linux-kernel - Re: [PATCH v5] PCI: vmd: Add the module param to adjust MSI mode

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]

Date:   Fri, 28 Apr 2023 14:58:22 -0500
From:   Bjorn Helgaas <helgaas@...nel.org>
To:     korantwork@...il.com
Cc:     nirmal.patel@...ux.intel.com, kbusch@...nel.org,
        jonathan.derrick@...ux.dev, lpieralisi@...nel.org,
        linux-pci@...r.kernel.org, linux-kernel@...r.kernel.org,
        Xinghui Li <korantli@...cent.com>
Subject: Re: [PATCH v5] PCI: vmd: Add the module param to adjust MSI mode

On Fri, Apr 28, 2023 at 01:40:36PM -0500, Bjorn Helgaas wrote:
> On Thu, Apr 20, 2023 at 03:09:14PM +0800, korantwork@...il.com wrote:
> > From: Xinghui Li <korantli@...cent.com>

> What if you made boolean parameters like these:
> 
>   no_msi_remap
> 
>     If the VMD supports it, disable VMD MSI-X remapping.  This
>     improves interrupt performance because child device interrupts
>     avoid the VMD MSI-X domain interrupt handler.
> 
>   msi_remap
> 
>     Remap child MSI-X interrupts into VMD MSI-X interrupts.  This
>     limits the number of MSI-X vectors available to the whole child
>     device domain to the number of VMD MSI-X interrupts.

I guess having two parameters that affect the same feature is also
confusing.  Maybe just "msi_remap=0" or "msi_remap=1" or something?

I think what makes "disable_msi_bypass=0" hard is that "MSI bypass" by
itself is a negative feature (the positive activity is MSI remapping),
and disabling bypass gets us back to the positive "MSI remapping"
situation, and "disable_msi_bypass=0" negates that again, so we're
back to ... uh ... let's see ... we are not disabling the bypass of
MSI remapping, so I guess MSI remapping would be *enabled*?  Is that
right?

Bjorn