Received-SPF: pass (google.com: domain of linux-wireless-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: quoted-printable
From:   Felix Fietkau <nbd@nbd.name>
Mime-Version: 1.0 (1.0)
Subject: Re: [PATCH v2] wifi: mt76: fix potential memory leakage
Date:   Wed, 21 Dec 2022 16:12:40 +0100
Message-Id: <B2689D6F-E04E-4E9B-9244-4E5069DC464D@nbd.name>
References: <3940cc5ccb4b17d51b5176e722daef0d1af66d6a.camel@mediatek.com>
Cc:     lorenzo@kernel.org, linux-wireless@vger.kernel.org,
        Shayne Chen <shayne.chen@mediatek.com>,
        Bo Jiao <bo.jiao@mediatek.com>,
        Evelyn Tsai <evelyn.tsai@mediatek.com>,
        Ryder Lee <ryder.lee@mediatek.com>,
        linux-mediatek@lists.infradead.org
In-Reply-To: <3940cc5ccb4b17d51b5176e722daef0d1af66d6a.camel@mediatek.com>
To:     Sujuan Chen <sujuan.chen@mediatek.com>
Precedence: bulk

Hi Sujuan,

> Yes, it is so expensive, but if no memcopy, it will casue memory
> fragmentation (we hit this issue in internal SQC).
>=20
> as we know, wed needs to exchange rx pkt(belongs to wed rx buffer
> manager) with wifi driver(dma rx data queue) by exchanging wfdma dmad
> to ensure the free wed rx buffer.
>=20
> it is possiable that a large number of buffer has been exchanged to wed
> and can not come back to wlan driver. So, the memory from the same 32K
> page cache is unable to be released, and it will be failed at
> page_frag_alloc in mt76_dma_rx_fill.
>=20
> Any ideas but memcopy?
A simple solution would be to simply allocate single pages, or half-page fra=
gments.

- Felix