Re: [Linux-stable-mirror] [PATCH v3] mmc: sdhci: Implement an SDHCI-specific bounce buffer

[...]

@@ -3713,6 +3751,43 @@ int sdhci_setup_host(struct sdhci_host *host) */ mmc->max_blk_count = (host->quirks & SDHCI_QUIRK_NO_MULTIBLOCK) ? 1 : 65535;

•   if (mmc->max_segs == 1) {
  

•           unsigned int max_blocks;
  

•           unsigned int max_seg_size;
  

•           max_seg_size = mmc->max_req_size;
  

•           max_blocks = max_seg_size / 512;
  

•           dev_info(mmc->parent, "host only supports SDMA, activate bounce buffer\n");
  

•           /*
  

•            * When we just support one segment, we can get significant speedups
  

•            * by the help of a bounce buffer to group scattered reads/writes
  

•            * together.
  

•            *
  

•            * TODO: is this too big? Stealing too much memory? The old bounce
  

•            * buffer is max 64K. This should be the 512K that SDMA can handle
  

•            * if I read the code above right. Anyways let's try this.
  

•            * FIXME: use devm_*
  

•            */
  

•           host->bounce_buffer = dma_alloc_coherent(mmc->parent, max_seg_size,
  

•                                                    &host->bounce_addr, GFP_KERNEL);
  

Why do you need dma_alloc_coherent() - allocated bounce buffer? Isn't a regular kmalloc-ed buffer fine as a bounce buffer?

I am worried, also according to your above comment, that it's not always going to work to request a large buffer like this, especially with dma_alloc_coherent().

•           if (!host->bounce_buffer) {
  

•                   dev_err(mmc->parent,
  

•                           "failed to allocate %u bytes for bounce buffer\n",
  

•                           max_seg_size);
  

•                   return -ENOMEM;
  

Why not fall back to use the max_segs = 1 here. It may be slow, as reported, but it works.

•           }
  

•           host->bounce_buffer_size = max_seg_size;
  

•           /* Lie about this since we're bouncing */
  

•           mmc->max_segs = max_blocks;
  

•           mmc->max_seg_size = max_seg_size;
  

•           dev_info(mmc->parent,
  

•                    "bounce buffer: bounce up to %u segments into one, max segment size %u bytes\n",
  

•                    max_blocks, max_seg_size);
  

•   }
  

•   return 0;
  
  

[...]

Kind regards Uffe