All of lore.kernel.org
 help / color / mirror / Atom feed
From: Mike Travis <travis@sgi.com>
To: David Rientjes <rientjes@google.com>
Cc: Yinghai Lu <yhlu.kernel@gmail.com>,
	Andrew Morton <akpm@linux-foundation.org>,
	Andi Kleen <ak@suse.de>,
	mingo@elte.hu, Christoph Lameter <clameter@sgi.com>,
	linux-mm@kvack.org, linux-kernel@vger.kernel.org,
	Eric Dumazet <dada1@cosmosbay.com>
Subject: Re: [PATCH 1/5] x86: Change size of node ids from u8 to u16 fixup
Date: Sat, 19 Jan 2008 13:25:32 -0800	[thread overview]
Message-ID: <47926ACC.4060707@sgi.com> (raw)
In-Reply-To: <alpine.DEB.0.9999.0801182026130.32726@chino.kir.corp.google.com>

David Rientjes wrote:
> On Fri, 18 Jan 2008, Yinghai Lu wrote:
> 
>>> +#if MAX_NUMNODES > 256
>>> +typedef u16 numanode_t;
>>> +#else
>>> +typedef u8 numanode_t;
>>> +#endif
>>> +
>>>  #endif /* _LINUX_NUMA_H */
>> that is wrong, you can not change pxm_to_node_map from int to u8 or u16.
>>

Thanks for finding this!

> 
> Yeah, NID_INVAL is negative so no unsigned type will work here, 
> unfortunately.  And that reduces the intended savings of your change since 
> the smaller type can only be used with a smaller CONFIG_NODES_SHIFT.
> 

Excuse my ignorance but why wouldn't this work:

static numanode_t pxm_to_node_map[MAX_PXM_DOMAINS]
                                = { [0 ... MAX_PXM_DOMAINS - 1] = NUMA_NO_NODE };
...
>> int acpi_map_pxm_to_node(int pxm)
>> {
>         int node = pxm_to_node_map[pxm];
> 
>         if (node < 0)

	   numanode_t node = pxm_to_node_map[pxm];

	   if (node != NUMA_NO_NODE) {
>>                 if (nodes_weight(nodes_found_map) >= MAX_NUMNODES)
>>                         return NID_INVAL;
>>                 node = first_unset_node(nodes_found_map);
>>                 __acpi_map_pxm_to_node(pxm, node);
>>                 node_set(node, nodes_found_map);
>>         }

or change:
	#define NID_INVAL       (-1)
to
	#define NID_INVAL       ((numanode_t)(-1))
...
	   if (node != NID_INVAL) {
>>                 if (nodes_weight(nodes_found_map) >= MAX_NUMNODES)
>>                         return NID_INVAL;
>>                 node = first_unset_node(nodes_found_map);
>>                 __acpi_map_pxm_to_node(pxm, node);
>>                 node_set(node, nodes_found_map);
>>         }

Though why there two "node invalid" values I'm not sure... ?

>>
>>         return node;
>> }

And btw, shouldn't the pxm value be sized to numanode_t size as well?
Will it ever be larger than the largest node id?

Thanks,
Mike

WARNING: multiple messages have this Message-ID (diff)
From: Mike Travis <travis@sgi.com>
To: David Rientjes <rientjes@google.com>
Cc: Yinghai Lu <yhlu.kernel@gmail.com>,
	Andrew Morton <akpm@linux-foundation.org>,
	Andi Kleen <ak@suse.de>,
	mingo@elte.hu, Christoph Lameter <clameter@sgi.com>,
	linux-mm@kvack.org, linux-kernel@vger.kernel.org,
	Eric Dumazet <dada1@cosmosbay.com>
Subject: Re: [PATCH 1/5] x86: Change size of node ids from u8 to u16 fixup
Date: Sat, 19 Jan 2008 13:25:32 -0800	[thread overview]
Message-ID: <47926ACC.4060707@sgi.com> (raw)
In-Reply-To: <alpine.DEB.0.9999.0801182026130.32726@chino.kir.corp.google.com>

David Rientjes wrote:
> On Fri, 18 Jan 2008, Yinghai Lu wrote:
> 
>>> +#if MAX_NUMNODES > 256
>>> +typedef u16 numanode_t;
>>> +#else
>>> +typedef u8 numanode_t;
>>> +#endif
>>> +
>>>  #endif /* _LINUX_NUMA_H */
>> that is wrong, you can not change pxm_to_node_map from int to u8 or u16.
>>

Thanks for finding this!

> 
> Yeah, NID_INVAL is negative so no unsigned type will work here, 
> unfortunately.  And that reduces the intended savings of your change since 
> the smaller type can only be used with a smaller CONFIG_NODES_SHIFT.
> 

Excuse my ignorance but why wouldn't this work:

static numanode_t pxm_to_node_map[MAX_PXM_DOMAINS]
                                = { [0 ... MAX_PXM_DOMAINS - 1] = NUMA_NO_NODE };
...
>> int acpi_map_pxm_to_node(int pxm)
>> {
>         int node = pxm_to_node_map[pxm];
> 
>         if (node < 0)

	   numanode_t node = pxm_to_node_map[pxm];

	   if (node != NUMA_NO_NODE) {
>>                 if (nodes_weight(nodes_found_map) >= MAX_NUMNODES)
>>                         return NID_INVAL;
>>                 node = first_unset_node(nodes_found_map);
>>                 __acpi_map_pxm_to_node(pxm, node);
>>                 node_set(node, nodes_found_map);
>>         }

or change:
	#define NID_INVAL       (-1)
to
	#define NID_INVAL       ((numanode_t)(-1))
...
	   if (node != NID_INVAL) {
>>                 if (nodes_weight(nodes_found_map) >= MAX_NUMNODES)
>>                         return NID_INVAL;
>>                 node = first_unset_node(nodes_found_map);
>>                 __acpi_map_pxm_to_node(pxm, node);
>>                 node_set(node, nodes_found_map);
>>         }

Though why there two "node invalid" values I'm not sure... ?

>>
>>         return node;
>> }

And btw, shouldn't the pxm value be sized to numanode_t size as well?
Will it ever be larger than the largest node id?

Thanks,
Mike

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

  parent reply	other threads:[~2008-01-19 21:25 UTC|newest]

Thread overview: 81+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2008-01-18 18:30 [PATCH 0/5] x86: Reduce memory usage for large count NR_CPUs fixup travis
2008-01-18 18:30 ` travis
2008-01-18 18:30 ` [PATCH 1/5] x86: Change size of node ids from u8 to u16 fixup travis
2008-01-18 18:30   ` travis
2008-01-18 19:56   ` Jan Engelhardt
2008-01-18 19:56     ` Jan Engelhardt
2008-01-18 19:59     ` Mike Travis
2008-01-18 19:59       ` Mike Travis
2008-01-19  4:03   ` Yinghai Lu
2008-01-19  4:03     ` Yinghai Lu
2008-01-19  4:36     ` David Rientjes
2008-01-19  4:36       ` David Rientjes
2008-01-19  4:43       ` Yinghai Lu
2008-01-19  4:43         ` Yinghai Lu
2008-01-19  5:17         ` David Rientjes
2008-01-19  5:17           ` David Rientjes
2008-01-19  6:20           ` Yinghai Lu
2008-01-19  6:20             ` Yinghai Lu
2008-01-19 21:25       ` Mike Travis [this message]
2008-01-19 21:25         ` Mike Travis
2008-01-19 22:33         ` David Rientjes
2008-01-19 22:33           ` David Rientjes
2008-01-20  0:41           ` Mike Travis
2008-01-20  0:41             ` Mike Travis
2008-01-20  1:31             ` Yinghai Lu
2008-01-20  1:31               ` Yinghai Lu
2008-01-20  6:22             ` David Rientjes
2008-01-20  6:22               ` David Rientjes
2008-01-18 18:30 ` [PATCH 2/5] x86: Change NR_CPUS arrays in numa_64 fixup travis
2008-01-18 18:30   ` travis
2008-01-18 18:30 ` [PATCH 3/5] x86: Change bios_cpu_apicid to percpu data variable fixup travis
2008-01-18 18:30   ` travis
2008-01-18 18:30 ` [PATCH 4/5] x86: Add config variables for SMP_MAX travis
2008-01-18 18:30   ` travis
2008-01-18 20:04   ` Ingo Oeser
2008-01-18 20:04     ` Ingo Oeser
2008-01-18 20:10     ` Christoph Lameter
2008-01-18 20:10       ` Christoph Lameter
2008-01-18 20:14     ` Mike Travis
2008-01-18 20:14       ` Mike Travis
2008-01-18 20:36       ` Andi Kleen
2008-01-18 20:36         ` Andi Kleen
2008-01-18 20:48         ` Mike Travis
2008-01-18 20:48           ` Mike Travis
2008-01-18 21:02         ` [PATCH 4/5] x86: Add config variables for SMP_MAX II Andi Kleen
2008-01-18 21:02           ` Andi Kleen
2008-01-18 20:48       ` [PATCH 4/5] x86: Add config variables for SMP_MAX Ingo Molnar
2008-01-18 20:48         ` Ingo Molnar
2008-01-18 20:55         ` Mike Travis
2008-01-18 20:55           ` Mike Travis
2008-01-18 20:58           ` Andi Kleen
2008-01-18 20:58             ` Andi Kleen
2008-01-28 16:45       ` Paul Jackson
2008-01-28 16:45         ` Paul Jackson
2008-01-28 17:00         ` Andi Kleen
2008-01-28 17:00           ` Andi Kleen
2008-01-18 20:46   ` Ingo Molnar
2008-01-18 20:46     ` Ingo Molnar
2008-01-19 14:52   ` Ingo Molnar
2008-01-19 15:15     ` Ingo Molnar
2008-01-19 15:15       ` Ingo Molnar
2008-01-19 15:24       ` Ingo Molnar
2008-01-19 15:24         ` Ingo Molnar
2008-01-19 21:52         ` Mike Travis
2008-01-19 21:52           ` Mike Travis
2008-01-19 23:24         ` Mike Travis
2008-01-19 23:24           ` Mike Travis
2008-01-20  1:14         ` Mike Travis
2008-01-20  1:14           ` Mike Travis
2008-01-19 21:39     ` Mike Travis
2008-01-19 21:39       ` Mike Travis
2008-01-18 18:30 ` [PATCH 5/5] x86: Add debug of invalid per_cpu map accesses travis
2008-01-18 18:30   ` travis
2008-01-18 18:33   ` Andi Kleen
2008-01-18 18:33     ` Andi Kleen
2008-01-18 18:49     ` Mike Travis
2008-01-18 18:49       ` Mike Travis
2008-01-18 18:56     ` Christoph Lameter
2008-01-18 18:56       ` Christoph Lameter
2008-01-18 20:49     ` Ingo Molnar
2008-01-18 20:49       ` Ingo Molnar

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=47926ACC.4060707@sgi.com \
    --to=travis@sgi.com \
    --cc=ak@suse.de \
    --cc=akpm@linux-foundation.org \
    --cc=clameter@sgi.com \
    --cc=dada1@cosmosbay.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=mingo@elte.hu \
    --cc=rientjes@google.com \
    --cc=yhlu.kernel@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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.