On Mon, Dec 07, 2020 at 01:42:04PM -0800, Jianxiong Gao wrote:
> NVMe driver and other applications depend on the data offset
> to operate correctly. Currently when unaligned data is mapped via
> SWIOTLB, the data is mapped as slab aligned with the SWIOTLB. When
> booting with --swiotlb=force option and using NVMe as interface,
> running mkfs.xfs on Rhel fails because of the unalignment issue.
> This patch makes sure the mapped data preserves
> its offset of the orginal address. Tested on latest kernel that
> this patch fixes the issue.
>
> Signed-off-by: Jianxiong Gao <jxgao@google.com>
> Acked-by: David Rientjes <rientjes@google.com>

This breaks DHCP with upstream kernel (applied this on top v5.10-rc7)
and used swiotlb=262144,force and now the dhclient is not working:

[  119.300502] bnxt_en 0000:3b:00.0 eno2np0: NIC Link is Up, 25000 Mbps full duplex, Flow control: ON - receive & transmit
[  119.437573] bnxt_en 0000:3b:00.0 eno2np0: FEC autoneg off encoding: None
[   90.064220] dracut-initqueue[1477]: Warning: dhcp for interface eno2np0 failed
[  101.155295] dracut-initqueue[1477]: Warning: dhcp for interfa[  142.361359] bnxt_en 0000:3b:00.1 eno3np1: NIC Link is Up, 25000 Mbps full duplex, Flow control: ON - receive & transmit
ce eno2np0 faile[  142.501860] bnxt_en 0000:3b:00.1 eno3np1: FEC autoneg off encoding: None
d
[  113.054108] dracut-initqueue[1477]: Warning: dhcp for interface eno3np1 failed
[  123.867108] dracut-initqueue[1477]: Warning: dhcp for interface eno3np1 failed
[  251.888002] dracut-initqueue[1477]: Warning: dracut-initqueue timeout - starting timeout scripts

Dropping from linux-next.

> ---
>  kernel/dma/swiotlb.c | 11 +++++++++++
>  1 file changed, 11 insertions(+)
>
> diff --git a/kernel/dma/swiotlb.c b/kernel/dma/swiotlb.c
> index 781b9dca197c..56a35e71b3fd 100644
> --- a/kernel/dma/swiotlb.c
> +++ b/kernel/dma/swiotlb.c
> @@ -483,6 +483,12 @@ phys_addr_t swiotlb_tbl_map_single(struct device *hwdev, phys_addr_t orig_addr,
>       max_slots = mask + 1
>                   ? ALIGN(mask + 1, 1 << IO_TLB_SHIFT) >> IO_TLB_SHIFT
>                   : 1UL << (BITS_PER_LONG - IO_TLB_SHIFT);
> +
> +     /*
> +      * We need to keep the offset when mapping, so adding the offset
> +      * to the total set we need to allocate in SWIOTLB
> +      */
> +     alloc_size += offset_in_page(orig_addr);
> 
>       /*
>        * For mappings greater than or equal to a page, we limit the stride
> @@ -567,6 +573,11 @@ phys_addr_t swiotlb_tbl_map_single(struct device *hwdev, phys_addr_t orig_addr,
>        */
>       for (i = 0; i < nslots; i++)
>               io_tlb_orig_addr[index+i] = orig_addr + (i << IO_TLB_SHIFT);
> +     /*
> +      * When keeping the offset of the original data, we need to advance
> +      * the tlb_addr by the offset of orig_addr.
> +      */
> +     tlb_addr += orig_addr & (PAGE_SIZE - 1);
>       if (!(attrs & DMA_ATTR_SKIP_CPU_SYNC) &&
>           (dir == DMA_TO_DEVICE || dir == DMA_BIDIRECTIONAL))
>               swiotlb_bounce(orig_addr, tlb_addr, mapping_size, DMA_TO_DEVICE);
> --
> 2.27.0
>
>