Re: [PATCH 09/15] swiotlb: make the swiotlb_init interface more useful

cleanup swiotlb initialization v8 · Christoph Hellwig <hch@lst.de> · 2022-04-04
[PATCH 01/15] dma-direct: use is_swiotlb_active in dma_direct_map_page · Christoph Hellwig <hch@lst.de> · 2022-04-04
[PATCH 02/15] swiotlb: make swiotlb_exit a no-op if SWIOTLB_FORCE is set · Christoph Hellwig <hch@lst.de> · 2022-04-04
[PATCH 03/15] swiotlb: simplify swiotlb_max_segment · Christoph Hellwig <hch@lst.de> · 2022-04-04
[PATCH 04/15] swiotlb: rename swiotlb_late_init_with_default_size · Christoph Hellwig <hch@lst.de> · 2022-04-04
[PATCH 05/15] arm/xen: don't check for xen_initial_domain() in xen_create_contiguous_region · Christoph Hellwig <hch@lst.de> · 2022-04-04
[PATCH 06/15] MIPS/octeon: use swiotlb_init instead of open coding it · Christoph Hellwig <hch@lst.de> · 2022-04-04
[PATCH 07/15] x86: remove the IOMMU table infrastructure · Christoph Hellwig <hch@lst.de> · 2022-04-04
Re: [PATCH 07/15] x86: remove the IOMMU table infrastructure · Konrad Rzeszutek Wilk <hidden> · 2022-04-07
[PATCH 08/15] x86: centralize setting SWIOTLB_FORCE when guest memory encryption is enabled · Christoph Hellwig <hch@lst.de> · 2022-04-04
[PATCH 09/15] swiotlb: make the swiotlb_init interface more useful · Christoph Hellwig <hch@lst.de> · 2022-04-04
RE: [PATCH 09/15] swiotlb: make the swiotlb_init interface more useful · Michael Kelley (LINUX) <hidden> · 2022-04-04
Re: [PATCH 09/15] swiotlb: make the swiotlb_init interface more useful · Nathan Chancellor <nathan@kernel.org> · 2022-06-01
Re: [PATCH 09/15] swiotlb: make the swiotlb_init interface more useful · Christoph Hellwig <hch@lst.de> · 2022-06-01
Re: [PATCH 09/15] swiotlb: make the swiotlb_init interface more useful · Nathan Chancellor <nathan@kernel.org> · 2022-06-01
Re: [PATCH 09/15] swiotlb: make the swiotlb_init interface more useful · Christoph Hellwig <hch@lst.de> · 2022-06-01
Re: [PATCH 09/15] swiotlb: make the swiotlb_init interface more useful · Nathan Chancellor <nathan@kernel.org> · 2022-06-01
Re: [PATCH 09/15] swiotlb: make the swiotlb_init interface more useful · Christoph Hellwig <hch@lst.de> · 2022-06-01
Re: [PATCH 09/15] swiotlb: make the swiotlb_init interface more useful · Nathan Chancellor <nathan@kernel.org> · 2022-06-01
[PATCH 10/15] swiotlb: add a SWIOTLB_ANY flag to lift the low memory restriction · Christoph Hellwig <hch@lst.de> · 2022-04-04
Re: [PATCH 10/15] swiotlb: add a SWIOTLB_ANY flag to lift the low memory restriction · Konrad Rzeszutek Wilk <hidden> · 2022-04-07
Re: [PATCH 10/15] swiotlb: add a SWIOTLB_ANY flag to lift the low memory restriction · Christoph Hellwig <hch@lst.de> · 2022-04-12
[PATCH 11/15] swiotlb: pass a gfp_mask argument to swiotlb_init_late · Christoph Hellwig <hch@lst.de> · 2022-04-04
[PATCH 12/15] swiotlb: provide swiotlb_init variants that remap the buffer · Christoph Hellwig <hch@lst.de> · 2022-04-04
Re: [PATCH 12/15] swiotlb: provide swiotlb_init variants that remap the buffer · Dongli Zhang <hidden> · 2022-04-04
[PATCH 13/15] swiotlb: merge swiotlb-xen initialization into swiotlb · Christoph Hellwig <hch@lst.de> · 2022-04-04
[PATCH 14/15] swiotlb: remove swiotlb_init_with_tbl and swiotlb_init_late_with_tbl · Christoph Hellwig <hch@lst.de> · 2022-04-04
[PATCH 15/15] x86: remove cruft from <asm/dma-mapping.h> · Christoph Hellwig <hch@lst.de> · 2022-04-04
Re: cleanup swiotlb initialization v8 · Boris Ostrovsky <boris.ostrovsky@oracle.com> · 2022-04-05
Re: cleanup swiotlb initialization v8 · Konrad Rzeszutek Wilk <hidden> · 2022-04-07
Re: cleanup swiotlb initialization v8 · Christoph Hellwig <hch@lst.de> · 2022-04-13

From: Christoph Hellwig <hch@lst.de>
Date: 2022-06-01 17:58:58
Also in: linux-arm-kernel, linux-hyperv, linux-iommu, linux-mips, linux-pci, linux-riscv, linux-s390, xen-devel

On Wed, Jun 01, 2022 at 10:46:54AM -0700, Nathan Chancellor wrote:

On Wed, Jun 01, 2022 at 07:34:41PM +0200, Christoph Hellwig wrote:

quoted

Can you send me the full dmesg and the content of
/sys/kernel/debug/swiotlb/io_tlb_nslabs for a good and a bad boot?

Sure thing, they are attached! If there is anything else I can provide
or test, I am more than happy to do so.

Nothing interesting.  But the performance numbers almost look like
swiotlb=force got ignored before (even if I can't explain why).
Do you get a similar performance with the new kernel without
swiotlb=force as the old one with that argument by any chance?

`h`	back out one level
`j`	next message in thread
`k`	previous message in thread
`l`	drill in
`Esc`	close help / fold thread tree
`?`	toggle this help