From: Andrew Morton <akpm@linux-foundation.org>
To: "Uladzislau Rezki (Sony)" <urezki@gmail.com>
Cc: Daniel Wagner <dwagner@suse.de>,
Sebastian Andrzej Siewior <bigeasy@linutronix.de>,
Thomas Gleixner <tglx@linutronix.de>,
linux-mm@kvack.org, LKML <linux-kernel@vger.kernel.org>,
Peter Zijlstra <peterz@infradead.org>,
Hillf Danton <hdanton@sina.com>, Michal Hocko <mhocko@suse.com>,
Matthew Wilcox <willy@infradead.org>,
Oleksiy Avramchenko <oleksiy.avramchenko@sonymobile.com>,
Steven Rostedt <rostedt@goodmis.org>
Subject: Re: [PATCH 1/1] mm/vmalloc: remove preempt_disable/enable when do preloading
Date: Wed, 9 Oct 2019 15:19:01 -0700 [thread overview]
Message-ID: <20191009151901.1be5f7211db291e4bd2da8ca@linux-foundation.org> (raw)
In-Reply-To: <20191009164934.10166-1-urezki@gmail.com>
On Wed, 9 Oct 2019 18:49:34 +0200 "Uladzislau Rezki (Sony)" <urezki@gmail.com> wrote:
> Get rid of preempt_disable() and preempt_enable() when the
> preload is done for splitting purpose. The reason is that
> calling spin_lock() with disabled preemtion is forbidden in
> CONFIG_PREEMPT_RT kernel.
>
> Therefore, we do not guarantee that a CPU is preloaded, instead
> we minimize the case when it is not with this change.
>
> For example i run the special test case that follows the preload
> pattern and path. 20 "unbind" threads run it and each does
> 1000000 allocations. Only 3.5 times among 1000000 a CPU was
> not preloaded thus. So it can happen but the number is rather
> negligible.
>
> ...
>
A few questions about the resulting alloc_vmap_area():
: static struct vmap_area *alloc_vmap_area(unsigned long size,
: unsigned long align,
: unsigned long vstart, unsigned long vend,
: int node, gfp_t gfp_mask)
: {
: struct vmap_area *va, *pva;
: unsigned long addr;
: int purged = 0;
:
: BUG_ON(!size);
: BUG_ON(offset_in_page(size));
: BUG_ON(!is_power_of_2(align));
:
: if (unlikely(!vmap_initialized))
: return ERR_PTR(-EBUSY);
:
: might_sleep();
:
: va = kmem_cache_alloc_node(vmap_area_cachep,
: gfp_mask & GFP_RECLAIM_MASK, node);
Why does this use GFP_RECLAIM_MASK? Please add a comment explaining
this.
: if (unlikely(!va))
: return ERR_PTR(-ENOMEM);
:
: /*
: * Only scan the relevant parts containing pointers to other objects
: * to avoid false negatives.
: */
: kmemleak_scan_area(&va->rb_node, SIZE_MAX, gfp_mask & GFP_RECLAIM_MASK);
:
: retry:
: /*
: * Preload this CPU with one extra vmap_area object. It is used
: * when fit type of free area is NE_FIT_TYPE. Please note, it
: * does not guarantee that an allocation occurs on a CPU that
: * is preloaded, instead we minimize the case when it is not.
: * It can happen because of migration, because there is a race
: * until the below spinlock is taken.
: *
: * The preload is done in non-atomic context, thus it allows us
: * to use more permissive allocation masks to be more stable under
: * low memory condition and high memory pressure.
: *
: * Even if it fails we do not really care about that. Just proceed
: * as it is. "overflow" path will refill the cache we allocate from.
: */
: if (!this_cpu_read(ne_fit_preload_node)) {
Readability nit: local `pva' should be defined here, rather than having
function-wide scope.
: pva = kmem_cache_alloc_node(vmap_area_cachep, GFP_KERNEL, node);
Why doesn't this honour gfp_mask? If it's not a bug, please add
comment explaining this.
The kmem_cache_alloc() in adjust_va_to_fit_type() omits the caller's
gfp_mask also. If not a bug, please document the unexpected behaviour.
:
: if (this_cpu_cmpxchg(ne_fit_preload_node, NULL, pva)) {
: if (pva)
: kmem_cache_free(vmap_area_cachep, pva);
: }
: }
:
: spin_lock(&vmap_area_lock);
:
: /*
: * If an allocation fails, the "vend" address is
: * returned. Therefore trigger the overflow path.
: */
As for the intent of this patch, why not preallocate the vmap_area
outside the spinlock and use it within the spinlock? Does spin_lock()
disable preemption on RT? I forget, but it doesn't matter much anyway
- doing this will make the code better in the regular kernel I think?
Something like this:
struct vmap_area *pva = NULL;
...
if (!this_cpu_read(ne_fit_preload_node))
pva = kmem_cache_alloc_node(vmap_area_cachep, ...);
spin_lock(&vmap_area_lock);
if (pva && __this_cpu_cmpxchg(ne_fit_preload_node, NULL, pva))
kmem_cache_free(vmap_area_cachep, pva);
next prev parent reply other threads:[~2019-10-09 22:19 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-10-09 16:49 [PATCH 1/1] mm/vmalloc: remove preempt_disable/enable when do preloading Uladzislau Rezki (Sony)
2019-10-09 18:21 ` Steven Rostedt
2019-10-09 22:19 ` Andrew Morton [this message]
2019-10-10 2:17 ` Steven Rostedt
2019-10-10 15:17 ` Uladzislau Rezki
2019-10-11 23:55 ` Andrew Morton
2019-10-14 14:27 ` Uladzislau Rezki
2019-10-14 16:30 ` Michal Hocko
2019-10-15 9:54 ` Uladzislau Rezki
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20191009151901.1be5f7211db291e4bd2da8ca@linux-foundation.org \
--to=akpm@linux-foundation.org \
--cc=bigeasy@linutronix.de \
--cc=dwagner@suse.de \
--cc=hdanton@sina.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=mhocko@suse.com \
--cc=oleksiy.avramchenko@sonymobile.com \
--cc=peterz@infradead.org \
--cc=rostedt@goodmis.org \
--cc=tglx@linutronix.de \
--cc=urezki@gmail.com \
--cc=willy@infradead.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).