:00.0: mhi_gen_tre - buf_info = ffff800080d4d038, offset_of(buf_info->used) = 34

> And I can indeed imagine that leading to the NULL deref in case the
> endpoint is registered before being fully set up.

I've been able to reproduce the issue twice now by instrumenting the
code to increase the race window. Specifically, I added a sleep in
mhi_init_chan_ctxt() after allocating the ring buffers but before
initialising the wp pointers. And when the in-kernel pd-mapper is
started in that window, we hit the NULL-deref:

        [    8.593582] mhi-pci-generic 0005:01:00.0: mhi_init_chan_ctxt - ring allocated (IPCR:20), buf_ring->base = ffff800080d55000
        [    8.598902] mhi_net mhi0_IP_SW0: mhi_prepare_channel - channel started (46), dir = 1
        [    8.612888] mhi_net mhi0_IP_SW0: mhi_prepare_channel - channel started (47), dir = 2
        [    8.614767] qcom_pdm_start - adding server
        [    8.615302] ctrl_cmd_new_server - service = 0x40, instance = 0x101, node_id = 1, port = 0
        [    8.615388] mhi-pci-generic 0005:01:00.0: mhi_gen_tre - buf_info = 0000000000000000 (IPCR:20)
        [    8.615402] Unable to handle kernel NULL pointer dereference at virtual address 0000000000000034
        ...
        [    8.615541] Call trace:
        [    8.615542]  mhi_gen_tre+0x68/0x248 [mhi]
        [    8.615544]  mhi_queue+0x74/0x194 [mhi]
        [    8.615546]  mhi_queue_skb+0x5c/0x8c [mhi]
        [    8.615549]  qcom_mhi_qrtr_send+0x6c/0x160 [qrtr_mhi]
        [    8.615551]  qrtr_node_enqueue+0xd0/0x4a0 [qrtr]
        [    8.615553]  qrtr_bcast_enqueue+0x78/0xe8 [qrtr]
        [    8.615554]  qrtr_sendmsg+0x15c/0x33c [qrtr]
        [    8.615555]  sock_sendmsg+0xc0/0xec
        [    8.615560]  kernel_sendmsg+0x30/0x40
        [    8.615561]  service_announce_new+0xbc/0x1c4 [qrtr]
        [    8.615563]  qrtr_ns_worker+0x754/0x7d4 [qrtr]

Johan