linux-kernel - Re: [PATCH 1/4] acpi,pci,irq: reduce resource requirements

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

Message-ID: <20160315023632.GA23940@localhost>
Date:	Mon, 14 Mar 2016 21:36:32 -0500
From:	Bjorn Helgaas <helgaas@...nel.org>
To:	Sinan Kaya <okaya@...eaurora.org>
Cc:	linux-acpi@...r.kernel.org, timur@...eaurora.org,
	cov@...eaurora.org, linux-pci@...r.kernel.org,
	ravikanth.nalla@....com, lenb@...nel.org, harish.k@....com,
	ashwin.reghunandanan@....com, bhelgaas@...gle.com,
	rjw@...ysocki.net, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 1/4] acpi,pci,irq: reduce resource requirements

On Mon, Mar 14, 2016 at 10:28:11PM -0400, Sinan Kaya wrote:
> On 3/14/2016 9:48 PM, Bjorn Helgaas wrote:
> 
> Yes. I was talking about PCIe on ARM64. 
> 
> > If you go to Fry's and buy a conventional PCI card, it is going to
> > pull INTA# low to assert an interrupt.  It doesn't matter whether you
> > put it in an x86 system or an arm64 system.
> 
> I don't see INTA# of the PCIe card at the system level. The PCIe wire 
> interrupt gets converted to the system level interrupt by the PCIe controller.

That's why I said *conventional PCI*.  If you have a conventional PCI
device below either a conventional PCI host controller or a PCIe-to-PCI
bridge, there are real INTx wires, not virtual wires, and they are
level/low.  But I think you pointed out the key below (that the
Interrupt resource in a PNP0C0F device encodes the trigger type).

> >> > I pasted the code here again. It looks like you want to validate that
> >> > PCI interrupts are always level low. 
> > I don't really care whether PCI interrupts are always level low.  What
> > matters is that the PCI interrupt line matches the configuration of
> > the interrupt controller input.
> > 
> 
> Agreed. But the interrupt controller configuration is system specific. How would
> you check interrupt line against what the interrupt controller requires
> on each architecture as this is common code?
> 
> 
> > If the PCI interrupt can be a different type, e.g., level high, and
> > there's a way to discover that, we can check that against the
> > interrupt controller configuration.
> > 
> > This is all in the context of conventional PCI, and we're probably
> > talking about arm64 PCIe systems, not conventional PCI. 
> 
> INTx interrupts are TLP messages on PCIe as you already know. There is no INTA 
> interrupt wire.

Yes, that's why I mentioned PCIe sec 2.2.8.1 below.

> "6.1.2. PCI Compatible INTx Emulation" section of the PCIe spec describes
> INTx emulation on PCIe.
>  ...
> 
> > I'm not sure what an Interrupt Link device means in PCIe.  I suppose it would have
> > to connect an INTx virtual wire to a system interrupt?  The PCIe spec
> > says this sort of mapping is system implementation specific (r3.0, sec
> > 2.2.8.1).
> 
> The INTx messages are converted to the system interrupt by the PCIe controller.
> The interrupt type between the PCIe controller and the ARM GIC interrupt 
> controller is dictated by the ARM GIC Interrupt Controller Specification for
> ARM64.
> 
> Here is what ACPI table looks like for reference
> 
> 	Name(_PRT, Package(){
> 		Package(){0x0FFFF, 0, \_SB.LN0A, 0},  // Slot 0, INTA
> 		Package(){0x0FFFF, 1, \_SB.LN0B, 0},  // Slot 0, INTB
> 		Package(){0x0FFFF, 2, \_SB.LN0C, 0},  // Slot 0, INTC
> 		Package(){0x0FFFF, 3, \_SB.LN0D, 0}   // Slot 0, INTD
> 	}) 
> 
> Device(LN0A){
> 	Name(_HID, EISAID("PNP0C0F")) // PCI interrupt link
> 	Name(_UID, 1)
> 	Name(_PRS, ResourceTemplate(){
> 	Interrupt(ResourceProducer, Level, ActiveHigh, Exclusive, , ,) {0x123}
> 	})

I forgot that the link already include the trigger mode in it.  Maybe we
can check for that instead of assuming level/low.

Bjorn