Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Date:   Fri, 15 Mar 2019 10:21:18 -0700 (PDT)
From:   Hugh Dickins <hughd@google.com>
To:     Oscar Salvador <osalvador@suse.de>
cc:     Michal Hocko <mhocko@kernel.org>, akpm@linux-foundation.org,
        anshuman.khandual@arm.com, william.kucharski@oracle.com,
        linux-mm@kvack.org, linux-kernel@vger.kernel.org,
        Jan Kara <jack@suse.cz>, Hugh Dickins <hughd@google.com>
Subject: Re: [PATCH] mm: Fix __dump_page when mapping->host is not set
In-Reply-To: <20190315143304.pkuvj4qwtlzgm7iq@d104.suse.de>
Message-ID: <alpine.LSU.2.11.1903150952270.2934@eggly.anvils>
References: <20190315121826.23609-1-osalvador@suse.de> <20190315124733.GE15672@dhcp22.suse.cz> <20190315143304.pkuvj4qwtlzgm7iq@d104.suse.de>
User-Agent: Alpine 2.11 (LSU 23 2013-08-11)
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Fri, 15 Mar 2019, Oscar Salvador wrote:
> On Fri, Mar 15, 2019 at 01:47:33PM +0100, Michal Hocko wrote:
> > diff --git a/mm/debug.c b/mm/debug.c
> > index 1611cf00a137..499c26d5ebe5 100644
> > --- a/mm/debug.c
> > +++ b/mm/debug.c
> > @@ -78,6 +78,9 @@ void __dump_page(struct page *page, const char *reason)
> >  	else if (PageKsm(page))
> >  		pr_warn("ksm ");
> >  	else if (mapping) {
> > +		if (PageSwapCache(page))
> > +			mapping = page_swap_info(page)->swap_file->f_mapping;
> > +
> >  		pr_warn("%ps ", mapping->a_ops);
> >  		if (mapping->host->i_dentry.first) {
> >  			struct dentry *dentry;
> 
> This looks like a much nicer fix, indeed.
> I gave it a spin and it works.
> 
> Since the mapping is set during the swapon, I would assume that this should
> always work for swap.
> Although I am not sure if once you start playing with e.g zswap the picture can
> change.
> 
> Let us wait for Hugh and Jan.
> 
> Thanks Michal

Sorry, I don't agree that Michal's more sophisticated patch is nicer:
the appropriate patch was your original, just checking for NULL.

Though, would I be too snarky to suggest that your patch description
would be better at 2 lines than 90?  Swap mapping->host is NULL,
so of course __dump_page() needs to be careful about that.

I was a little disturbed to see __dump_page() now getting into dentries,
but admit that it can sometimes be very helpful to see the name of the
file involved; so if that is not in danger of breaking anything, okay.

It is very often useful to see if a page is PageSwapCache (typically
because that should account for 1 of its refcount); I cannot think of
a time when it's been useful to know the name of the underlying swap
device (if that's indeed what f_mapping leads to: it's new to me).
And if you need swp_type and swp_offset, they're in the raw output.

The cleverer __dump_page() tries to get, the more likely that it will
itself crash just when you need it most. Please just keep it simple.

Thanks,
Hugh