From: Fiona Ebner <f.ebner@proxmox.com>
To: pve-devel@lists.proxmox.com
Subject: [PATCH qemu-server 4/4] cpu config: warn that OVMF might limit phys-bits to 40 without pdpe1gb CPU flag
Date: Mon, 23 Mar 2026 14:27:24 +0100 [thread overview]
Message-ID: <20260323132857.125747-5-f.ebner@proxmox.com> (raw)
In-Reply-To: <20260323132857.125747-1-f.ebner@proxmox.com>
OVMF limits phys-bits to 40 if no 1 GiB pages are available. When
built in debug mode, OVMF logs a debug message about this. But
otherwise, there is no user-visible log of this happening.
In practice, VMs with a lot of memory or VMs with GPU passthrough with
a lot of vRAM might not boot when the pdpe1gb CPU flag is not present.
In particular, this is the case for the built-in x86-64-vX models that
are based on the qemu64 model as well as the backend default kvm64.
It can't be expected that Proxmox VE users check the EDK II source
code to figure this out, so log a warning. The warning is limited to
CPU models based on the qemu64 and kvm64 models, since checking for
other models would require the CPU model expansion which is currently
not cheap to get. This should cover most problematic scenarios.
Signed-off-by: Fiona Ebner <f.ebner@proxmox.com>
---
src/PVE/QemuServer/CPUConfig.pm | 26 ++++++++++++++++++++++++++
1 file changed, 26 insertions(+)
diff --git a/src/PVE/QemuServer/CPUConfig.pm b/src/PVE/QemuServer/CPUConfig.pm
index 8da9010e..fb9af277 100644
--- a/src/PVE/QemuServer/CPUConfig.pm
+++ b/src/PVE/QemuServer/CPUConfig.pm
@@ -879,6 +879,21 @@ sub parse_cpuflag_list {
return $res;
}
+my sub check_phys_bits_above_40_compat {
+ my ($bios, $cpu_type, $cpu_flags) = @_;
+
+ # Would need to check CPU model expansion for others, but that information is not cheap to get
+ # right now. Checking with 'qemu64' and 'kvm64' should cover most problematic scenarios.
+ return if $cpu_type ne 'qemu64' && $cpu_type ne 'kvm64';
+
+ return if !$bios || $bios ne 'ovmf';
+
+ if (!$cpu_flags->{pdpe1gb} || $cpu_flags->{pdpe1gb}->{op} eq '-') {
+ log_warn("OVMF firmware might limit CPU 'phys-bits' to 40"
+ . " - enable the 'pdpe1gb' CPU flag to avoid this");
+ }
+}
+
# Calculate QEMU's '-cpu' argument from a given VM configuration
sub get_cpu_options {
my ($conf, $arch, $kvm, $kvm_off, $machine_version, $winversion, $gpu_passthrough) = @_;
@@ -973,6 +988,8 @@ sub get_cpu_options {
);
$cpu_str .= print_cpu_flags($resolved_flags);
+ my $using_phys_bits_above_40;
+
for my $phys_bits_opt (qw(guest-phys-bits phys-bits)) {
my $phys_bits = '';
for my $cpu_conf ($custom_cpu, $cpu) {
@@ -982,14 +999,23 @@ sub get_cpu_options {
if ($conf_val eq 'host') {
die "unexpected value 'host' for guest-phys-bits"
if $phys_bits_opt eq 'guest-phys-bits';
+
$phys_bits = ",host-phys-bits=true";
+
+ my $host_phys_bits = PVE::QemuServer::Helpers::get_host_phys_address_bits();
+ $using_phys_bits_above_40 = 1 if defined($host_phys_bits) && $host_phys_bits > 40;
} else {
$phys_bits = ",${phys_bits_opt}=${conf_val}";
+
+ $using_phys_bits_above_40 = 1 if $phys_bits_opt eq 'phys-bits' && $conf_val > 40;
}
}
$cpu_str .= $phys_bits;
}
+ check_phys_bits_above_40_compat($conf->{bios}, $cputype, $resolved_flags)
+ if $using_phys_bits_above_40;
+
return ('-cpu', $cpu_str);
}
--
2.47.3
prev parent reply other threads:[~2026-03-23 13:28 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-03-23 13:27 [PATCH-SERIES qemu-server 0/4] " Fiona Ebner
2026-03-23 13:27 ` [PATCH qemu-server 1/4] cpu config: split resolve_cpu_flags() function Fiona Ebner
2026-03-23 13:27 ` [PATCH qemu-server 2/4] move get_host_phys_address_bits() to helpers module Fiona Ebner
2026-03-23 13:27 ` [PATCH qemu-server 3/4] cpu config: get cpu options: avoid shadowing $conf variable Fiona Ebner
2026-03-23 13:27 ` Fiona Ebner [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=20260323132857.125747-5-f.ebner@proxmox.com \
--to=f.ebner@proxmox.com \
--cc=pve-devel@lists.proxmox.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.