linux-api.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Thomas Gleixner <tglx@linutronix.de>
To: Peter Zijlstra <peterz@infradead.org>, axboe@kernel.dk
Cc: linux-kernel@vger.kernel.org, peterz@infradead.org,
	mingo@redhat.com, dvhart@infradead.org, dave@stgolabs.net,
	andrealmeid@igalia.com, Andrew Morton <akpm@linux-foundation.org>,
	urezki@gmail.com, hch@infradead.org, lstoakes@gmail.com,
	Arnd Bergmann <arnd@arndb.de>,
	linux-api@vger.kernel.org, linux-mm@kvack.org,
	linux-arch@vger.kernel.org, malteskarupke@web.de
Subject: Re: [PATCH v1 11/14] futex: Implement FUTEX2_NUMA
Date: Mon, 31 Jul 2023 19:36:21 +0200	[thread overview]
Message-ID: <87pm48m19m.ffs@tglx> (raw)
In-Reply-To: <20230721105744.434742902@infradead.org>

On Fri, Jul 21 2023 at 12:22, Peter Zijlstra wrote:
>  struct futex_hash_bucket *futex_hash(union futex_key *key)
>  {
> -	u32 hash = jhash2((u32 *)key, offsetof(typeof(*key), both.offset) / 4,
> +	u32 hash = jhash2((u32 *)key,
> +			  offsetof(typeof(*key), both.offset) / sizeof(u32),
>  			  key->both.offset);
> +	int node = key->both.node;
>  
> -	return &futex_queues[hash & (futex_hashsize - 1)];
> +	if (node == -1) {
> +		/*
> +		 * In case of !FLAGS_NUMA, use some unused hash bits to pick a
> +		 * node -- this ensures regular futexes are interleaved across
> +		 * the nodes and avoids having to allocate multiple
> +		 * hash-tables.
> +		 *
> +		 * NOTE: this isn't perfectly uniform, but it is fast and
> +		 * handles sparse node masks.
> +		 */
> +		node = (hash >> futex_hashshift) % nr_node_ids;

Is nr_node_ids guaranteed to be stable after init? It's marked
__read_mostly, but not __ro_after_init.

> +		if (!node_possible(node)) {
> +			node = find_next_bit_wrap(node_possible_map.bits,
> +						  nr_node_ids, node);
> +		}
> +	}
> +
> +	return &futex_queues[node][hash & (futex_hashsize - 1)];
>  }
>  	fshared = flags & FLAGS_SHARED;
> +	size = futex_size(flags);
>  
>  	/*
>  	 * The futex address must be "naturally" aligned.
>  	 */
>  	key->both.offset = address % PAGE_SIZE;
> -	if (unlikely((address % sizeof(u32)) != 0))
> +	if (unlikely((address % size) != 0))
>  		return -EINVAL;

Hmm. Shouldn't that have changed with the allowance of the 1 and 2 byte
futexes?

>  	address -= key->both.offset;
>  
> -	if (unlikely(!access_ok(uaddr, sizeof(u32))))
> +	if (flags & FLAGS_NUMA)
> +		size *= 2;
> +
> +	if (unlikely(!access_ok(uaddr, size)))
>  		return -EFAULT;
>  
>  	if (unlikely(should_fail_futex(fshared)))
>  		return -EFAULT;
>  
> +	key->both.node = -1;

Please put this into an else path.

> +	if (flags & FLAGS_NUMA) {
> +		void __user *naddr = uaddr + size/2;

size / 2;

> +
> +		if (futex_get_value(&node, naddr, flags))
> +			return -EFAULT;
> +
> +		if (node == -1) {
> +			node = numa_node_id();
> +			if (futex_put_value(node, naddr, flags))
> +				return -EFAULT;
> +		}
> +
> +		if (node >= MAX_NUMNODES || !node_possible(node))
> +			return -EINVAL;

That's clearly an else path too. No point in checking whether
numa_node_id() is valid. 

> +		key->both.node = node;
> +	}
>  
> +static inline unsigned int futex_size(unsigned int flags)
> +{
> +	unsigned int size = flags & FLAGS_SIZE_MASK;
> +	return 1 << size; /* {0,1,2,3} -> {1,2,4,8} */
> +}
> +
>  static inline bool futex_flags_valid(unsigned int flags)
>  {
>  	/* Only 64bit futexes for 64bit code */
> @@ -77,13 +83,19 @@ static inline bool futex_flags_valid(uns
>  	if ((flags & FLAGS_SIZE_MASK) != FLAGS_SIZE_32)
>  		return false;
>  
> -	return true;
> -}
> +	/*
> +	 * Must be able to represent both NUMA_NO_NODE and every valid nodeid
> +	 * in a futex word.
> +	 */
> +	if (flags & FLAGS_NUMA) {
> +		int bits = 8 * futex_size(flags);
> +		u64 max = ~0ULL;
> +		max >>= 64 - bits;
Your newline key is broken, right?
> +		if (nr_node_ids >= max)
> +			return false;
> +	}

Thanks,

        tglx

  parent reply	other threads:[~2023-07-31 17:36 UTC|newest]

Thread overview: 60+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-07-21 10:22 [PATCH v1 00/14] futex: More futex2 bits Peter Zijlstra
2023-07-21 10:22 ` [PATCH v1 01/14] futex: Clarify FUTEX2 flags Peter Zijlstra
2023-07-31 16:08   ` Thomas Gleixner
2023-07-21 10:22 ` [PATCH v1 02/14] futex: Extend the " Peter Zijlstra
2023-07-21 15:47   ` Arnd Bergmann
2023-07-21 18:52     ` Peter Zijlstra
2023-07-31 16:11   ` Thomas Gleixner
2023-07-31 16:25     ` Peter Zijlstra
2023-07-31 17:16     ` Thomas Gleixner
2023-07-31 17:35       ` Peter Zijlstra
2023-07-31 20:52         ` Thomas Gleixner
2023-07-31 17:42     ` Thomas Gleixner
2023-07-31 19:20       ` Peter Zijlstra
2023-07-31 21:14         ` Thomas Gleixner
2023-07-31 21:33           ` Peter Zijlstra
2023-07-31 22:43             ` Thomas Gleixner
2023-07-31 22:59               ` Peter Zijlstra
2023-08-01  8:49                 ` Thomas Gleixner
2023-08-01  6:02               ` Arnd Bergmann
2023-07-21 10:22 ` [PATCH v1 03/14] futex: Flag conversion Peter Zijlstra
2023-07-31 16:21   ` Thomas Gleixner
2023-07-31 16:26     ` Peter Zijlstra
2023-07-21 10:22 ` [PATCH v1 04/14] futex: Validate futex value against futex size Peter Zijlstra
2023-07-31 17:12   ` Thomas Gleixner
2023-07-21 10:22 ` [PATCH v1 05/14] futex: Add sys_futex_wake() Peter Zijlstra
2023-07-21 15:41   ` Arnd Bergmann
2023-07-21 18:54     ` Peter Zijlstra
2023-07-21 21:23       ` Arnd Bergmann
2023-07-25  7:22   ` Geert Uytterhoeven
2023-07-21 10:22 ` [PATCH v1 06/14] futex: Add sys_futex_wait() Peter Zijlstra
2023-07-25  7:22   ` Geert Uytterhoeven
2023-07-31 16:35   ` Thomas Gleixner
2023-07-21 10:22 ` [PATCH v1 07/14] futex: Propagate flags into get_futex_key() Peter Zijlstra
2023-07-31 16:36   ` Thomas Gleixner
2023-07-21 10:22 ` [PATCH v1 08/14] futex: Add flags2 argument to futex_requeue() Peter Zijlstra
2023-07-31 16:43   ` Thomas Gleixner
2023-07-21 10:22 ` [PATCH v1 09/14] futex: Add sys_futex_requeue() Peter Zijlstra
2023-07-25  7:23   ` Geert Uytterhoeven
2023-07-31 17:19   ` Thomas Gleixner
2023-07-31 17:38     ` Peter Zijlstra
2023-07-21 10:22 ` [PATCH v1 10/14] mm: Add vmalloc_huge_node() Peter Zijlstra
2023-07-24 13:46   ` Christoph Hellwig
2023-07-21 10:22 ` [PATCH v1 11/14] futex: Implement FUTEX2_NUMA Peter Zijlstra
2023-07-21 12:16   ` Peter Zijlstra
2023-07-31 17:36   ` Thomas Gleixner [this message]
2023-07-31 18:03     ` Peter Zijlstra
2023-07-31 21:26       ` Thomas Gleixner
2024-06-12 17:07       ` Christoph Lameter (Ampere)
2024-06-12 17:23   ` Christoph Lameter (Ampere)
2024-06-12 17:44     ` Peter Zijlstra
2024-06-18 16:53       ` Christoph Lameter (Ampere)
2024-10-25  8:58     ` Peter Zijlstra
2024-10-25 19:36       ` Christoph Lameter (Ampere)
2024-10-26  7:21         ` Peter Zijlstra
2024-10-28 22:32           ` Christoph Lameter (Ampere)
2023-07-21 10:22 ` [PATCH v1 12/14] futex: Propagate flags into futex_get_value_locked() Peter Zijlstra
2023-07-21 10:22 ` [PATCH v1 13/14] futex: Enable FUTEX2_{8,16} Peter Zijlstra
2023-07-21 10:22 ` [PATCH v1 14/14] futex,selftests: Extend the futex selftests Peter Zijlstra
2023-07-21 14:42 ` [PATCH v1 00/14] futex: More futex2 bits Jens Axboe
2023-07-21 15:49 ` Arnd Bergmann

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=87pm48m19m.ffs@tglx \
    --to=tglx@linutronix.de \
    --cc=akpm@linux-foundation.org \
    --cc=andrealmeid@igalia.com \
    --cc=arnd@arndb.de \
    --cc=axboe@kernel.dk \
    --cc=dave@stgolabs.net \
    --cc=dvhart@infradead.org \
    --cc=hch@infradead.org \
    --cc=linux-api@vger.kernel.org \
    --cc=linux-arch@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=lstoakes@gmail.com \
    --cc=malteskarupke@web.de \
    --cc=mingo@redhat.com \
    --cc=peterz@infradead.org \
    --cc=urezki@gmail.com \
    /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).