Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::3:4 as permitted sender) client-ip=2620:137:e000::3:4;
MIME-Version: 1.0
References: <20230828075420.2009568-1-anle.pan@nxp.com> <DB9PR04MB9284AA58AA71655C9C0C7C9A87E6A@DB9PR04MB9284.eurprd04.prod.outlook.com>
 <CAAFQd5DYxwX+JdXQ_-ba5B9y+gyWPjvUpCBnKM1zz1W2EkO_vA@mail.gmail.com>
 <DB9PR04MB928460023FB3CCCBC3EACE1487E9A@DB9PR04MB9284.eurprd04.prod.outlook.com>
 <CAAFQd5BGJX7=Z1ukFRq_ktaQ0d7FbvV-ob5gs8hfGaNHUXPTww@mail.gmail.com>
 <DB9PR04MB9284A0CDB1FC7CAADE0A394F87EFA@DB9PR04MB9284.eurprd04.prod.outlook.com>
 <CAAFQd5AugUGh7Z=9Qh7SS4=-0ddGBmRAOOyCGDfPdT-=eurtUg@mail.gmail.com> <DB9PR04MB9284A45033B3E24F44C5AA3987F2A@DB9PR04MB9284.eurprd04.prod.outlook.com>
In-Reply-To: <DB9PR04MB9284A45033B3E24F44C5AA3987F2A@DB9PR04MB9284.eurprd04.prod.outlook.com>
From:   Tomasz Figa <tfiga@chromium.org>
Date:   Tue, 12 Sep 2023 11:22:12 +0900
Message-ID: <CAAFQd5Bk0DUq5jgFDHV2wGYEcfQzNVQ2KYJ2ZYLubwFLgqW4kQ@mail.gmail.com>
Subject: Re: [EXT] Re: [PATCH] media: videobuf2-dma-sg: limit the sg segment size
To:     Hui Fang <hui.fang@nxp.com>
Cc:     Anle Pan <anle.pan@nxp.com>,
        "m.szyprowski@samsung.com" <m.szyprowski@samsung.com>,
        "mchehab@kernel.org" <mchehab@kernel.org>,
        "linux-media@vger.kernel.org" <linux-media@vger.kernel.org>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
        Jindong Yue <jindong.yue@nxp.com>,
        Xuegang Liu <xuegang.liu@nxp.com>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Mon, Sep 11, 2023 at 3:13=E2=80=AFPM Hui Fang <hui.fang@nxp.com> wrote:
>
> On Wed, Sep 6, 2023 at 18:28=E2=80=AFPM Tomasz Figa <tfiga@chromium.org> =
wrote:
> > That all makes sense, but it still doesn't answer the real question on =
why
> > swiotlb ends up being used. I think you may want to trace what happens =
in
> > the DMA mapping ops implementation on your system causing it to use
> > swiotlb.
>
> Add log and feed invalid data to low buffer on purpose,
> it's confirmed that swiotlb is actually used.
>

Yes, that we already know. But why?

> Got log as
> "[  846.570271][  T138] software IO TLB: =3D=3D=3D=3D swiotlb_bounce: DMA=
_TO_DEVICE,
>  dst 000000004589fa38, src 00000000c6d7e8d8, srcPhy 5504139264, size 4096=
".
>
> " srcPhy 5504139264" is larger than 4G (8mp has DRAM over 5G).
> And "CONFIG_ZONE_DMA32=3Dy" in kernel config, so swiotlb static is used.
> Also, the host (win10) side can't get valid image.
>
> Code as below.
> diff --git a/drivers/media/common/videobuf2/videobuf2-dma-sg.c b/drivers/=
media/common/videobuf2/videobuf2-dma-sg.c
> index 7f83a86e6810..de03704ce695 100644
> --- a/drivers/media/common/videobuf2/videobuf2-dma-sg.c
> +++ b/drivers/media/common/videobuf2/videobuf2-dma-sg.c
> @@ -98,6 +98,7 @@ static int vb2_dma_sg_alloc_compacted(struct vb2_dma_sg=
_buf *buf,
>         return 0;
>  }
>
> +bool g_v4l2 =3D false;
>  static void *vb2_dma_sg_alloc(struct vb2_buffer *vb, struct device *dev,
>                               unsigned long size)
>  {
> @@ -144,6 +145,7 @@ static void *vb2_dma_sg_alloc(struct vb2_buffer *vb, =
struct device *dev,
>         if (ret)
>                 goto fail_table_alloc;
>
> +       g_v4l2 =3D true;
>         pr_info("=3D=3D=3D=3D vb2_dma_sg_alloc, call sg_alloc_table_from_=
pages_segment,
>                         size %d, max_segment %d\n", (int)size, (int)max_s=
egment);
> diff --git a/kernel/dma/swiotlb.c b/kernel/dma/swiotlb.c
> index dac01ace03a0..a2cda646a02f 100644
> --- a/kernel/dma/swiotlb.c
> +++ b/kernel/dma/swiotlb.c
> @@ -523,6 +523,7 @@ static unsigned int swiotlb_align_offset(struct devic=
e *dev, u64 addr)
>         return addr & dma_get_min_align_mask(dev) & (IO_TLB_SIZE - 1);
>  }
>
> +extern bool g_v4l2;
>  /*
>   * Bounce: copy the swiotlb buffer from or back to the original dma loca=
tion
>   */
> @@ -591,8 +592,19 @@ static void swiotlb_bounce(struct device *dev, phys_=
addr_t tlb_addr, size_t size
>                 }
>         } else if (dir =3D=3D DMA_TO_DEVICE) {
>                 memcpy(vaddr, phys_to_virt(orig_addr), size);
> +               if (g_v4l2) {
> +                       static unsigned char val;
> +                       val++;
> +                       memset(vaddr, val, size);
> +
> +                       pr_info("=3D=3D=3D=3Dxx %s: DMA_TO_DEVICE, dst %p=
, src %p, srcPhy %llu, size %zu\n",
> +                               __func__, vaddr, phys_to_virt(orig_addr),=
 orig_addr, size);
> +               }
>         } else {
>                 memcpy(phys_to_virt(orig_addr), vaddr, size);
>         }
>  }
>
>
> BRs,
> Fang Hui
>