qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: "“William Roche" <william.roche@oracle.com>
To: pbonzini@redhat.com, peterx@redhat.com, david@redhat.com,
	philmd@linaro.org, marcandre.lureau@redhat.com,
	berrange@redhat.com, thuth@redhat.com,
	richard.henderson@linaro.org, peter.maydell@linaro.org,
	mtosatti@redhat.com, qemu-devel@nongnu.org
Cc: kvm@vger.kernel.org, qemu-arm@nongnu.org,
	william.roche@oracle.com, joao.m.martins@oracle.com
Subject: [RFC RESEND 3/6] system/physmem: Remap memory pages on reset based on the page size
Date: Tue, 10 Sep 2024 10:02:13 +0000	[thread overview]
Message-ID: <20240910100216.2744078-4-william.roche@oracle.com> (raw)
In-Reply-To: <20240910100216.2744078-1-william.roche@oracle.com>

From: William Roche <william.roche@oracle.com>

When the VM reboots, a memory reset is performed calling
qemu_ram_remap() on all hwpoisoned pages.
We take into account the recorded page size to adjust the
size and location of the memory hole.
In case of a largepage used, we also need to punch a hole
in the backend file to regenerate a usable memory, cleaning
the HW poisoned section.

Signed-off-by: William Roche <william.roche@oracle.com>
---
 system/physmem.c | 13 +++++++++++++
 1 file changed, 13 insertions(+)

diff --git a/system/physmem.c b/system/physmem.c
index 94600a33ec..5c176146c0 100644
--- a/system/physmem.c
+++ b/system/physmem.c
@@ -2195,6 +2195,11 @@ void qemu_ram_remap(ram_addr_t addr, ram_addr_t length)
     RAMBLOCK_FOREACH(block) {
         offset = addr - block->offset;
         if (offset < block->max_length) {
+            /* addr could be a subpage of a former large page */
+            if (length < block->page_size) {
+                offset = ROUND_DOWN(offset, block->page_size);
+                length = block->page_size;
+            }
             vaddr = ramblock_ptr(block, offset);
             if (block->flags & RAM_PREALLOC) {
                 ;
@@ -2208,6 +2213,14 @@ void qemu_ram_remap(ram_addr_t addr, ram_addr_t length)
                 prot = PROT_READ;
                 prot |= block->flags & RAM_READONLY ? 0 : PROT_WRITE;
                 if (block->fd >= 0) {
+                    if (length > TARGET_PAGE_SIZE && fallocate(block->fd,
+                        FALLOC_FL_KEEP_SIZE | FALLOC_FL_PUNCH_HOLE,
+                        offset + block->fd_offset, length) != 0) {
+                        error_report("Could not recreate the file hole for "
+                                     "addr: " RAM_ADDR_FMT "@" RAM_ADDR_FMT "",
+                                     length, addr);
+                        exit(1);
+                    }
                     area = mmap(vaddr, length, prot, flags, block->fd,
                                 offset + block->fd_offset);
                 } else {
-- 
2.43.5



  parent reply	other threads:[~2024-09-10 10:03 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   ` “William Roche [this message]
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                               ` [PATCH v2 5/7] hostmem: Factor out applying settings “William Roche
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=20240910100216.2744078-4-william.roche@oracle.com \
    --to=william.roche@oracle.com \
    --cc=berrange@redhat.com \
    --cc=david@redhat.com \
    --cc=joao.m.martins@oracle.com \
    --cc=kvm@vger.kernel.org \
    --cc=marcandre.lureau@redhat.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=thuth@redhat.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).