From: "“William Roche" <william.roche@oracle.com>
To: david@redhat.com, kvm@vger.kernel.org, qemu-devel@nongnu.org,
qemu-arm@nongnu.org
Cc: william.roche@oracle.com, peterx@redhat.com, pbonzini@redhat.com,
richard.henderson@linaro.org, philmd@linaro.org,
peter.maydell@linaro.org, mtosatti@redhat.com,
imammedo@redhat.com, eduardo@habkost.net,
marcel.apfelbaum@gmail.com, wangyanan55@huawei.com,
zhao1.liu@intel.com, joao.m.martins@oracle.com
Subject: [PATCH v2 5/7] hostmem: Factor out applying settings
Date: Thu, 7 Nov 2024 10:21:24 +0000 [thread overview]
Message-ID: <20241107102126.2183152-6-william.roche@oracle.com> (raw)
In-Reply-To: <20241107102126.2183152-1-william.roche@oracle.com>
From: David Hildenbrand <david@redhat.com>
We want to reuse the functionality when remapping or resizing RAM.
Signed-off-by: David Hildenbrand <david@redhat.com>
Signed-off-by: William Roche <william.roche@oracle.com>
---
backends/hostmem.c | 155 ++++++++++++++++++++++++---------------------
1 file changed, 82 insertions(+), 73 deletions(-)
diff --git a/backends/hostmem.c b/backends/hostmem.c
index 181446626a..bf85d716e5 100644
--- a/backends/hostmem.c
+++ b/backends/hostmem.c
@@ -36,6 +36,87 @@ QEMU_BUILD_BUG_ON(HOST_MEM_POLICY_BIND != MPOL_BIND);
QEMU_BUILD_BUG_ON(HOST_MEM_POLICY_INTERLEAVE != MPOL_INTERLEAVE);
#endif
+static void host_memory_backend_apply_settings(HostMemoryBackend *backend,
+ void *ptr, uint64_t size,
+ Error **errp)
+{
+ bool async = !phase_check(PHASE_LATE_BACKENDS_CREATED);
+
+ if (backend->merge) {
+ qemu_madvise(ptr, size, QEMU_MADV_MERGEABLE);
+ }
+ if (!backend->dump) {
+ qemu_madvise(ptr, size, QEMU_MADV_DONTDUMP);
+ }
+#ifdef CONFIG_NUMA
+ unsigned long lastbit = find_last_bit(backend->host_nodes, MAX_NODES);
+ /* lastbit == MAX_NODES means maxnode = 0 */
+ unsigned long maxnode = (lastbit + 1) % (MAX_NODES + 1);
+ /*
+ * Ensure policy won't be ignored in case memory is preallocated
+ * before mbind(). note: MPOL_MF_STRICT is ignored on hugepages so
+ * this doesn't catch hugepage case.
+ */
+ unsigned flags = MPOL_MF_STRICT | MPOL_MF_MOVE;
+ int mode = backend->policy;
+
+ /*
+ * Check for invalid host-nodes and policies and give more verbose
+ * error messages than mbind().
+ */
+ if (maxnode && backend->policy == MPOL_DEFAULT) {
+ error_setg(errp, "host-nodes must be empty for policy default,"
+ " or you should explicitly specify a policy other"
+ " than default");
+ return;
+ } else if (maxnode == 0 && backend->policy != MPOL_DEFAULT) {
+ error_setg(errp, "host-nodes must be set for policy %s",
+ HostMemPolicy_str(backend->policy));
+ return;
+ }
+
+ /*
+ * We can have up to MAX_NODES nodes, but we need to pass maxnode+1
+ * as argument to mbind() due to an old Linux bug (feature?) which
+ * cuts off the last specified node. This means backend->host_nodes
+ * must have MAX_NODES+1 bits available.
+ */
+ assert(sizeof(backend->host_nodes) >=
+ BITS_TO_LONGS(MAX_NODES + 1) * sizeof(unsigned long));
+ assert(maxnode <= MAX_NODES);
+
+#ifdef HAVE_NUMA_HAS_PREFERRED_MANY
+ if (mode == MPOL_PREFERRED && numa_has_preferred_many() > 0) {
+ /*
+ * Replace with MPOL_PREFERRED_MANY otherwise the mbind() below
+ * silently picks the first node.
+ */
+ mode = MPOL_PREFERRED_MANY;
+ }
+#endif
+
+ if (maxnode &&
+ mbind(ptr, size, mode, backend->host_nodes, maxnode + 1, flags)) {
+ if (backend->policy != MPOL_DEFAULT || errno != ENOSYS) {
+ error_setg_errno(errp, errno,
+ "cannot bind memory to host NUMA nodes");
+ return;
+ }
+ }
+#endif
+ /*
+ * Preallocate memory after the NUMA policy has been instantiated.
+ * This is necessary to guarantee memory is allocated with
+ * specified NUMA policy in place.
+ */
+ if (backend->prealloc &&
+ !qemu_prealloc_mem(memory_region_get_fd(&backend->mr),
+ ptr, size, backend->prealloc_threads,
+ backend->prealloc_context, async, errp)) {
+ return;
+ }
+}
+
char *
host_memory_backend_get_name(HostMemoryBackend *backend)
{
@@ -337,7 +418,6 @@ host_memory_backend_memory_complete(UserCreatable *uc, Error **errp)
void *ptr;
uint64_t sz;
size_t pagesize;
- bool async = !phase_check(PHASE_LATE_BACKENDS_CREATED);
if (!bc->alloc) {
return;
@@ -357,78 +437,7 @@ host_memory_backend_memory_complete(UserCreatable *uc, Error **errp)
return;
}
- if (backend->merge) {
- qemu_madvise(ptr, sz, QEMU_MADV_MERGEABLE);
- }
- if (!backend->dump) {
- qemu_madvise(ptr, sz, QEMU_MADV_DONTDUMP);
- }
-#ifdef CONFIG_NUMA
- unsigned long lastbit = find_last_bit(backend->host_nodes, MAX_NODES);
- /* lastbit == MAX_NODES means maxnode = 0 */
- unsigned long maxnode = (lastbit + 1) % (MAX_NODES + 1);
- /*
- * Ensure policy won't be ignored in case memory is preallocated
- * before mbind(). note: MPOL_MF_STRICT is ignored on hugepages so
- * this doesn't catch hugepage case.
- */
- unsigned flags = MPOL_MF_STRICT | MPOL_MF_MOVE;
- int mode = backend->policy;
-
- /* check for invalid host-nodes and policies and give more verbose
- * error messages than mbind(). */
- if (maxnode && backend->policy == MPOL_DEFAULT) {
- error_setg(errp, "host-nodes must be empty for policy default,"
- " or you should explicitly specify a policy other"
- " than default");
- return;
- } else if (maxnode == 0 && backend->policy != MPOL_DEFAULT) {
- error_setg(errp, "host-nodes must be set for policy %s",
- HostMemPolicy_str(backend->policy));
- return;
- }
-
- /*
- * We can have up to MAX_NODES nodes, but we need to pass maxnode+1
- * as argument to mbind() due to an old Linux bug (feature?) which
- * cuts off the last specified node. This means backend->host_nodes
- * must have MAX_NODES+1 bits available.
- */
- assert(sizeof(backend->host_nodes) >=
- BITS_TO_LONGS(MAX_NODES + 1) * sizeof(unsigned long));
- assert(maxnode <= MAX_NODES);
-
-#ifdef HAVE_NUMA_HAS_PREFERRED_MANY
- if (mode == MPOL_PREFERRED && numa_has_preferred_many() > 0) {
- /*
- * Replace with MPOL_PREFERRED_MANY otherwise the mbind() below
- * silently picks the first node.
- */
- mode = MPOL_PREFERRED_MANY;
- }
-#endif
-
- if (maxnode &&
- mbind(ptr, sz, mode, backend->host_nodes, maxnode + 1, flags)) {
- if (backend->policy != MPOL_DEFAULT || errno != ENOSYS) {
- error_setg_errno(errp, errno,
- "cannot bind memory to host NUMA nodes");
- return;
- }
- }
-#endif
- /*
- * Preallocate memory after the NUMA policy has been instantiated.
- * This is necessary to guarantee memory is allocated with
- * specified NUMA policy in place.
- */
- if (backend->prealloc && !qemu_prealloc_mem(memory_region_get_fd(&backend->mr),
- ptr, sz,
- backend->prealloc_threads,
- backend->prealloc_context,
- async, errp)) {
- return;
- }
+ host_memory_backend_apply_settings(backend, ptr, sz, errp);
}
static bool
--
2.43.5
next prev parent reply other threads:[~2024-11-07 10:23 UTC|newest]
Thread overview: 119+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-09-10 9:07 [RFC 0/6] hugetlbfs largepage RAS project “William Roche
2024-09-10 9:07 ` [RFC 1/6] accel/kvm: SIGBUS handler should also deal with si_addr_lsb “William Roche
2024-09-10 9:07 ` [RFC 2/6] accel/kvm: Keep track of the HWPoisonPage sizes “William Roche
2024-09-10 9:07 ` [RFC 3/6] system/physmem: Remap memory pages on reset based on the page size “William Roche
2024-09-10 9:07 ` [RFC 4/6] system: Introducing hugetlbfs largepage RAS feature “William Roche
2024-09-10 9:07 ` [RFC 5/6] system/hugetlb_ras: Handle madvise SIGBUS signal on listener “William Roche
2024-09-10 9:07 ` [RFC 6/6] system/hugetlb_ras: Replay lost BUS_MCEERR_AO signals on VM resume “William Roche
2024-09-10 10:02 ` [RFC RESEND 0/6] hugetlbfs largepage RAS project “William Roche
2024-09-10 10:02 ` [RFC RESEND 1/6] accel/kvm: SIGBUS handler should also deal with si_addr_lsb “William Roche
2024-09-10 10:02 ` [RFC RESEND 2/6] accel/kvm: Keep track of the HWPoisonPage sizes “William Roche
2024-09-10 10:02 ` [RFC RESEND 3/6] system/physmem: Remap memory pages on reset based on the page size “William Roche
2024-09-10 10:02 ` [RFC RESEND 4/6] system: Introducing hugetlbfs largepage RAS feature “William Roche
2024-09-10 10:02 ` [RFC RESEND 5/6] system/hugetlb_ras: Handle madvise SIGBUS signal on listener “William Roche
2024-09-10 10:02 ` [RFC RESEND 6/6] system/hugetlb_ras: Replay lost BUS_MCEERR_AO signals on VM resume “William Roche
2024-09-10 11:36 ` [RFC RESEND 0/6] hugetlbfs largepage RAS project David Hildenbrand
2024-09-10 16:24 ` William Roche
2024-09-11 22:07 ` David Hildenbrand
2024-09-12 17:07 ` William Roche
2024-09-19 16:52 ` William Roche
2024-10-09 15:45 ` Peter Xu
2024-10-10 20:35 ` William Roche
2024-10-22 21:34 ` [PATCH v1 0/4] hugetlbfs memory HW error fixes “William Roche
2024-10-22 21:35 ` [PATCH v1 1/4] accel/kvm: SIGBUS handler should also deal with si_addr_lsb “William Roche
2024-10-22 21:35 ` [PATCH v1 2/4] accel/kvm: Keep track of the HWPoisonPage page_size “William Roche
2024-10-23 7:28 ` David Hildenbrand
2024-10-25 23:27 ` William Roche
2024-10-28 16:42 ` David Hildenbrand
2024-10-30 1:56 ` William Roche
2024-11-04 14:10 ` David Hildenbrand
2024-10-25 23:30 ` William Roche
2024-10-22 21:35 ` [PATCH v1 3/4] system/physmem: Largepage punch hole before reset of memory pages “William Roche
2024-10-23 7:30 ` David Hildenbrand
2024-10-25 23:27 ` William Roche
2024-10-28 17:01 ` David Hildenbrand
2024-10-30 1:56 ` William Roche
2024-11-04 13:30 ` David Hildenbrand
2024-11-07 10:21 ` [PATCH v2 0/7] hugetlbfs memory HW error fixes “William Roche
2024-11-07 10:21 ` [PATCH v2 1/7] accel/kvm: Keep track of the HWPoisonPage page_size “William Roche
2024-11-12 10:30 ` David Hildenbrand
2024-11-12 18:17 ` William Roche
2024-11-12 21:35 ` David Hildenbrand
2024-11-07 10:21 ` [PATCH v2 2/7] system/physmem: poisoned memory discard on reboot “William Roche
2024-11-12 11:07 ` David Hildenbrand
2024-11-12 18:17 ` William Roche
2024-11-12 22:06 ` David Hildenbrand
2024-11-07 10:21 ` [PATCH v2 3/7] accel/kvm: Report the loss of a large memory page “William Roche
2024-11-12 11:13 ` David Hildenbrand
2024-11-12 18:17 ` William Roche
2024-11-12 22:22 ` David Hildenbrand
2024-11-15 21:03 ` William Roche
2024-11-18 9:45 ` David Hildenbrand
2024-11-07 10:21 ` [PATCH v2 4/7] numa: Introduce and use ram_block_notify_remap() “William Roche
2024-11-07 10:21 ` “William Roche [this message]
2024-11-07 10:21 ` [PATCH v2 6/7] hostmem: Handle remapping of RAM “William Roche
2024-11-12 13:45 ` David Hildenbrand
2024-11-12 18:17 ` William Roche
2024-11-12 22:24 ` David Hildenbrand
2024-11-07 10:21 ` [PATCH v2 7/7] system/physmem: Memory settings applied on remap notification “William Roche
2024-10-22 21:35 ` [PATCH v1 4/4] accel/kvm: Report the loss of a large memory page “William Roche
2024-10-28 16:32 ` [RFC RESEND 0/6] hugetlbfs largepage RAS project David Hildenbrand
2024-11-25 14:27 ` [PATCH v3 0/7] hugetlbfs memory HW error fixes “William Roche
2024-11-25 14:27 ` [PATCH v3 1/7] hwpoison_page_list and qemu_ram_remap are based of pages “William Roche
2024-11-25 14:27 ` [PATCH v3 2/7] system/physmem: poisoned memory discard on reboot “William Roche
2024-11-25 14:27 ` [PATCH v3 3/7] accel/kvm: Report the loss of a large memory page “William Roche
2024-11-25 14:27 ` [PATCH v3 4/7] numa: Introduce and use ram_block_notify_remap() “William Roche
2024-11-25 14:27 ` [PATCH v3 5/7] hostmem: Factor out applying settings “William Roche
2024-11-25 14:27 ` [PATCH v3 6/7] hostmem: Handle remapping of RAM “William Roche
2024-11-25 14:27 ` [PATCH v3 7/7] system/physmem: Memory settings applied on remap notification “William Roche
2024-12-02 15:41 ` [PATCH v3 0/7] hugetlbfs memory HW error fixes William Roche
2024-12-02 16:00 ` David Hildenbrand
2024-12-03 0:15 ` William Roche
2024-12-03 14:08 ` David Hildenbrand
2024-12-03 14:39 ` William Roche
2024-12-03 15:00 ` David Hildenbrand
2024-12-06 18:26 ` William Roche
2024-12-09 21:25 ` David Hildenbrand
2024-12-14 13:45 ` [PATCH v4 0/7] Poisoned memory recovery on reboot “William Roche
2024-12-14 13:45 ` [PATCH v4 1/7] hwpoison_page_list and qemu_ram_remap are based on pages “William Roche
2025-01-08 21:34 ` David Hildenbrand
2025-01-10 20:56 ` William Roche
2025-01-14 13:56 ` David Hildenbrand
2024-12-14 13:45 ` [PATCH v4 2/7] system/physmem: poisoned memory discard on reboot “William Roche
2025-01-08 21:44 ` David Hildenbrand
2025-01-10 20:56 ` William Roche
2025-01-14 14:00 ` David Hildenbrand
2025-01-27 21:15 ` William Roche
2024-12-14 13:45 ` [PATCH v4 3/7] accel/kvm: Report the loss of a large memory page “William Roche
2024-12-14 13:45 ` [PATCH v4 4/7] numa: Introduce and use ram_block_notify_remap() “William Roche
2024-12-14 13:45 ` [PATCH v4 5/7] hostmem: Factor out applying settings “William Roche
2025-01-08 21:58 ` David Hildenbrand
2025-01-10 20:56 ` William Roche
2024-12-14 13:45 ` [PATCH v4 6/7] hostmem: Handle remapping of RAM “William Roche
2025-01-08 21:51 ` [PATCH v4 6/7] c David Hildenbrand
2025-01-10 20:57 ` [PATCH v4 6/7] hostmem: Handle remapping of RAM William Roche
2024-12-14 13:45 ` [PATCH v4 7/7] system/physmem: Memory settings applied on remap notification “William Roche
2025-01-08 21:53 ` David Hildenbrand
2025-01-10 20:57 ` William Roche
2025-01-14 14:01 ` David Hildenbrand
2025-01-08 21:22 ` [PATCH v4 0/7] Poisoned memory recovery on reboot David Hildenbrand
2025-01-10 20:55 ` William Roche
2025-01-10 21:13 ` [PATCH v5 0/6] " “William Roche
2025-01-10 21:14 ` [PATCH v5 1/6] system/physmem: handle hugetlb correctly in qemu_ram_remap() “William Roche
2025-01-14 14:02 ` David Hildenbrand
2025-01-27 21:16 ` William Roche
2025-01-28 18:41 ` David Hildenbrand
2025-01-10 21:14 ` [PATCH v5 2/6] system/physmem: poisoned memory discard on reboot “William Roche
2025-01-14 14:07 ` David Hildenbrand
2025-01-27 21:16 ` William Roche
2025-01-10 21:14 ` [PATCH v5 3/6] accel/kvm: Report the loss of a large memory page “William Roche
2025-01-14 14:09 ` David Hildenbrand
2025-01-27 21:16 ` William Roche
2025-01-28 18:45 ` David Hildenbrand
2025-01-10 21:14 ` [PATCH v5 4/6] numa: Introduce and use ram_block_notify_remap() “William Roche
2025-01-10 21:14 ` [PATCH v5 5/6] hostmem: Factor out applying settings “William Roche
2025-01-10 21:14 ` [PATCH v5 6/6] hostmem: Handle remapping of RAM “William Roche
2025-01-14 14:11 ` David Hildenbrand
2025-01-27 21:16 ` William Roche
2025-01-14 14:12 ` [PATCH v5 0/6] Poisoned memory recovery on reboot David Hildenbrand
2025-01-27 21:16 ` William Roche
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=20241107102126.2183152-6-william.roche@oracle.com \
--to=william.roche@oracle.com \
--cc=david@redhat.com \
--cc=eduardo@habkost.net \
--cc=imammedo@redhat.com \
--cc=joao.m.martins@oracle.com \
--cc=kvm@vger.kernel.org \
--cc=marcel.apfelbaum@gmail.com \
--cc=mtosatti@redhat.com \
--cc=pbonzini@redhat.com \
--cc=peter.maydell@linaro.org \
--cc=peterx@redhat.com \
--cc=philmd@linaro.org \
--cc=qemu-arm@nongnu.org \
--cc=qemu-devel@nongnu.org \
--cc=richard.henderson@linaro.org \
--cc=wangyanan55@huawei.com \
--cc=zhao1.liu@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).