Content-Type: text/plain; charset="utf-8"
MIME-Version: 1.0
Content-Transfer-Encoding: 8BIT
From: Chris Wilson <chris@chris-wilson.co.uk>
User-Agent: alot/0.3.6
To: =?utf-8?q?Christian_K=C3=B6nig________?= 
        <ckoenig.leichtzumerken@gmail.com>, imirkin@alum.mit.edu,
        efault@gmx.de, michel@daenzer.net, konrad.wilk@oracle.com,
        linux-kernel@vger.kernel.org
References: <20180102121358.16215-1-christian.koenig@amd.com>
In-Reply-To: <20180102121358.16215-1-christian.koenig@amd.com>
Message-ID: <151490759196.21495.9033629958966748437@mail.alporthouse.com>
Subject: Re: [PATCH] swiotlb: suppress warning when __GFP_NOWARN is set v2
Date: Tue, 02 Jan 2018 15:39:51 +0000
Sender: linux-kernel-owner@vger.kernel.org

Quoting Christian König (2018-01-02 12:13:58)
> TTM tries to allocate coherent memory in chunks of 2MB first to improve
> TLB efficiency and falls back to allocating 4K pages if that fails.
> 
> Suppress the warning when the 2MB allocations fails since there is a
> valid fall back path.
> 
> v2: suppress warnings from swiotlb_tbl_map_single as well
> 
> Signed-off-by: Christian König <christian.koenig@amd.com>
> Reported-by: Mike Galbraith <efault@gmx.de>
> Bug: https://bugs.freedesktop.org/show_bug.cgi?id=104082
> CC: stable@vger.kernel.org
> ---
> @@ -989,7 +1002,8 @@ swiotlb_map_sg_attrs(struct device *hwdev, struct scatterlist *sgl, int nelems,
>                 if (swiotlb_force == SWIOTLB_FORCE ||
>                     !dma_capable(hwdev, dev_addr, sg->length)) {
>                         phys_addr_t map = map_single(hwdev, sg_phys(sg),
> -                                                    sg->length, dir, attrs);
> +                                                    sg->length, dir, attrs,
> +                                                    true);

This keeps the warning for dma_map_sg(), so seems quite ineffective.

>                         if (map == SWIOTLB_MAP_ERROR) {
>                                 /* Don't panic here, we expect map_sg users
>                                    to do proper error handling. */

And counter to the suggestion here?
-Chris