linux-kernel - Re: [PATCH 5/5] usb: dwc3: dwc3-omap: Disable/Enable core interrupts in Suspend/Resume

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, 14 May 2014 11:29:43 +0530
From:	George Cherian <george.cherian@...com>
To:	<balbi@...com>
CC:	<linux-kernel@...r.kernel.org>, <linux-omap@...r.kernel.org>,
	<linux-usb@...r.kernel.org>, <gregkh@...uxfoundation.org>,
	<rogerq@...com>, <kishon@...com>
Subject: Re: [PATCH 5/5] usb: dwc3: dwc3-omap: Disable/Enable core interrupts
 in Suspend/Resume

On 5/13/2014 9:20 PM, Felipe Balbi wrote:
> Hi,
>
> On Thu, May 08, 2014 at 03:03:07PM +0530, George Cherian wrote:
>> Enabling the core interrupts in complete is too late for XHCI, and stops
>> xhci from proper operation. So remove prepare and complete and disable/enable
> isn't this a bug in xhci ? I mean the driver should make no assumption
> as to when IRQs are enabled, why do we need to enable IRQs earlier when
> the device is only considered "ready for use" after ->complete()
> finishes executing ?
I dont think its a bug in xhci. In case of xhci-pci driver it actually 
does an
hcd->driver->pci_suspend (xhci_suspend) followed by synchronize_irq()
and the does a pci_disable_device().  In resume path it calls 
pci_enable_device()
followed by hcd->driver->pci_resume(xhci_resume).

In case of dwc3-omap we do have a wrapper register which can still 
disable the XHCI IRQs
even though the xhci driver enables the interrupts internally.

Now dwc3-omap wrapper driver should not actually fiddle with the core 
Interrupt
enable/disable except in probe/remove.

>  From documentation we have:
>
> 107  * @complete: Undo the changes made by @prepare().  This method is executed for
> 108  *      all kinds of resume transitions, following one of the resume callbacks:
> 109  *      @resume(), @thaw(), @restore().  Also called if the state transition
> 110  *      fails before the driver's suspend callback: @suspend(), @freeze() or
> 111  *      @poweroff(), can be executed (e.g. if the suspend callback fails for one
> 112  *      of the other devices that the PM core has unsuccessfully attempted to
> 113  *      suspend earlier).
> 114  *      The PM core executes subsystem-level @complete() after it has executed
> 115  *      the appropriate resume callbacks for all devices.
>
> which tells me that using ->complete() to reenable IRQs is ok here.
> Specially when you consider that the role of ->prepare() is to prevent
> new children from being created and, for a USB host, that means we
> should prevent hub port changes.
Probably the patch should have been to still keep the complete/prepare 
in place
but not disable the core interrupts, rather enable/disable only the 
wrapper interrupt.
> cheers
>


-- 
-George

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/