netdev - Re: [PATCH V3 03/10] PCI/TPH: Add pci=notph to prevent use of TPH

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: <20240723224102.GA779599@bhelgaas>
Date: Tue, 23 Jul 2024 17:41:02 -0500
From: Bjorn Helgaas <helgaas@...nel.org>
To: Wei Huang <wei.huang2@....com>
Cc: linux-pci@...r.kernel.org, linux-kernel@...r.kernel.org,
	linux-doc@...r.kernel.org, netdev@...r.kernel.org,
	Jonathan.Cameron@...wei.com, corbet@....net, davem@...emloft.net,
	edumazet@...gle.com, kuba@...nel.org, pabeni@...hat.com,
	alex.williamson@...hat.com, gospo@...adcom.com,
	michael.chan@...adcom.com, ajit.khaparde@...adcom.com,
	somnath.kotur@...adcom.com, andrew.gospodarek@...adcom.com,
	manoj.panicker2@....com, Eric.VanTassell@....com,
	vadim.fedorenko@...ux.dev, horms@...nel.org, bagasdotme@...il.com,
	bhelgaas@...gle.com
Subject: Re: [PATCH V3 03/10] PCI/TPH: Add pci=notph to prevent use of TPH

On Wed, Jul 17, 2024 at 03:55:04PM -0500, Wei Huang wrote:
> TLP headers with incorrect steering tags (e.g. caused by buggy driver)
> can potentially cause issues when the system hardware consumes the tags.

Hmm.  What kind of issues?  Crash?  Data corruption?  Poor
performance?

> Provide a kernel option, with related helper functions, to completely
> prevent TPH from being enabled.

Also would be nice to have a hint about the difference between "notph"
and "nostmode".  Maybe that goes in the "nostmode" patch?  I'm not
super clear on all the differences here.

> +++ b/Documentation/admin-guide/kernel-parameters.txt
> @@ -4655,6 +4655,7 @@
>  		nomio		[S390] Do not use MIO instructions.
>  		norid		[S390] ignore the RID field and force use of
>  				one PCI domain per PCI function
> +		notph		[PCIE] Do not use PCIe TPH

Expand acronym here since there's no helpful context.  Can also
include "(TPH)" if that's useful.

> @@ -322,8 +323,12 @@ static long local_pci_probe(void *_ddi)
>  	pm_runtime_get_sync(dev);
>  	pci_dev->driver = pci_drv;
>  	rc = pci_drv->probe(pci_dev, ddi->id);
> -	if (!rc)
> +	if (!rc) {
> +		if (pci_tph_disabled())
> +			pcie_tph_disable(pci_dev);

I'm not really a fan of cluttering probe() like this.  Can't we
disable it in pcie_tph_init() so all devices start off with TPH
disabled, and then check pci_tph_disabled() in whatever interface
drivers use to enable TPH?

> +bool pci_tph_disabled(void)
> +{
> +	return pcie_tph_disabled;
> +}
> +EXPORT_SYMBOL_GPL(pci_tph_disabled);

Other related interfaces use "pcie" prefix; I think this should match.

Do drivers need this?  Would be nice not to export it unless they do.

Bjorn