linux-kernel - Re: [PATCH v8 2/6] PCI: endpoint: Add RC-to-EP doorbell support using platform MSI controller

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: <20241127062004.oruhwkhj4zrvjx25@thinkpad>
Date: Wed, 27 Nov 2024 11:50:04 +0530
From: Manivannan Sadhasivam <manivannan.sadhasivam@...aro.org>
To: Frank Li <Frank.li@....com>
Cc: Krzysztof Wilczyński <kw@...ux.com>,
	Kishon Vijay Abraham I <kishon@...nel.org>,
	Bjorn Helgaas <bhelgaas@...gle.com>, Arnd Bergmann <arnd@...db.de>,
	Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
	linux-kernel@...r.kernel.org, linux-pci@...r.kernel.org,
	imx@...ts.linux.dev, Niklas Cassel <cassel@...nel.org>,
	dlemoal@...nel.org, maz@...nel.org, tglx@...utronix.de,
	jdmason@...zu.us
Subject: Re: [PATCH v8 2/6] PCI: endpoint: Add RC-to-EP doorbell support
 using platform MSI controller

On Tue, Nov 26, 2024 at 12:19:01PM -0500, Frank Li wrote:

[...]

> > > > > +	/* Only support one EPF for doorbell */
> > > > > +	epf = list_first_entry_or_null(&epc->pci_epf, struct pci_epf, list);
> > > >
> > > > Why don't you impose this restriction in pci_epf_alloc_doorbell() itself?
> > >
> > > This is callback from platform_device_msi_init_and_alloc_irqs(). So it is
> > > actually restriction at pci_epf_alloc_doorbell().
> > >
> >
> > I don't know how to parse this last sentence. But my question was why don't you
> > impose this one EPF restriction in pci_epf_alloc_doorbell() before allocating
> > the MSI domain using platform_device_msi_init_and_alloc_irqs()? This way if
> > someone calls pci_epf_alloc_doorbell() multi EPF, it will fail.
> 
> Yes, it is limitation for current platfrom msi framework. It is totally
> equal.
> 
> call stack is
> 	pci_epf_alloc_doorbell()
> 	     platform_device_msi_init_and_alloc_irqs()
> 		...
> 		pci_epc_write_msi_msg()
> 
> 
> 
> It is totally equal return at pci_epc_write_msi_msg() or return before
> platform_device_msi_init_and_alloc_irqs().
> 

No, these two are not the same. pci_epc_write_msi_msg() will only be called
when enabling the MSI, which is too late IMO. Why are you insisting in
calling platform_device_msi_init_and_alloc_irqs() for multi EPF? I don't quite
understand.

If the platform cannot support it, then it should not be called in first place.

> why check epf in pci_epc_write_msi_msg() is because it use epf in here.
> pci_epf_alloc_doorbell() never use epf variable. If check unused variable
> in pci_epf_alloc_doorbell(), user don't know why and what's exactly
> restrict it.
> 

This is not a good argument and doesn't make sense, sorry. You should not call
platform_device_msi_init_and_alloc_irqs() for multi EPF.

- Mani

-- 
மணிவண்ணன் சதாசிவம்