linux-kernel - dma_sync_*_for_cpu and direction=TO_DEVICE (was Re: [PATCH 02/20] dma-mapping: provide a generic dma-noncoherent implementation)

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]

Date:   Fri, 18 May 2018 10:20:02 -0700
From:   Vineet Gupta <Vineet.Gupta1@...opsys.com>
To:     Alexey Brodkin <Alexey.Brodkin@...opsys.com>,
        "hch@....de" <hch@....de>
CC:     "linux-arch@...r.kernel.org" <linux-arch@...r.kernel.org>,
        "linux-xtensa@...ux-xtensa.org" <linux-xtensa@...ux-xtensa.org>,
        "monstr@...str.eu" <monstr@...str.eu>,
        "linux-snps-arc@...ts.infradead.org" 
        <linux-snps-arc@...ts.infradead.org>,
        "linux-c6x-dev@...ux-c6x.org" <linux-c6x-dev@...ux-c6x.org>,
        "linux-parisc@...r.kernel.org" <linux-parisc@...r.kernel.org>,
        "linux-sh@...r.kernel.org" <linux-sh@...r.kernel.org>,
        "linux-hexagon@...r.kernel.org" <linux-hexagon@...r.kernel.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "iommu@...ts.linux-foundation.org" <iommu@...ts.linux-foundation.org>,
        "linux-m68k@...ts.linux-m68k.org" <linux-m68k@...ts.linux-m68k.org>,
        "openrisc@...ts.librecores.org" <openrisc@...ts.librecores.org>,
        "green.hu@...il.com" <green.hu@...il.com>,
        "linux-alpha@...r.kernel.org" <linux-alpha@...r.kernel.org>,
        "sparclinux@...r.kernel.org" <sparclinux@...r.kernel.org>,
        "nios2-dev@...ts.rocketboards.org" <nios2-dev@...ts.rocketboards.org>,
        "deanbo422@...il.com" <deanbo422@...il.com>,
        "linux-arm-kernel@...ts.infradead.org" 
        <linux-arm-kernel@...ts.infradead.org>,
        "linux-mm@...ck.org" <linux-mm@...ck.org>,
        Andrew Morton <akpm@...ux-foundation.org>
Subject: dma_sync_*_for_cpu and direction=TO_DEVICE (was Re: [PATCH 02/20]
 dma-mapping: provide a generic dma-noncoherent implementation)

On 05/18/2018 06:11 AM, Alexey Brodkin wrote:
>   void arch_sync_dma_for_device(struct device *dev, phys_addr_t paddr,
>                  size_t size, enum dma_data_direction dir)
>   {
> +       if (dir != DMA_TO_DEVICE){
> +               dump_stack();
> +               printk(" *** %s@%d: DMA direction is %s instead of %s\n",
> +                      __func__, __LINE__, dir_to_str(dir), dir_to_str(DMA_TO_DEVICE));
> +       }
> +
>          return _dma_cache_sync(dev, paddr, size, dir);
>   }
>   
>   void arch_sync_dma_for_cpu(struct device *dev, phys_addr_t paddr,
>                  size_t size, enum dma_data_direction dir)
>   {
> +       if (dir != DMA_FROM_DEVICE) {
> +               dump_stack();
> +               printk(" *** %s@%d: DMA direction is %s instead of %s\n",
> +                      __func__, __LINE__, dir_to_str(dir), dir_to_str(DMA_FROM_DEVICE));
> +       }
> +
>          return _dma_cache_sync(dev, paddr, size, dir);
>   }

...
> In case of MMC/DW_MCI (AKA DesignWare MobileStorage controller) that's an execution flow:
> 1) __dw_mci_start_request()
> 2) dw_mci_pre_dma_transfer()
> 3) dma_map_sg(..., mmc_get_dma_dir(data))
>
> Note mmc_get_dma_dir() is just "data->flags & MMC_DATA_WRITE ? DMA_TO_DEVICE : DMA_FROM_DEVICE".
> I.e. if we're preparing for sending data dma_noncoherent_map_sg() will have DMA_TO_DEVICE which
> is quite OK for passing to dma_noncoherent_sync_sg_for_device() but in case of reading we'll have
> DMA_FROM_DEVICE which we'll pass to dma_noncoherent_sync_sg_for_device() in dma_noncoherent_map_sg().
>
> I'd say this is not entirely correct because IMHO arch_sync_dma_for_cpu() is supposed to only be used
> in case of DMA_FROM_DEVICE and arch_sync_dma_for_device() only in case of DMA_TO_DEVICE.

So roughly 10 years ago, some kernel rookie name Vineet Gupta, asked the exact 
same question :-)

http://kernelnewbies.kernelnewbies.narkive.com/aGW1QcDv/query-about-dma-sync-for-cpu-and-direction-to-device

I never understood the need for this direction. And if memory serves me right, at 
that time I was seeing twice the amount of cache flushing !

> But the real fix of my problem is:
> ---------------------------------------->8------------------------------------
> --- a/lib/dma-noncoherent.c
> +++ b/lib/dma-noncoherent.c
> @@ -35,7 +35,7 @@ static dma_addr_t dma_noncoherent_map_page(struct device *dev, struct page *page
>   
>          addr = dma_direct_map_page(dev, page, offset, size, dir, attrs);
>          if (!dma_mapping_error(dev, addr) && !(attrs & DMA_ATTR_SKIP_CPU_SYNC))
> -               arch_sync_dma_for_device(dev, page_to_phys(page), size, dir);
> +               arch_sync_dma_for_device(dev, page_to_phys(page) + offset, size, dir);
>          return addr;
>   }
> ---------------------------------------->8------------------------------------
>
> You seem to lost an offset in the page so if we happen to have a buffer not aligned to
> a page boundary then we were obviously corrupting data outside our data :)

Neat !