Re: [PATCH v11 1/8] dax: Introduce holder for dax_device

From: Shiyang Ruan <ruansy.fnst@fujitsu.com>
To: Christoph Hellwig <hch@infradead.org>
Cc: Dan Williams <dan.j.williams@intel.com>,
	Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
	linux-xfs <linux-xfs@vger.kernel.org>,
	Linux NVDIMM <nvdimm@lists.linux.dev>,
	Linux MM <linux-mm@kvack.org>,
	linux-fsdevel <linux-fsdevel@vger.kernel.org>,
	"Darrick J. Wong" <djwong@kernel.org>,
	david <david@fromorbit.com>, Jane Chu <jane.chu@oracle.com>
Subject: Re: [PATCH v11 1/8] dax: Introduce holder for dax_device
Date: Wed, 30 Mar 2022 18:58:21 +0800	[thread overview]
Message-ID: <4ed8baf7-7eb9-71e5-58ea-7c73b7e5bb73@fujitsu.com> (raw)
In-Reply-To: <YkQtOO/Z3SZ2Pksg@infradead.org>

在 2022/3/30 18:13, Christoph Hellwig 写道:
> On Wed, Mar 30, 2022 at 06:03:01PM +0800, Shiyang Ruan wrote:
>>
>> Because I am not sure if the offset between each layer is page aligned.  For
>> example, when pmem dirver handles ->memory_failure(), it should subtract its
>> ->data_offset when it calls dax_holder_notify_failure().
> 
> If they aren't, none of the DAX machinery would work.

OK. Got it.

So, use page-based function signature for ->memory_failure():

int (*memory_failure)(struct dev_pagemap *pgmap, unsigned long pfn,
		      unsigned long nr_pfns, int flags);

As the code I pasted before, pmem driver will subtract its 
->data_offset, which is byte-based. And the filesystem who implements 
->notify_failure() will calculate the offset in unit of byte again.

So, leave its function signature byte-based, to avoid repeated conversions.

int (*notify_failure)(struct dax_device *dax_dev, u64 offset,
		      u64 len, int mf_flags);

What do you think?

--
Thanks,
Ruan.