linux-kernel - RE: [PATCH v2 1/1] swiotlb: Reduce swiotlb pool lookups

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:
 <SN6PR02MB4157144A1C28D63E00B4F9B2D4D92@SN6PR02MB4157.namprd02.prod.outlook.com>
Date: Sun, 7 Jul 2024 02:35:57 +0000
From: Michael Kelley <mhklinux@...look.com>
To: Christoph Hellwig <hch@....de>
CC: "robin.murphy@....com" <robin.murphy@....com>, "joro@...tes.org"
	<joro@...tes.org>, "will@...nel.org" <will@...nel.org>, "jgross@...e.com"
	<jgross@...e.com>, "sstabellini@...nel.org" <sstabellini@...nel.org>,
	"oleksandr_tyshchenko@...m.com" <oleksandr_tyshchenko@...m.com>,
	"m.szyprowski@...sung.com" <m.szyprowski@...sung.com>, "petr@...arici.cz"
	<petr@...arici.cz>, "iommu@...ts.linux.dev" <iommu@...ts.linux.dev>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
	"xen-devel@...ts.xenproject.org" <xen-devel@...ts.xenproject.org>
Subject: RE: [PATCH v2 1/1] swiotlb: Reduce swiotlb pool lookups

From: Michael Kelley Sent: Saturday, July 6, 2024 7:12 PM

[ ... ]
> >
> > If we then stub out swiotlb_find_pool to return NULL for !CONFIG_SWIOTLB,
> > we also don't need extra stubs for all the __swiotlb_ helpers as the
> > compiler will eliminate the calls as dead code.
> 
> Yes, this works as long as the declarations for the __swiotlb_foo
> functions are *not* under CONFIG_SWIOTLB. But when compiling with
> !CONFIG_SWIOTLB on arm64 with gcc-8.5.0, two tangentially related
> compile errors occur. iommu_dma_map_page() references
> swiotlb_tlb_map_single(). The declaration for the latter is under
> CONFIG_SWIOTLB. A similar problem occurs with dma_direct_map_page()
> and swiotlb_map(). Do later versions of gcc not complain when the
> reference is in dead code? Or are these just bugs that occurred because
> !CONFIG_SWIOTLB is rare? If the latter, I can submit a separate patch to
> move the declarations out from under CONFIG_SWIOTLB.
> 

Ignore the "two tangentially related compile errors".  I schlepped some
code around incorrectly and caused the problem myself. :-(

Michael