linux-kernel - Re: [RFT PATCH 1/2] xhci: Reconfigure endpoint 0 max packet size only during endpoint reset

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]

Date:   Wed, 29 Nov 2023 16:57:23 +0200
From:   Mathias Nyman <mathias.nyman@...ux.intel.com>
To:     Kuen-Han Tsai <khtsai@...gle.com>
Cc:     gregkh@...uxfoundation.org, linux-usb@...r.kernel.org,
        linux-kernel@...r.kernel.org, stern@...land.harvard.edu
Subject: Re: [RFT PATCH 1/2] xhci: Reconfigure endpoint 0 max packet size only
 during endpoint reset

>> +       ep_index = xhci_get_endpoint_index(&host_ep->desc);
>> +
>> +       /*
>> +        * Usb core assumes a max packet value for ep0 on FS devices until the
>> +        * real value is read from the descriptor. Core resets Ep0 if values
>> +        * mismatch. Reconfigure the xhci ep0 endpoint context here in that case
>> +        */
>> +       if (usb_endpoint_xfer_control(&host_ep->desc) && ep_index == 0) {
>> +               udev = container_of(host_ep, struct usb_device, ep0);
>> +               if (udev->speed == USB_SPEED_FULL)
>> +                       xhci_check_ep0_maxpacket(xhci, xhci->devs[udev->slot_id]);
>> +               /* Nothing else should be done here for ep0 during ep reset */
>> +               return;
>> +       }
>> +
> 
> Could there be a race condition between the xhci_endpoint_reset() and
> xhci_free_dev() functions, resulting in the xhci->devs[udev->slot_id]
> becoming null?
> If so, a null pointer dereference will happen in
> xhci_check_ep0_maxpacket() when accessing vdev->out_ctx.

should not race. xhci_free_dev() and xhci_endpoint_reset() for endpoint 0 should only
be called by hub driver hub_free_dev() and usb_ep0_reinit() respectively.

Hub driver takes care of concurrency for these

Thanks
Mathias