[Intel-gfx] [PATCH 6/8] iommu: allow the dma-iommu api to use bounce buffers
kbuild test robot
lkp at intel.com
Tue Dec 24 10:20:41 UTC 2019
Hi Tom,
Thank you for the patch! Perhaps something to improve:
[auto build test WARNING on rockchip/for-next]
[cannot apply to iommu/next tegra/for-next vfio/next linus/master v5.5-rc3 next-20191219]
[if your patch is applied to the wrong git tree, please drop us a note to help
improve the system. BTW, we also suggest to use '--base' option to specify the
base tree in git format-patch, please see https://stackoverflow.com/a/37406982]
url: https://github.com/0day-ci/linux/commits/Tom-Murphy/Convert-the-intel-iommu-driver-to-the-dma-iommu-api/20191224-171249
base: https://git.kernel.org/pub/scm/linux/kernel/git/mmind/linux-rockchip.git for-next
config: x86_64-defconfig (attached as .config)
compiler: gcc-7 (Debian 7.5.0-3) 7.5.0
reproduce:
# save the attached .config to linux build tree
make ARCH=x86_64
If you fix the issue, kindly add following tag
Reported-by: kbuild test robot <lkp at intel.com>
All warnings (new ones prefixed by >>):
drivers//iommu/dma-iommu.c: In function '__iommu_dma_map':
>> drivers//iommu/dma-iommu.c:568:3: warning: ISO C90 forbids mixed declarations and code [-Wdeclaration-after-statement]
void *padding_start = phys_to_virt(phys);
^~~~
vim +568 drivers//iommu/dma-iommu.c
537
538 static dma_addr_t __iommu_dma_map(struct device *dev, phys_addr_t phys,
539 size_t org_size, dma_addr_t dma_mask, bool coherent,
540 enum dma_data_direction dir, unsigned long attrs)
541 {
542 int prot = dma_info_to_prot(dir, coherent, attrs);
543 struct iommu_domain *domain = iommu_get_dma_domain(dev);
544 struct iommu_dma_cookie *cookie = domain->iova_cookie;
545 struct iova_domain *iovad = &cookie->iovad;
546 size_t iova_off = iova_offset(iovad, phys);
547 size_t aligned_size = iova_align(iovad, org_size + iova_off);
548 dma_addr_t iova;
549
550 if (unlikely(iommu_dma_deferred_attach(dev, domain)))
551 return DMA_MAPPING_ERROR;
552
553 #ifdef CONFIG_SWIOTLB
554 /*
555 * If both the physical buffer start address and size are
556 * page aligned, we don't need to use a bounce page.
557 */
558 if (iommu_needs_bounce_buffer(dev)
559 && !iova_offset(iovad, phys | org_size)) {
560 phys = swiotlb_tbl_map_single(dev,
561 __phys_to_dma(dev, io_tlb_start),
562 phys, org_size, aligned_size, dir, attrs);
563
564 if (phys == DMA_MAPPING_ERROR)
565 return DMA_MAPPING_ERROR;
566
567 /* Cleanup the padding area. */
> 568 void *padding_start = phys_to_virt(phys);
569 size_t padding_size = aligned_size;
570
571 if (!(attrs & DMA_ATTR_SKIP_CPU_SYNC) &&
572 (dir == DMA_TO_DEVICE ||
573 dir == DMA_BIDIRECTIONAL)) {
574 padding_start += org_size;
575 padding_size -= org_size;
576 }
577
578 memset(padding_start, 0, padding_size);
579 }
580 #endif
581
582 iova = iommu_dma_alloc_iova(domain, aligned_size, dma_mask, dev);
583 if (!iova)
584 return DMA_MAPPING_ERROR;
585
586 if (iommu_map_atomic(domain, iova, phys - iova_off, aligned_size,
587 prot)) {
588
589 if (unlikely(is_swiotlb_buffer(phys)))
590 swiotlb_tbl_unmap_single(dev, phys, aligned_size,
591 aligned_size, dir, attrs);
592 iommu_dma_free_iova(cookie, iova, aligned_size, NULL);
593 return DMA_MAPPING_ERROR;
594 }
595 return iova + iova_off;
596 }
597
---
0-DAY kernel test infrastructure Open Source Technology Center
https://lists.01.org/hyperkitty/list/kbuild-all@lists.01.org Intel Corporation
-------------- next part --------------
A non-text attachment was scrubbed...
Name: .config.gz
Type: application/gzip
Size: 28860 bytes
Desc: not available
URL: <https://lists.freedesktop.org/archives/intel-gfx/attachments/20191224/91f89e3b/attachment-0001.gz>
More information about the Intel-gfx
mailing list