From: "Aneesh Kumar K.V" <aneesh.kumar@linux.ibm.com>
To: David Hildenbrand <david@redhat.com>,
	linux-mm@kvack.org, akpm@linux-foundation.org,
	mpe@ellerman.id.au, linuxppc-dev@lists.ozlabs.org,
	npiggin@gmail.com, christophe.leroy@csgroup.eu
Cc: Vishal Verma <vishal.l.verma@intel.com>,
	Michal Hocko <mhocko@suse.com>,
	Oscar Salvador <osalvador@suse.de>
Subject: Re: [PATCH v3 4/7] mm/hotplug: Allow pageblock alignment via altmap reservation
Date: Wed, 12 Jul 2023 19:20:14 +0530	[thread overview]
Message-ID: <87wmz56xyh.fsf@linux.ibm.com> (raw)
In-Reply-To: <57dd0568-ee56-ff8d-3ba3-a9089a2ab386@redhat.com>
David Hildenbrand <david@redhat.com> writes:
> On 12.07.23 05:16, Aneesh Kumar K V wrote:
>> On 7/11/23 10:49 PM, David Hildenbrand wrote:
>>> On 11.07.23 06:48, Aneesh Kumar K.V wrote:
>>>> Add a new kconfig option that can be selected if we want to allow
>>>> pageblock alignment by reserving pages in the vmemmap altmap area.
>>>> This implies we will be reserving some pages for every memoryblock
>>>> This also allows the memmap on memory feature to be widely useful
>>>> with different memory block size values.
>>>
>>> "reserving pages" is a nice way of saying "wasting memory". :) Let's spell that out.
>>>
>>> I think we have to find a better name for this, and I think we should have a toggle similar to memory_hotplug.memmap_on_memory. This should be an admin decision, not some kernel config option.
>>>
>>>
>>> memory_hotplug.force_memmap_on_memory
>>>
>>> "Enable the memmap on memory feature even if it could result in memory waste due to memmap size limitations. For example, if the memmap for a memory block requires 1 MiB, but the pageblock size is 2 MiB, 1 MiB
>>> of hotplugged memory will be wasted. Note that there are still cases where the feature cannot be enforced: for example, if the memmap is smaller than a single page, or if the architecture does not support the forced mode in all configurations."
>>>
>>> Thoughts?
>>>
>> 
>> With module parameter, do we still need the Kconfig option?
>
> No.
>
> Sleeping over this, maybe we can convert the existing 
> memory_hotplug.memmap_on_memory parameter to also accept "force".
>
How about this? 
modified   mm/memory_hotplug.c
@@ -45,13 +45,67 @@
 /*
  * memory_hotplug.memmap_on_memory parameter
  */
-static bool memmap_on_memory __ro_after_init;
-module_param(memmap_on_memory, bool, 0444);
-MODULE_PARM_DESC(memmap_on_memory, "Enable memmap on memory for memory hotplug");
+enum {
+	MEMMAP_ON_MEMORY_DISABLE = 0,
+	MEMMAP_ON_MEMORY_ENABLE,
+	FORCE_MEMMAP_ON_MEMORY,
+};
+static int memmap_mode __read_mostly = MEMMAP_ON_MEMORY_DISABLE;
+static const char *memmap_on_memory_to_str[] = {
+	[MEMMAP_ON_MEMORY_DISABLE]  = "disable",
+	[MEMMAP_ON_MEMORY_ENABLE]   = "enable",
+	[FORCE_MEMMAP_ON_MEMORY]    = "force",
+};
+
+static inline unsigned long memory_block_align_base(unsigned long size)
+{
+	if (memmap_mode == FORCE_MEMMAP_ON_MEMORY) {
+		unsigned long align;
+		unsigned long nr_vmemmap_pages = size >> PAGE_SHIFT;
+		unsigned long vmemmap_size;
+
+		vmemmap_size = DIV_ROUND_UP(nr_vmemmap_pages * sizeof(struct page), PAGE_SIZE);
+		align = pageblock_align(vmemmap_size) - vmemmap_size;
+		return align;
+	} else
+		return 0;
+}
+
+static int set_memmap_mode(const char *val, const struct kernel_param *kp)
+{
+	int ret = sysfs_match_string(memmap_on_memory_to_str, val);
+
+	if (ret < 0)
+		return ret;
+	*((int *)kp->arg) = ret;
+	if (ret == FORCE_MEMMAP_ON_MEMORY) {
+		pr_info("Memory hotplug will reserve %ld pages in each memory block\n",
+			memory_block_align_base(memory_block_size_bytes()));
+	}
+	return 0;
+}
+
+static int get_memmap_mode(char *buffer, const struct kernel_param *kp)
+{
+	return sprintf(buffer, "%s\n", memmap_on_memory_to_str[*((int *)kp->arg)]);
+}
+
+static const struct kernel_param_ops memmap_mode_ops = {
+	.set = set_memmap_mode,
+	.get = get_memmap_mode,
+};
+module_param_cb(memmap_on_memory, &memmap_mode_ops, &memmap_mode, 0644);
+MODULE_PARM_DESC(memmap_on_memory, "Enable memmap on memory for memory hotplug\n"
+	"With value \"force\" it could result in memory waste due to memmap size limitations \n"
+	"For example, if the memmap for a memory block requires 1 MiB, but the pageblock \n"
+	"size is 2 MiB, 1 MiB of hotplugged memory will be wasted. Note that there are \n"
+	"still cases where the feature cannot be enforced: for example, if the memmap is \n"
+	"smaller than a single page, or if the architecture does not support the forced \n"
+	"mode in all configurations. (disable/enable/force)");
 
 static inline bool mhp_memmap_on_memory(void)
 {
-	return memmap_on_memory;
+	return !!memmap_mode;
 }
 #else
We can also enable runtime enable/disable/force the feature. We just
need to make sure on try_remove_memory we lookup for altmap correctly. 
next prev parent reply	other threads:[~2023-07-12 13:51 UTC|newest]
Thread overview: 28+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-07-11  4:48 [PATCH v3 0/7] Add support for memmap on memory feature on ppc64 Aneesh Kumar K.V
2023-07-11  4:48 ` [PATCH v3 1/7] mm/hotplug: Simplify ARCH_MHP_MEMMAP_ON_MEMORY_ENABLE kconfig Aneesh Kumar K.V
2023-07-11  4:48 ` [PATCH v3 2/7] mm/hotplug: Allow memmap on memory hotplug request to fallback Aneesh Kumar K.V
2023-07-11 10:23   ` David Hildenbrand
2023-07-11 15:58     ` Aneesh Kumar K V
2023-07-11  4:48 ` [PATCH v3 3/7] mm/hotplug: Allow architecture to override memmap on memory support check Aneesh Kumar K.V
2023-07-11 10:36   ` David Hildenbrand
2023-07-11 16:07     ` Aneesh Kumar K V
2023-07-11 16:09       ` David Hildenbrand
2023-07-12 20:07         ` John Hubbard
2023-07-13  9:08           ` David Hildenbrand
2023-07-14 23:14             ` John Hubbard
2023-07-11  4:48 ` [PATCH v3 4/7] mm/hotplug: Allow pageblock alignment via altmap reservation Aneesh Kumar K.V
2023-07-11  6:21   ` Huang, Ying
2023-07-11  8:20     ` Aneesh Kumar K V
2023-07-11 17:19   ` David Hildenbrand
2023-07-12  3:16     ` Aneesh Kumar K V
2023-07-12  7:22       ` David Hildenbrand
2023-07-12 13:50         ` Aneesh Kumar K.V [this message]
2023-07-12 19:06           ` David Hildenbrand
2023-07-11  4:48 ` [PATCH v3 5/7] powerpc/book3s64/memhotplug: Enable memmap on memory for radix Aneesh Kumar K.V
2023-07-11 15:26   ` David Hildenbrand
2023-07-11 15:40     ` Aneesh Kumar K V
2023-07-11 15:44       ` David Hildenbrand
2023-07-11 15:46         ` Aneesh Kumar K V
2023-07-11  4:48 ` [PATCH v3 6/7] dax/kmem: Always enroll hotplugged memory for memmap_on_memory Aneesh Kumar K.V
2023-07-11 10:21   ` David Hildenbrand
2023-07-11  4:48 ` [PATCH v3 7/7] mm/hotplug: Embed vmem_altmap details in memory block Aneesh Kumar K.V
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=87wmz56xyh.fsf@linux.ibm.com \
    --to=aneesh.kumar@linux.ibm.com \
    --cc=akpm@linux-foundation.org \
    --cc=christophe.leroy@csgroup.eu \
    --cc=david@redhat.com \
    --cc=linux-mm@kvack.org \
    --cc=linuxppc-dev@lists.ozlabs.org \
    --cc=mhocko@suse.com \
    --cc=mpe@ellerman.id.au \
    --cc=npiggin@gmail.com \
    --cc=osalvador@suse.de \
    --cc=vishal.l.verma@intel.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).