| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <A846613E-A2B4-4B56-B368-5786F572F168@nvidia.com>
Date: Tue, 24 Sep 2024 15:27:52 +0000
From: Bruno Faccini <bfaccini@...dia.com>
To: Mike Rapoport <rppt@...nel.org>
CC: "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
"linux-mm@...ck.org" <linux-mm@...ck.org>, "akpm@...ux-foundation.org"
<akpm@...ux-foundation.org>, Zi Yan <ziy@...dia.com>, Timur Tabi
<ttabi@...dia.com>, John Hubbard <jhubbard@...dia.com>
Subject: Re: [PATCH] mm/fake-numa: per-phys node fake size
> I'm not sure I understand the problem you are trying to solve.
> Can you provide more specific example?
Thank you for your feedback already.
I will try to be more precise about the situation I have encountered with your original set of patches and how I thought it could be solved.
On a system with 2 physical Numa nodes each with 480GB local memory, where the biggest part of reserved memory (~ 309MB) is from node 0 with a small part (~ 51MB) from node 1, leading to the fake node size of ~ <120GB being determined.
But when allocating fake nodes from physical nodes, with let say fake=8 boot parameter being used, we ended with less (7) than expected, because there was not enough room to allocate 8/2 fake nodes in physical node 0, due to too big size evaluation.
I don't think that fake=N allocation method is intended to get fake nodes with equal size, but to get this exact number of nodes.
This is why I think we should use a per-phys node size for the fake nodes it will host.
Hope this clarifies the reason and intent for my patch, have a good day,
Bruno
On 24/09/2024 12:43, "Mike Rapoport" <rppt@...nel.org <mailto:rppt@...nel.org>> wrote:
External email: Use caution opening links or attachments
On Sat, Sep 21, 2024 at 01:13:49AM -0700, Bruno Faccini wrote:
> Determine fake numa node size on a per-phys node basis to
> handle cases where there are big differences of reserved
> memory size inside physical nodes, this will allow to get
> the expected number of nodes evenly interleaved.
>
> Consider a system with 2 physical Numa nodes where almost
> all reserved memory sits into a single node, computing the
> fake-numa nodes (fake=N) size as the ratio of all
> available/non-reserved memory can cause the inability to
> create N/2 fake-numa nodes in the physical node.
I'm not sure I understand the problem you are trying to solve.
Can you provide more specific example?
> Signed-off-by: Bruno Faccini <bfaccini@...dia.com <mailto:bfaccini@...dia.com>>
> ---
> mm/numa_emulation.c | 66 ++++++++++++++++++++++++++-------------------
> 1 file changed, 39 insertions(+), 27 deletions(-)
>
> diff --git a/mm/numa_emulation.c b/mm/numa_emulation.c
> index 031fb9961bf7..0c72c85cfc10 100644
> --- a/mm/numa_emulation.c
> +++ b/mm/numa_emulation.c
> @@ -77,20 +77,19 @@ static int __init emu_setup_memblk(struct numa_meminfo *ei,
> }
>
> /*
> - * Sets up nr_nodes fake nodes interleaved over physical nodes ranging from addr
> - * to max_addr.
> + * Sets up nr_nodes fake nodes interleaved over all physical nodes
> *
> * Returns zero on success or negative on error.
> */
> static int __init split_nodes_interleave(struct numa_meminfo *ei,
> struct numa_meminfo *pi,
> - u64 addr, u64 max_addr, int nr_nodes)
> + int nr_nodes)
> {
> nodemask_t physnode_mask = numa_nodes_parsed;
> - u64 size;
> - int big;
> - int nid = 0;
> - int i, ret;
> + int nid = 0, physnodes_with_mem = 0;
> + int i, ret, phys_blk;
> + static u64 sizes[MAX_NUMNODES] __initdata;
> + static int bigs[MAX_NUMNODES] __initdata;
>
> if (nr_nodes <= 0)
> return -1;
> @@ -100,25 +99,41 @@ static int __init split_nodes_interleave(struct numa_meminfo *ei,
> nr_nodes = MAX_NUMNODES;
> }
>
> - /*
> - * Calculate target node size. x86_32 freaks on __udivdi3() so do
> - * the division in ulong number of pages and convert back.
> - */
> - size = max_addr - addr - mem_hole_size(addr, max_addr);
> - size = PFN_PHYS((unsigned long)(size >> PAGE_SHIFT) / nr_nodes);
> + /* count physical nodes with memory */
> + for_each_node_mask(i, physnode_mask) {
> + phys_blk = emu_find_memblk_by_nid(i, pi);
> + if (phys_blk < 0)
> + continue;
> + physnodes_with_mem++;
> + }
>
> /*
> - * Calculate the number of big nodes that can be allocated as a result
> - * of consolidating the remainder.
> + * Calculate target fake nodes sizes for each physical node with memory.
> + * x86_32 freaks on __udivdi3() so do the division in ulong number of
> + * pages and convert back.
> */
> - big = ((size & ~FAKE_NODE_MIN_HASH_MASK) * nr_nodes) /
> - FAKE_NODE_MIN_SIZE;
> + for_each_node_mask(i, physnode_mask) {
> + phys_blk = emu_find_memblk_by_nid(i, pi);
> + if (phys_blk < 0)
> + continue;
>
> - size &= FAKE_NODE_MIN_HASH_MASK;
> - if (!size) {
> - pr_err("Not enough memory for each node. "
> - "NUMA emulation disabled.\n");
> - return -1;
> + sizes[i] = pi->blk[phys_blk].end - pi->blk[phys_blk].start -
> + mem_hole_size(pi->blk[phys_blk].start, pi->blk[phys_blk].end);
> + sizes[i] = PFN_PHYS((unsigned long)(sizes[i] >> PAGE_SHIFT) /
> + nr_nodes * physnodes_with_mem);
> +
> + /*
> + * Calculate the number of big nodes that can be allocated as a result
> + * of consolidating the remainder.
> + */
> + bigs[i] = ((sizes[i] & ~FAKE_NODE_MIN_HASH_MASK) * nr_nodes) / physnodes_with_mem /
> + FAKE_NODE_MIN_SIZE;
> + sizes[i] &= FAKE_NODE_MIN_HASH_MASK;
> + if (!sizes[i]) {
> + pr_err("Not enough memory for each node inside physical numa node %d. NUMA emulation disabled.\n",
> + i);
> + return -1;
> + }
> }
>
> /*
> @@ -138,16 +150,16 @@ static int __init split_nodes_interleave(struct numa_meminfo *ei,
> }
> start = pi->blk[phys_blk].start;
> limit = pi->blk[phys_blk].end;
> - end = start + size;
> + end = start + sizes[i];
>
> - if (nid < big)
> + if (nid < bigs[i])
> end += FAKE_NODE_MIN_SIZE;
>
> /*
> * Continue to add memory to this fake node if its
> * non-reserved memory is less than the per-node size.
> */
> - while (end - start - mem_hole_size(start, end) < size) {
> + while (end - start - mem_hole_size(start, end) < sizes[i]) {
> end += FAKE_NODE_MIN_SIZE;
> if (end > limit) {
> end = limit;
> @@ -169,7 +181,7 @@ static int __init split_nodes_interleave(struct numa_meminfo *ei,
> * next node, this one must extend to the end of the
> * physical node.
> */
> - if (limit - end - mem_hole_size(end, limit) < size)
> + if (limit - end - mem_hole_size(end, limit) < sizes[i])
> end = limit;
>
> ret = emu_setup_memblk(ei, pi, nid++ % nr_nodes,
> @@ -432,7 +444,7 @@ void __init numa_emulation(struct numa_meminfo *numa_meminfo, int numa_dist_cnt)
> unsigned long n;
>
> n = simple_strtoul(emu_cmdline, &emu_cmdline, 0);
> - ret = split_nodes_interleave(&ei, &pi, 0, max_addr, n);
> + ret = split_nodes_interleave(&ei, &pi, n);
> }
> if (*emu_cmdline == ':')
> emu_cmdline++;
> --
> 2.34.1
>
--
Sincerely yours,
Mike.
Powered by blists - more mailing lists