From: Mike Rapoport <rppt@kernel.org>
To: priyanshukumarpu@gmail.com
Cc: akpm@linux-foundation.org, changyuanl@google.com,
linux-mm@kvack.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH v2] tools/testing/memblock: fix stale NUMA reservation tests
Date: Wed, 15 Apr 2026 19:16:44 +0300 [thread overview]
Message-ID: <ad-57Iq8yat01OiN@kernel.org> (raw)
In-Reply-To: <20260415122731.1768912-1-priyanshukumarpu@gmail.com>
Hi,
Please next time send v2 as a different mail rather than replying to v1 and
add description of the changes between the versions:
https://docs.kernel.org/process/submitting-patches.html#commentary
On Wed, Apr 15, 2026 at 12:27:31PM +0000, priyanshukumarpu@gmail.com wrote:
> From: Priyanshu Kumar <priyanshukumarpu@gmail.com>
>
> memblock allocations now reserve memory with MEMBLOCK_RSRV_KERN and,
> on NUMA configurations, record the requested node on the reserved
> region. Several memblock simulator NUMA tests still expected merges
> that only worked before those reservation semantics changed, so the
> suite aborted even though the allocator behavior was correct.
>
> Update the NUMA merge expectations in the memblock_alloc_try_nid()
> and memblock_alloc_exact_nid_raw() tests to match the current reserved
> region metadata rules. For cases that should still merge, create the
> pre-existing reservation with matching nid and MEMBLOCK_RSRV_KERN
> metadata. Also strengthen the memblock_alloc_node() coverage by
> checking the newly created reserved region directly instead of
> re-reading the source memory node descriptor.
>
> Finally, drop the stale README/TODO notes that still claimed
> memblock_alloc_node() could not be tested.
>
> The memblock simulator passes again with NUMA enabled after these
> updates.
>
> Signed-off-by: Priyanshu Kumar <priyanshukumarpu@gmail.com>
> ---
> tools/testing/memblock/README | 5 +----
> tools/testing/memblock/TODO | 4 ++--
> .../memblock/tests/alloc_exact_nid_api.c | 6 +++---
> tools/testing/memblock/tests/alloc_nid_api.c | 21 +++++++++++++------
> 4 files changed, 21 insertions(+), 15 deletions(-)
>
> diff --git a/tools/testing/memblock/README b/tools/testing/memblock/README
> index 7ca437d81806..b435f48d8a70 100644
> --- a/tools/testing/memblock/README
> +++ b/tools/testing/memblock/README
> @@ -104,10 +104,7 @@ called at the beginning of each test.
> Known issues
> ============
>
> -1. Requesting a specific NUMA node via memblock_alloc_node() does not work as
> - intended. Once the fix is in place, tests for this function can be added.
> -
> -2. Tests for memblock_alloc_low() can't be easily implemented. The function uses
> +1. Tests for memblock_alloc_low() can't be easily implemented. The function uses
> ARCH_LOW_ADDRESS_LIMIT marco, which can't be changed to point at the low
> memory of the memory_block.
>
> diff --git a/tools/testing/memblock/TODO b/tools/testing/memblock/TODO
> index e306c90c535f..c13ad0dae776 100644
> --- a/tools/testing/memblock/TODO
> +++ b/tools/testing/memblock/TODO
> @@ -1,5 +1,5 @@
> TODO
> =====
>
> -1. Add tests for memblock_alloc_node() to check if the correct NUMA node is set
> - for the new region
> +1. Add tests for memblock_alloc_low() once the simulator can model
> + ARCH_LOW_ADDRESS_LIMIT against the low memory in memory_block
> diff --git a/tools/testing/memblock/tests/alloc_exact_nid_api.c b/tools/testing/memblock/tests/alloc_exact_nid_api.c
> index 6e14447da6e1..0c46c73b5e04 100644
> --- a/tools/testing/memblock/tests/alloc_exact_nid_api.c
> +++ b/tools/testing/memblock/tests/alloc_exact_nid_api.c
> @@ -368,7 +368,7 @@ static int alloc_exact_nid_bottom_up_numa_part_reserved_check(void)
> max_addr = memblock_end_of_DRAM();
> total_size = size + r1.size;
>
> - memblock_reserve(r1.base, r1.size);
> + __memblock_reserve(r1.base, r1.size, nid_req, MEMBLOCK_RSRV_KERN);
> allocated_ptr = memblock_alloc_exact_nid_raw(size, SMP_CACHE_BYTES,
> min_addr, max_addr,
> nid_req);
> @@ -861,8 +861,8 @@ static int alloc_exact_nid_numa_reserved_full_merge_generic_check(void)
> min_addr = r2.base + r2.size;
> max_addr = r1.base;
>
> - memblock_reserve(r1.base, r1.size);
> - memblock_reserve(r2.base, r2.size);
> + __memblock_reserve(r1.base, r1.size, nid_req, MEMBLOCK_RSRV_KERN);
> + __memblock_reserve(r2.base, r2.size, nid_req, MEMBLOCK_RSRV_KERN);
>
> allocated_ptr = memblock_alloc_exact_nid_raw(size, SMP_CACHE_BYTES,
> min_addr, max_addr,
> diff --git a/tools/testing/memblock/tests/alloc_nid_api.c b/tools/testing/memblock/tests/alloc_nid_api.c
> index 562e4701b0e0..c23652727976 100644
> --- a/tools/testing/memblock/tests/alloc_nid_api.c
> +++ b/tools/testing/memblock/tests/alloc_nid_api.c
> @@ -1965,7 +1965,7 @@ static int alloc_nid_bottom_up_numa_part_reserved_check(void)
> max_addr = memblock_end_of_DRAM();
> total_size = size + r1.size;
>
> - memblock_reserve(r1.base, r1.size);
> + __memblock_reserve(r1.base, r1.size, nid_req, MEMBLOCK_RSRV_KERN);
> allocated_ptr = run_memblock_alloc_nid(size, SMP_CACHE_BYTES,
> min_addr, max_addr, nid_req);
>
> @@ -2412,8 +2412,8 @@ static int alloc_nid_numa_reserved_full_merge_generic_check(void)
> min_addr = r2.base + r2.size;
> max_addr = r1.base;
>
> - memblock_reserve(r1.base, r1.size);
> - memblock_reserve(r2.base, r2.size);
> + __memblock_reserve(r1.base, r1.size, nid_req, MEMBLOCK_RSRV_KERN);
> + __memblock_reserve(r2.base, r2.size, nid_req, MEMBLOCK_RSRV_KERN);
>
> allocated_ptr = run_memblock_alloc_nid(size, SMP_CACHE_BYTES,
> min_addr, max_addr, nid_req);
> @@ -2496,15 +2496,18 @@ static int alloc_nid_numa_split_all_reserved_generic_check(void)
>
> /*
> * A simple test that tries to allocate a memory region through the
> - * memblock_alloc_node() on a NUMA node with id `nid`. Expected to have the
> - * correct NUMA node set for the new region.
> + * memblock_alloc_node() on a NUMA node with id `nid`. Expected to allocate
> + * the region within the requested node and mark the new reservation with the
> + * correct NUMA node.
This change is not related to the fix and I think it's adding too much
noise to the test in any case.
> */
> static int alloc_node_on_correct_nid(void)
> {
> int nid_req = 2;
> void *allocated_ptr = NULL;
> #ifdef CONFIG_NUMA
> + struct memblock_region *new_rgn = &memblock.reserved.regions[0];
> struct memblock_region *req_node = &memblock.memory.regions[nid_req];
> + phys_addr_t req_node_end = region_end(req_node);
> #endif
> phys_addr_t size = SZ_512;
>
> @@ -2515,7 +2518,13 @@ static int alloc_node_on_correct_nid(void)
>
> ASSERT_NE(allocated_ptr, NULL);
> #ifdef CONFIG_NUMA
> - ASSERT_EQ(nid_req, req_node->nid);
> + ASSERT_EQ(memblock.reserved.cnt, 1);
> + ASSERT_EQ(new_rgn->size, size);
> + ASSERT_EQ(new_rgn->base, (phys_addr_t)allocated_ptr);
> + ASSERT_EQ(new_rgn->flags, MEMBLOCK_RSRV_KERN);
> + ASSERT_EQ(nid_req, memblock_get_region_node(new_rgn));
> + ASSERT_LE(req_node->base, new_rgn->base);
> + ASSERT_LE(region_end(new_rgn), req_node_end);
> #endif
>
> test_pass_pop();
> --
> 2.43.0
>
--
Sincerely yours,
Mike.
prev parent reply other threads:[~2026-04-15 16:16 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-04-13 9:14 [PATCH] " priyanshukumarpu
2026-04-14 15:14 ` Mike Rapoport
2026-04-15 12:27 ` [PATCH v2] " priyanshukumarpu
2026-04-15 16:16 ` Mike Rapoport [this message]
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=ad-57Iq8yat01OiN@kernel.org \
--to=rppt@kernel.org \
--cc=akpm@linux-foundation.org \
--cc=changyuanl@google.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=priyanshukumarpu@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