Re: [PATCH v4 bpf 0/4] vmalloc: bpf: introduce VM_ALLOW_HUGE_VMAP

From: Linus Torvalds <torvalds@linux-foundation.org>
To: Nicholas Piggin <npiggin@gmail.com>
Cc: Mike Rapoport <rppt@kernel.org>,
	"akpm@linux-foundation.org" <akpm@linux-foundation.org>,
	"ast@kernel.org" <ast@kernel.org>, "bp@alien8.de" <bp@alien8.de>,
	"bpf@vger.kernel.org" <bpf@vger.kernel.org>,
	"daniel@iogearbox.net" <daniel@iogearbox.net>,
	"dborkman@redhat.com" <dborkman@redhat.com>,
	"edumazet@google.com" <edumazet@google.com>,
	"hch@infradead.org" <hch@infradead.org>,
	"hpa@zytor.com" <hpa@zytor.com>,
	"imbrenda@linux.ibm.com" <imbrenda@linux.ibm.com>,
	Kernel Team <Kernel-team@fb.com>,
	"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
	"linux-mm@kvack.org" <linux-mm@kvack.org>,
	"mbenes@suse.cz" <mbenes@suse.cz>,
	"mcgrof@kernel.org" <mcgrof@kernel.org>,
	"pmladek@suse.com" <pmladek@suse.com>,
	"Edgecombe, Rick P" <rick.p.edgecombe@intel.com>,
	"song@kernel.org" <song@kernel.org>,
	Song Liu <songliubraving@fb.com>
Subject: Re: [PATCH v4 bpf 0/4] vmalloc: bpf: introduce VM_ALLOW_HUGE_VMAP
Date: Wed, 20 Apr 2022 23:02:48 -0700	[thread overview]
Message-ID: <CAHk-=wjYabTPnKiHgVzeKCaRkQaGVunwPbS+QeVb09Bm=YUEow@mail.gmail.com> (raw)
In-Reply-To: <CAHk-=wiQ5=S3m2+xRbm-1H8fuQwWfQxnO7tHhKg8FjegxzdVaQ@mail.gmail.com>

On Wed, Apr 20, 2022 at 10:48 PM Linus Torvalds
<torvalds@linux-foundation.org> wrote:
>
> The lagepage thing needs to be opt-in, and needs a lot more care.

Side note: part of the opt-in really should be about the performance impact.

It clearly can be quite noticeable, as outlined by that powerpc case
in commit 8abddd968a30 ("powerpc/64s/radix: Enable huge vmalloc
mappings"), but it presumably is some _particular_ case that actually
matters.

But it's equalyl clearly not the module code/data case, since
__module_alloc() explicitly disables largepages on powerpc.

At a guess, it's one or more of the large hash-table allocations.

And it would actually be interesting to hear *which*one*. From the
'git diff' workload, I'd expect it to be the dentry lookup hash table
- I can't think of anything else that would be vmalloc'ed that would
be remotely interesting - but who knows.

So I think the whole "opt in" isn't _purely_ about the "oh, random
cases are broken for odd reasons, so let's not enable it by default".

I think it would actually be good to literally mark the cases that
matter (and have the performance numbers for those cases).

               Linus