linux-kernel - Re: [patch V2 21/36] x86/apic/msi: Use device MSI properties

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: <20211208154502.GV6385@nvidia.com>
Date:   Wed, 8 Dec 2021 11:45:02 -0400
From:   Jason Gunthorpe <jgg@...dia.com>
To:     Thomas Gleixner <tglx@...utronix.de>
Cc:     LKML <linux-kernel@...r.kernel.org>,
        Bjorn Helgaas <helgaas@...nel.org>,
        Marc Zygnier <maz@...nel.org>,
        Alex Williamson <alex.williamson@...hat.com>,
        Kevin Tian <kevin.tian@...el.com>,
        Megha Dey <megha.dey@...el.com>,
        Ashok Raj <ashok.raj@...el.com>, linux-pci@...r.kernel.org,
        Cedric Le Goater <clg@...d.org>,
        xen-devel@...ts.xenproject.org, Juergen Gross <jgross@...e.com>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        Will Deacon <will@...nel.org>,
        Santosh Shilimkar <ssantosh@...nel.org>,
        iommu@...ts.linux-foundation.org, dmaengine@...r.kernel.org,
        Stuart Yoder <stuyoder@...il.com>,
        Laurentiu Tudor <laurentiu.tudor@....com>,
        Nishanth Menon <nm@...com>, Tero Kristo <kristo@...nel.org>,
        linux-arm-kernel@...ts.infradead.org,
        Vinod Koul <vkoul@...nel.org>,
        Mark Rutland <mark.rutland@....com>,
        Robin Murphy <robin.murphy@....com>,
        Sinan Kaya <okaya@...nel.org>
Subject: Re: [patch V2 21/36] x86/apic/msi: Use device MSI properties

On Mon, Dec 06, 2021 at 11:39:29PM +0100, Thomas Gleixner wrote:
> instead of fiddling with MSI descriptors.
> 
> Signed-off-by: Thomas Gleixner <tglx@...utronix.de>
> Reviewed-by: Greg Kroah-Hartman <gregkh@...uxfoundation.org>
> Reviewed-by: Jason Gunthorpe <jgg@...dia.com>
> ---
>  arch/x86/kernel/apic/msi.c |    5 +----
>  1 file changed, 1 insertion(+), 4 deletions(-)
> 
> --- a/arch/x86/kernel/apic/msi.c
> +++ b/arch/x86/kernel/apic/msi.c
> @@ -160,11 +160,8 @@ static struct irq_chip pci_msi_controlle
>  int pci_msi_prepare(struct irq_domain *domain, struct device *dev, int nvec,
>  		    msi_alloc_info_t *arg)
>  {
> -	struct pci_dev *pdev = to_pci_dev(dev);
> -	struct msi_desc *desc = first_pci_msi_entry(pdev);
> -
>  	init_irq_alloc_info(arg, NULL);
> -	if (desc->pci.msi_attrib.is_msix) {
> +	if (msi_device_has_property(dev, MSI_PROP_PCI_MSIX)) {
>  		arg->type = X86_IRQ_ALLOC_TYPE_PCI_MSIX;
>  	} else {
>  		arg->type = X86_IRQ_ALLOC_TYPE_PCI_MSI;
>

Just thought for future

It looks like the only use of this is to link to the irq_remapping
which is only using it to get back to the physical device:

	case X86_IRQ_ALLOC_TYPE_PCI_MSI:
	case X86_IRQ_ALLOC_TYPE_PCI_MSIX:
		set_msi_sid(irte,
			    pci_real_dma_dev(msi_desc_to_pci_dev(info->desc)));

	case X86_IRQ_ALLOC_TYPE_PCI_MSI:
	case X86_IRQ_ALLOC_TYPE_PCI_MSIX:
		return get_device_id(msi_desc_to_dev(info->desc));

And this is super confusing:

static inline int get_device_id(struct device *dev)
{
	int devid;

	if (dev_is_pci(dev))
		devid = get_pci_device_id(dev);
	else
		devid = get_acpihid_device_id(dev, NULL);

	return devid;
}

How does an ACPI device have a *PCI* MSI or MSI-X ??

IMHO this makes more sense written as:

  struct device *origin_device = msi_desc_get_origin_dev(info->desc);

  if (dev_is_pci(origin_device)
      devid = get_pci_device_id(origin_device);
  else if (dev_is_acpi(origin_device))
      devid = get_acpihid_device_id(dev, NULL);

And similar in all places touching X86_IRQ_ALLOC_TYPE_PCI_MSI/X

Like this oddball thing in AMD too:

	} else if (info->type == X86_IRQ_ALLOC_TYPE_PCI_MSI ||
		   info->type == X86_IRQ_ALLOC_TYPE_PCI_MSIX) {
		bool align = (info->type == X86_IRQ_ALLOC_TYPE_PCI_MSI);

		index = alloc_irq_index(devid, nr_irqs, align,
					msi_desc_to_pci_dev(info->desc));
	} else {
		index = alloc_irq_index(devid, nr_irqs, false, NULL);

This should just use a dev and inside alloc_irq_table do the dev_is_pci()
thing to guard the pci_for_each_dma_alias()

Then just call it X86_IRQ_ALLOC_TYPE_DEVICE (ie allocated for a struct device)

Jason