Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.38 as permitted sender) client-ip=23.128.96.38;
Date:   Mon, 6 Nov 2023 12:53:31 +0800
From:   Ming Lei <ming.lei@redhat.com>
To:     Ed Tsai =?utf-8?B?KOiUoeWul+i7kik=?= <Ed.Tsai@mediatek.com>
Cc:     Will Shiu =?utf-8?B?KOioseaBreeRnCk=?= <Will.Shiu@mediatek.com>,
        "linux-mediatek@lists.infradead.org" 
        <linux-mediatek@lists.infradead.org>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
        Peter Wang =?utf-8?B?KOeOi+S/oeWPiyk=?= 
        <peter.wang@mediatek.com>,
        "linux-block@vger.kernel.org" <linux-block@vger.kernel.org>,
        Alice Chao =?utf-8?B?KOi2meePruWdhyk=?= 
        <Alice.Chao@mediatek.com>,
        wsd_upstream <wsd_upstream@mediatek.com>,
        "axboe@kernel.dk" <axboe@kernel.dk>,
        Casper Li =?utf-8?B?KOadjuS4reamrik=?= <casper.li@mediatek.com>,
        Chun-Hung Wu =?utf-8?B?KOW3q+mnv+Wujyk=?= 
        <Chun-hung.Wu@mediatek.com>,
        Powen Kao =?utf-8?B?KOmrmOS8r+aWhyk=?= <Powen.Kao@mediatek.com>,
        Naomi Chu =?utf-8?B?KOacseipoOeUsCk=?= <Naomi.Chu@mediatek.com>,
        "linux-arm-kernel@lists.infradead.org" 
        <linux-arm-kernel@lists.infradead.org>,
        Stanley Chu =?utf-8?B?KOacseWOn+mZnik=?= 
        <stanley.chu@mediatek.com>,
        "matthias.bgg@gmail.com" <matthias.bgg@gmail.com>,
        "angelogioacchino.delregno@collabora.com" 
        <angelogioacchino.delregno@collabora.com>, ming.lei@redhat.com
Subject: Re: [PATCH 1/1] block: Check the queue limit before bio submitting
Message-ID: <ZUhxS9JMyPK+v6Ec@fedora>
References: <20231025092255.27930-1-ed.tsai@mediatek.com>
 <64db8f5406571c2f89b70f852eb411320201abe6.camel@mediatek.com>
 <ZUUd48QF/TEGFzPy@fedora>
 <cf09a768d0e116bfaf01a1592a7ae95a10b4c2cf.camel@mediatek.com>
 <ZUW96Ha5GoJePD8Y@fedora>
 <2bc847a83849973b7658145f2efdda86cc47e3d5.camel@mediatek.com>
 <5ecedad658bf28abf9bbeeb70dcac09b4b404cf5.camel@mediatek.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: <5ecedad658bf28abf9bbeeb70dcac09b4b404cf5.camel@mediatek.com>
Precedence: bulk

On Mon, Nov 06, 2023 at 01:40:12AM +0000, Ed Tsai (蔡宗軒) wrote:
> On Mon, 2023-11-06 at 09:33 +0800, Ed Tsai wrote:
> > On Sat, 2023-11-04 at 11:43 +0800, Ming Lei wrote:

...

> Sorry for missing out on my dd command. Here it is:
> dd if=/data/test_file of=/dev/null bs=64m count=1 iflag=direct

OK, thanks for the sharing.

I understand the issue now, but not sure if it is one good idea to check
queue limit in __bio_iov_iter_get_pages():

1) bio->bi_bdev may not be set

2) what matters is actually bio's alignment, and bio size still can
be big enough

So I cooked one patch, and it should address your issue:


diff --git a/block/bio.c b/block/bio.c
index 816d412c06e9..7d982e74c65d 100644
--- a/block/bio.c
+++ b/block/bio.c
@@ -1266,6 +1266,24 @@ static int __bio_iov_iter_get_pages(struct bio *bio, struct iov_iter *iter)
 		size -= trim;
 	}
 
+	/*
+	 * Try to make bio aligned with 128KB if there are pages left so we
+	 * can avoid small bio in case of big chunk sequential IO
+	 */
+	if (iov_iter_count(iter)) {
+		unsigned curr_size = (bio->bi_iter.bi_size + size) &
+			~((128U << 10) - 1);
+		if (curr_size <= bio->bi_iter.bi_size) {
+			ret = left = size;
+			goto revert;
+		} else {
+			curr_size -= bio->bi_iter.bi_size;
+			ret = size - curr_size;
+			iov_iter_revert(iter, ret);
+			size = curr_size;
+		}
+	}
+
 	if (unlikely(!size)) {
 		ret = -EFAULT;
 		goto out;
@@ -1285,7 +1303,7 @@ static int __bio_iov_iter_get_pages(struct bio *bio, struct iov_iter *iter)
 
 		offset = 0;
 	}
-
+revert:
 	iov_iter_revert(iter, left);
 out:
 	while (i < nr_pages)


Thanks,
Ming