linux-kernel - Re: [PATCH] xhci: Fix soft lockup in xhci_pci_probe path when XHCI_STATE

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 for Android: free password hash cracker in your pocket

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

Date:	Mon, 01 Aug 2016 13:27:38 -0300
From:	Gabriel Krisman Bertazi <krisman@...ux.vnet.ibm.com>
To:	Kamal Mostafa <kamal@...onical.com>
Cc:	Mathias Nyman <mathias.nyman@...el.com>,
	Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
	linux-usb@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] xhci: Fix soft lockup in xhci_pci_probe path when XHCI_STATE_HALTED

Kamal Mostafa <kamal@...onical.com> writes:

> Commit 27a41a83ec54 ("xhci: Cleanup only when releasing primary hcd")
> causes a soft lockup at boot when XHCI_STATE_HALTED, preventing
> VirtualBox 5.1.x from booting if USB3.0 is enabled.
>
> Revert to allowing xhci_irq to handle the interrupt when
> XHCI_STATE_HALTED but not XHCI_STATE_DYING.
>
> Fixes: 27a41a83ec54 ("xhci: Cleanup only when releasing primary hcd")
> BugLink: https://bugs.launchpad.net/bugs/1604058
> Signed-off-by: Kamal Mostafa <kamal@...onical.com>
> Cc: <stable@...r.kernel.org> #v4.3+
> Cc: Gabriel Krisman Bertazi <krisman@...ux.vnet.ibm.com>
> Acked-by: Tim Gardner <tim.gardner@...onical.com>
> Acked-by: Stefan Bader <stefan.bader@...onical.com>
> ---
>  drivers/usb/host/xhci-ring.c | 5 +++--
>  1 file changed, 3 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/usb/host/xhci-ring.c b/drivers/usb/host/xhci-ring.c
> index 918e0c7..1d36a5f 100644
> --- a/drivers/usb/host/xhci-ring.c
> +++ b/drivers/usb/host/xhci-ring.c
> @@ -2751,8 +2751,7 @@ hw_died:
>  		writel(irq_pending, &xhci->ir_set->irq_pending);
>  	}
>
> -	if (xhci->xhc_state & XHCI_STATE_DYING ||
> -	    xhci->xhc_state & XHCI_STATE_HALTED) {
> +	if (xhci->xhc_state & XHCI_STATE_DYING) {
>  		xhci_dbg(xhci, "xHCI dying, ignoring interrupt. "
>  				"Shouldn't IRQs be disabled?\n");
>  		/* Clear the event handler busy flag (RW1C);
> @@ -2764,6 +2763,8 @@ hw_died:
>  		spin_unlock(&xhci->lock);
>
>  		return IRQ_HANDLED;
> +	} else if (xhci->xhc_state & XHCI_STATE_HALTED) {
> +		xhci_dbg(xhci, "xHCI halted, handling interrupt.\n");
>  	}

Returning early when XHCI_STATE_HALTED is no longer necessary since my
commit ensures we only free the HCD after disabling the primary HCD.

I think we can drop the 'else if' leg all together and only test for
XHCI_STATE_DYING here.

-- 
Gabriel Krisman Bertazi