linux-kernel - Re: [PATCH] fs-writeback: Flush plug before next iteration in wb

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]

Message-ID: <20220415063920.GB24262@lst.de>
Date:   Fri, 15 Apr 2022 08:39:20 +0200
From:   Christoph Hellwig <hch@....de>
To:     Zhihao Cheng <chengzhihao1@...wei.com>
Cc:     viro@...iv.linux.org.uk, torvalds@...ux-foundation.org, hch@....de,
        linux-fsdevel@...r.kernel.org, linux-kernel@...r.kernel.org,
        yukuai3@...wei.com
Subject: Re: [PATCH] fs-writeback: Flush plug before next iteration in
 wb_writeback()

On Fri, Apr 15, 2022 at 09:37:35AM +0800, Zhihao Cheng wrote:
> +		if (progress) {
> +			/*
> +			 * The progress may be false postive in page redirty
> +			 * case (which is caused by failing to get buffer head
> +			 * lock), which will requeue dirty inodes and start
> +			 * next writeback iteration, and other tasks maybe
> +			 * stuck for getting tags for new requests. So, flush
> +			 * plug to schedule requests holding tags.
> +			 *
> +			 * The code can be removed after buffer head
> +			 * disappering from linux.
> +			 */
> +			blk_flush_plug(current->plug, false);

This basically removes plugging entirely, so we might as well stop
adding the plug if we can't solve it any other way.  But it seems
like that fake progress needs to be fixed instead.