[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-ID: <YlEZahPCTI/qh/6u@infradead.org>
Date: Fri, 8 Apr 2022 22:28:10 -0700
From: Christoph Hellwig <hch@...radead.org>
To: Song Liu <song@...nel.org>
Cc: bpf@...r.kernel.org, netdev@...r.kernel.org, linux-mm@...ck.org,
ast@...nel.org, daniel@...earbox.net, andrii@...nel.org,
kernel-team@...com, akpm@...ux-foundation.org,
rick.p.edgecombe@...el.com, hch@...radead.org,
imbrenda@...ux.ibm.com
Subject: Re: [PATCH bpf 1/2] vmalloc: replace VM_NO_HUGE_VMAP with
VM_ALLOW_HUGE_VMAP
On Fri, Apr 08, 2022 at 03:34:42PM -0700, Song Liu wrote:
> Huge page backed vmalloc memory could benefit performance in many cases.
> Since some users of vmalloc may not be ready to handle huge pages,
> VM_NO_HUGE_VMAP was introduced to allow vmalloc users to opt-out huge
> pages. However, it is not easy to add VM_NO_HUGE_VMAP to all the users
> that may try to allocate >= PMD_SIZE pages, but are not ready to handle
> huge pages properly.
>
> Replace VM_NO_HUGE_VMAP with an opt-in flag, VM_ALLOW_HUGE_VMAP, so that
> users that benefit from huge pages could ask specificially.
Given that the huge page backing was added explicitly for some big boot
time allocated hashed,those should probably have the VM_ALLOW_HUGE_VMAP
added from the start (maybe not in this patch, but certainly in this
series). We'll probably also need a vmalloc_huge interface for those.
Powered by blists - more mailing lists