Re: [linux-lvm] resend patch - bcache may mistakenly write data to another disk when writes error

From: David Teigland <teigland@redhat.com>
To: Heming Zhao <heming.zhao@suse.com>
Cc: "thornber@redhat.com" <thornber@redhat.com>,
	"linux-lvm@redhat.com" <linux-lvm@redhat.com>
Subject: Re: [linux-lvm] resend patch - bcache may mistakenly write data to another disk when writes error
Date: Tue, 12 Nov 2019 09:21:12 -0600	[thread overview]
Message-ID: <20191112152112.GA4422@redhat.com> (raw)
In-Reply-To: <872328cd-3d51-97bb-1c50-b54cc194c6f2@suse.com>

On Tue, Nov 12, 2019 at 08:38:00AM +0000, Heming Zhao wrote:
> There still have bug codes wait for fixing.
> Please see below item <2>. the _last_byte_xx (in bcache.c) should be restored to ZERO when fail/error.
> Do you have any plan/schedule to this issue?
> Do you have any plan/schedule to backport these fixes to stable-2.02 branch?

I don't think it should actually be a problem since bcache_unset_last_byte
checks for _last_byte_fd == fd.  But, I think it's better to not rely on that
check, so how about a change like this?

void dev_unset_last_byte(struct device *dev)
{
        if (dev->bcache_fd > 0)
                bcache_unset_last_byte(scan_bcache, dev->bcache_fd);
}