[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <87ldwuqa3q.ffs@tglx>
Date: Thu, 05 Dec 2024 19:05:13 +0100
From: Thomas Gleixner <tglx@...utronix.de>
To: Frank Li <Frank.li@....com>
Cc: Manivannan Sadhasivam <manivannan.sadhasivam@...aro.org>, Krzysztof
WilczyĆski <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>, "Rafael J. Wysocki"
<rafael@...nel.org>, Anup Patel <apatel@...tanamicro.com>, Marc Zyngier
<maz@...nel.org>, linux-kernel@...r.kernel.org, linux-pci@...r.kernel.org,
imx@...ts.linux.dev, Niklas Cassel <cassel@...nel.org>,
dlemoal@...nel.org, jdmason@...zu.us, linux-arm-kernel@...ts.infradead.org
Subject: Re: [PATCH v10 3/7] PCI: endpoint: pci-ep-msi: Add MSI address/data
pair mutable check
On Thu, Dec 05 2024 at 09:56, Frank Li wrote:
> On Thu, Dec 05, 2024 at 02:10:55PM +0100, Thomas Gleixner wrote:
>> You want a MSI_FLAG_MSG_IMMUTABLE and set that on the domains which
>> provide it. That way you ensure that someone looked at the domain to
>> validate it.
>
> Okay, at beginning I think most MSI controller is immutable. So I use
> MSI_FLAG_MSG_MUTABLE.
If you want to do that then _you_ have to go through every single
interrupt controllers, validate and opt-out in case it does change the
message. Otherwise that flag is completely pointless.
Instead of adding the IMMUTABLE flag for one controller you know and
then let others who want to utilize this amend their controllers.
Opt-in is less work and more safe than opt-out. See?
Thanks,
tglx
Powered by blists - more mailing lists