* [PATCH v7 1/1] acpi/virt: suppress SPCR when UART0 has no serial backend
@ 2026-02-19 13:04 Li Chen
2026-02-19 13:42 ` Igor Mammedov
2026-02-20 14:50 ` Michael S. Tsirkin
0 siblings, 2 replies; 4+ messages in thread
From: Li Chen @ 2026-02-19 13:04 UTC (permalink / raw)
To: qemu-devel
Cc: qemu-arm, qemu-riscv, Michael S . Tsirkin, Igor Mammedov,
Ani Sinha, Peter Maydell, Shannon Zhao, Song Gao, Bibo Mao,
Jiaxun Yang, Sunil V L, Palmer Dabbelt, Alistair Francis,
Weiwei Li, Daniel Henrique Barboza, Liu Zhiwei,
Philippe Mathieu-Daudé
From: Li Chen <chenl311@chinatelecom.cn>
virt machines always instantiate a PL011/16550 UART at slot 0 and describe
it in ACPI (DSDT and optional SPCR table). When the command line disables
the serial backend (e.g. "-serial none"), SPCR still points at UART0 and
the guest may pick it as the preferred console even though it is not
usable.
Only generate SPCR when UART0 has a chardev backend attached. Keep the
UART device description in DSDT unchanged so the guest-visible UART
topology does not change.
The bios-tables-test qtests use "-serial null" to keep SPCR enabled while
discarding firmware serial output, avoiding TAP stdout corruption.
Suggested-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Li Chen <chenl311@chinatelecom.cn>
Reviewed-by: Sunil V L <sunilvl@ventanamicro.com>
---
v6 -> v7:
Do not gate DSDT UART devices on serial_hd(0); only gate SPCR.
Drop the serial_exists() helper added by v6.
Keep bios-tables-test using "-serial null" (comment wording update).
v5 -> v6:
Fix TAP parsing errors by using "-serial null" in bios-tables-test.
v4 -> v5:
Also suppress UART device & SPCR when guest has no serial hardware on
loongarch.
Rename serial_exist to serial_exists.
Fix style issues.
hw/arm/virt-acpi-build.c | 3 ++-
hw/loongarch/virt-acpi-build.c | 2 +-
hw/riscv/virt-acpi-build.c | 3 ++-
tests/qtest/bios-tables-test.c | 7 +++++--
4 files changed, 10 insertions(+), 5 deletions(-)
diff --git a/hw/arm/virt-acpi-build.c b/hw/arm/virt-acpi-build.c
index c145678185..1411f98ea1 100644
--- a/hw/arm/virt-acpi-build.c
+++ b/hw/arm/virt-acpi-build.c
@@ -64,6 +64,7 @@
#include "hw/virtio/virtio-acpi.h"
#include "target/arm/cpu.h"
#include "target/arm/multiprocessing.h"
+#include "system/system.h"
#define ARM_SPI_BASE 32
@@ -1291,7 +1292,7 @@ void virt_acpi_build(VirtMachineState *vms, AcpiBuildTables *tables)
acpi_add_table(table_offsets, tables_blob);
- if (ms->acpi_spcr_enabled) {
+ if (ms->acpi_spcr_enabled && serial_hd(0)) {
spcr_setup(tables_blob, tables->linker, vms);
}
diff --git a/hw/loongarch/virt-acpi-build.c b/hw/loongarch/virt-acpi-build.c
index 54ac94e17d..dc8bbb9fdc 100644
--- a/hw/loongarch/virt-acpi-build.c
+++ b/hw/loongarch/virt-acpi-build.c
@@ -546,7 +546,7 @@ static void acpi_build(AcpiBuildTables *tables, MachineState *machine)
build_srat(tables_blob, tables->linker, machine);
acpi_add_table(table_offsets, tables_blob);
- if (machine->acpi_spcr_enabled)
+ if (machine->acpi_spcr_enabled && serial_hd(0))
spcr_setup(tables_blob, tables->linker, machine);
if (machine->numa_state->num_nodes) {
diff --git a/hw/riscv/virt-acpi-build.c b/hw/riscv/virt-acpi-build.c
index f1406cb683..29080d1125 100644
--- a/hw/riscv/virt-acpi-build.c
+++ b/hw/riscv/virt-acpi-build.c
@@ -39,6 +39,7 @@
#include "qapi/error.h"
#include "qemu/error-report.h"
#include "system/reset.h"
+#include "system/system.h"
#define ACPI_BUILD_TABLE_SIZE 0x20000
#define ACPI_BUILD_INTC_ID(socket, index) ((socket << 24) | (index))
@@ -890,7 +891,7 @@ static void virt_acpi_build(RISCVVirtState *s, AcpiBuildTables *tables)
acpi_add_table(table_offsets, tables_blob);
- if (ms->acpi_spcr_enabled) {
+ if (ms->acpi_spcr_enabled && serial_hd(0)) {
spcr_setup(tables_blob, tables->linker, s);
}
diff --git a/tests/qtest/bios-tables-test.c b/tests/qtest/bios-tables-test.c
index e489d94331..37f2b72d83 100644
--- a/tests/qtest/bios-tables-test.c
+++ b/tests/qtest/bios-tables-test.c
@@ -825,10 +825,13 @@ static char *test_acpi_create_args(test_data *data, const char *params)
/*
* TODO: convert '-drive if=pflash' to new syntax (see e33763be7cd3)
* when arm/virt boad starts to support it.
+ * NOTE: Explicitly add "-serial null" to keep SPCR enabled while
+ * discarding firmware serial output (avoid corrupting TAP
+ * stdout).
*/
if (data->cd) {
args = g_strdup_printf("-machine %s%s %s -accel tcg "
- "-nodefaults -nographic "
+ "-nodefaults -serial null -nographic "
"-drive if=pflash,format=raw,file=%s,readonly=on "
"-drive if=pflash,format=raw,file=%s,snapshot=on -cdrom %s %s",
data->machine, data->machine_param ?: "",
@@ -836,7 +839,7 @@ static char *test_acpi_create_args(test_data *data, const char *params)
data->uefi_fl1, data->uefi_fl2, data->cd, params ? params : "");
} else {
args = g_strdup_printf("-machine %s%s %s -accel tcg "
- "-nodefaults -nographic "
+ "-nodefaults -serial null -nographic "
"-drive if=pflash,format=raw,file=%s,readonly=on "
"-drive if=pflash,format=raw,file=%s,snapshot=on %s",
data->machine, data->machine_param ?: "",
--
2.52.0
^ permalink raw reply related [flat|nested] 4+ messages in thread
* Re: [PATCH v7 1/1] acpi/virt: suppress SPCR when UART0 has no serial backend
2026-02-19 13:04 [PATCH v7 1/1] acpi/virt: suppress SPCR when UART0 has no serial backend Li Chen
@ 2026-02-19 13:42 ` Igor Mammedov
2026-02-20 14:50 ` Michael S. Tsirkin
1 sibling, 0 replies; 4+ messages in thread
From: Igor Mammedov @ 2026-02-19 13:42 UTC (permalink / raw)
To: Li Chen
Cc: qemu-devel, qemu-arm, qemu-riscv, Michael S . Tsirkin, Ani Sinha,
Peter Maydell, Shannon Zhao, Song Gao, Bibo Mao, Jiaxun Yang,
Sunil V L, Palmer Dabbelt, Alistair Francis, Weiwei Li,
Daniel Henrique Barboza, Liu Zhiwei, Philippe Mathieu-Daudé
On Thu, 19 Feb 2026 21:04:02 +0800
Li Chen <me@linux.beauty> wrote:
> From: Li Chen <chenl311@chinatelecom.cn>
>
> virt machines always instantiate a PL011/16550 UART at slot 0 and describe
> it in ACPI (DSDT and optional SPCR table). When the command line disables
> the serial backend (e.g. "-serial none"), SPCR still points at UART0 and
> the guest may pick it as the preferred console even though it is not
> usable.
>
> Only generate SPCR when UART0 has a chardev backend attached. Keep the
> UART device description in DSDT unchanged so the guest-visible UART
> topology does not change.
>
I've though consensus was, if device (frontend) exists,
we should keep it described in ACPI (regardless of backend state).
> The bios-tables-test qtests use "-serial null" to keep SPCR enabled while
> discarding firmware serial output, avoiding TAP stdout corruption.
It's tests issue, I'd say fix it there.
See for more below.
> Suggested-by: Philippe Mathieu-Daudé <philmd@linaro.org>
> Signed-off-by: Li Chen <chenl311@chinatelecom.cn>
> Reviewed-by: Sunil V L <sunilvl@ventanamicro.com>
> ---
> v6 -> v7:
> Do not gate DSDT UART devices on serial_hd(0); only gate SPCR.
> Drop the serial_exists() helper added by v6.
> Keep bios-tables-test using "-serial null" (comment wording update).
>
> v5 -> v6:
> Fix TAP parsing errors by using "-serial null" in bios-tables-test.
>
> v4 -> v5:
> Also suppress UART device & SPCR when guest has no serial hardware on
> loongarch.
> Rename serial_exist to serial_exists.
> Fix style issues.
>
> hw/arm/virt-acpi-build.c | 3 ++-
> hw/loongarch/virt-acpi-build.c | 2 +-
> hw/riscv/virt-acpi-build.c | 3 ++-
> tests/qtest/bios-tables-test.c | 7 +++++--
> 4 files changed, 10 insertions(+), 5 deletions(-)
>
> diff --git a/hw/arm/virt-acpi-build.c b/hw/arm/virt-acpi-build.c
> index c145678185..1411f98ea1 100644
> --- a/hw/arm/virt-acpi-build.c
> +++ b/hw/arm/virt-acpi-build.c
> @@ -64,6 +64,7 @@
> #include "hw/virtio/virtio-acpi.h"
> #include "target/arm/cpu.h"
> #include "target/arm/multiprocessing.h"
> +#include "system/system.h"
>
> #define ARM_SPI_BASE 32
>
> @@ -1291,7 +1292,7 @@ void virt_acpi_build(VirtMachineState *vms, AcpiBuildTables *tables)
>
> acpi_add_table(table_offsets, tables_blob);
>
> - if (ms->acpi_spcr_enabled) {
> + if (ms->acpio_spcr_enabled && serial_hd(0)) {
if you really need to disable spcr,
add a knob to do so instead of gating it on backend state.
Having serial console configured and sent to /dev/null
is a legit config (ex: testing that writes to console do not explode machine,
but not really caring where writes end up)
> spcr_setup(tables_blob, tables->linker, vms);
> }
>
> diff --git a/hw/loongarch/virt-acpi-build.c b/hw/loongarch/virt-acpi-build.c
> index 54ac94e17d..dc8bbb9fdc 100644
> --- a/hw/loongarch/virt-acpi-build.c
> +++ b/hw/loongarch/virt-acpi-build.c
> @@ -546,7 +546,7 @@ static void acpi_build(AcpiBuildTables *tables, MachineState *machine)
> build_srat(tables_blob, tables->linker, machine);
> acpi_add_table(table_offsets, tables_blob);
>
> - if (machine->acpi_spcr_enabled)
> + if (machine->acpi_spcr_enabled && serial_hd(0))
> spcr_setup(tables_blob, tables->linker, machine);
>
> if (machine->numa_state->num_nodes) {
> diff --git a/hw/riscv/virt-acpi-build.c b/hw/riscv/virt-acpi-build.c
> index f1406cb683..29080d1125 100644
> --- a/hw/riscv/virt-acpi-build.c
> +++ b/hw/riscv/virt-acpi-build.c
> @@ -39,6 +39,7 @@
> #include "qapi/error.h"
> #include "qemu/error-report.h"
> #include "system/reset.h"
> +#include "system/system.h"
>
> #define ACPI_BUILD_TABLE_SIZE 0x20000
> #define ACPI_BUILD_INTC_ID(socket, index) ((socket << 24) | (index))
> @@ -890,7 +891,7 @@ static void virt_acpi_build(RISCVVirtState *s, AcpiBuildTables *tables)
>
> acpi_add_table(table_offsets, tables_blob);
>
> - if (ms->acpi_spcr_enabled) {
> + if (ms->acpi_spcr_enabled && serial_hd(0)) {
> spcr_setup(tables_blob, tables->linker, s);
> }
>
> diff --git a/tests/qtest/bios-tables-test.c b/tests/qtest/bios-tables-test.c
> index e489d94331..37f2b72d83 100644
> --- a/tests/qtest/bios-tables-test.c
> +++ b/tests/qtest/bios-tables-test.c
> @@ -825,10 +825,13 @@ static char *test_acpi_create_args(test_data *data, const char *params)
> /*
> * TODO: convert '-drive if=pflash' to new syntax (see e33763be7cd3)
> * when arm/virt boad starts to support it.
> + * NOTE: Explicitly add "-serial null" to keep SPCR enabled while
> + * discarding firmware serial output (avoid corrupting TAP
> + * stdout).
> */
> if (data->cd) {
> args = g_strdup_printf("-machine %s%s %s -accel tcg "
> - "-nodefaults -nographic "
> + "-nodefaults -serial null -nographic "
> "-drive if=pflash,format=raw,file=%s,readonly=on "
> "-drive if=pflash,format=raw,file=%s,snapshot=on -cdrom %s %s",
> data->machine, data->machine_param ?: "",
> @@ -836,7 +839,7 @@ static char *test_acpi_create_args(test_data *data, const char *params)
> data->uefi_fl1, data->uefi_fl2, data->cd, params ? params : "");
> } else {
> args = g_strdup_printf("-machine %s%s %s -accel tcg "
> - "-nodefaults -nographic "
> + "-nodefaults -serial null -nographic "
> "-drive if=pflash,format=raw,file=%s,readonly=on "
> "-drive if=pflash,format=raw,file=%s,snapshot=on %s",
> data->machine, data->machine_param ?: "",
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH v7 1/1] acpi/virt: suppress SPCR when UART0 has no serial backend
2026-02-19 13:04 [PATCH v7 1/1] acpi/virt: suppress SPCR when UART0 has no serial backend Li Chen
2026-02-19 13:42 ` Igor Mammedov
@ 2026-02-20 14:50 ` Michael S. Tsirkin
2026-02-22 3:25 ` Li Chen
1 sibling, 1 reply; 4+ messages in thread
From: Michael S. Tsirkin @ 2026-02-20 14:50 UTC (permalink / raw)
To: Li Chen
Cc: qemu-devel, qemu-arm, qemu-riscv, Igor Mammedov, Ani Sinha,
Peter Maydell, Shannon Zhao, Song Gao, Bibo Mao, Jiaxun Yang,
Sunil V L, Palmer Dabbelt, Alistair Francis, Weiwei Li,
Daniel Henrique Barboza, Liu Zhiwei, Philippe Mathieu-Daudé
On Thu, Feb 19, 2026 at 09:04:02PM +0800, Li Chen wrote:
> From: Li Chen <chenl311@chinatelecom.cn>
>
> virt machines always instantiate a PL011/16550 UART at slot 0 and describe
> it in ACPI (DSDT and optional SPCR table). When the command line disables
> the serial backend (e.g. "-serial none"), SPCR still points at UART0 and
> the guest may pick it as the preferred console even though it is not
> usable.
>
> Only generate SPCR when UART0 has a chardev backend attached. Keep the
> UART device description in DSDT unchanged so the guest-visible UART
> topology does not change.
>
> The bios-tables-test qtests use "-serial null" to keep SPCR enabled while
> discarding firmware serial output, avoiding TAP stdout corruption.
>
> Suggested-by: Philippe Mathieu-Daudé <philmd@linaro.org>
> Signed-off-by: Li Chen <chenl311@chinatelecom.cn>
> Reviewed-by: Sunil V L <sunilvl@ventanamicro.com>
> ---
So -serial null gives a device -serial none hides it from guest.
Sensible. And yet, out of abundance of caution, maybe keep
the legacy behaviour for existing machine types. No?
> v6 -> v7:
> Do not gate DSDT UART devices on serial_hd(0); only gate SPCR.
> Drop the serial_exists() helper added by v6.
> Keep bios-tables-test using "-serial null" (comment wording update).
>
> v5 -> v6:
> Fix TAP parsing errors by using "-serial null" in bios-tables-test.
>
> v4 -> v5:
> Also suppress UART device & SPCR when guest has no serial hardware on
> loongarch.
> Rename serial_exist to serial_exists.
> Fix style issues.
>
> hw/arm/virt-acpi-build.c | 3 ++-
> hw/loongarch/virt-acpi-build.c | 2 +-
> hw/riscv/virt-acpi-build.c | 3 ++-
> tests/qtest/bios-tables-test.c | 7 +++++--
> 4 files changed, 10 insertions(+), 5 deletions(-)
>
> diff --git a/hw/arm/virt-acpi-build.c b/hw/arm/virt-acpi-build.c
> index c145678185..1411f98ea1 100644
> --- a/hw/arm/virt-acpi-build.c
> +++ b/hw/arm/virt-acpi-build.c
> @@ -64,6 +64,7 @@
> #include "hw/virtio/virtio-acpi.h"
> #include "target/arm/cpu.h"
> #include "target/arm/multiprocessing.h"
> +#include "system/system.h"
>
> #define ARM_SPI_BASE 32
>
> @@ -1291,7 +1292,7 @@ void virt_acpi_build(VirtMachineState *vms, AcpiBuildTables *tables)
>
> acpi_add_table(table_offsets, tables_blob);
>
> - if (ms->acpi_spcr_enabled) {
> + if (ms->acpi_spcr_enabled && serial_hd(0)) {
> spcr_setup(tables_blob, tables->linker, vms);
> }
>
> diff --git a/hw/loongarch/virt-acpi-build.c b/hw/loongarch/virt-acpi-build.c
> index 54ac94e17d..dc8bbb9fdc 100644
> --- a/hw/loongarch/virt-acpi-build.c
> +++ b/hw/loongarch/virt-acpi-build.c
> @@ -546,7 +546,7 @@ static void acpi_build(AcpiBuildTables *tables, MachineState *machine)
> build_srat(tables_blob, tables->linker, machine);
> acpi_add_table(table_offsets, tables_blob);
>
> - if (machine->acpi_spcr_enabled)
> + if (machine->acpi_spcr_enabled && serial_hd(0))
> spcr_setup(tables_blob, tables->linker, machine);
>
> if (machine->numa_state->num_nodes) {
> diff --git a/hw/riscv/virt-acpi-build.c b/hw/riscv/virt-acpi-build.c
> index f1406cb683..29080d1125 100644
> --- a/hw/riscv/virt-acpi-build.c
> +++ b/hw/riscv/virt-acpi-build.c
> @@ -39,6 +39,7 @@
> #include "qapi/error.h"
> #include "qemu/error-report.h"
> #include "system/reset.h"
> +#include "system/system.h"
>
> #define ACPI_BUILD_TABLE_SIZE 0x20000
> #define ACPI_BUILD_INTC_ID(socket, index) ((socket << 24) | (index))
> @@ -890,7 +891,7 @@ static void virt_acpi_build(RISCVVirtState *s, AcpiBuildTables *tables)
>
> acpi_add_table(table_offsets, tables_blob);
>
> - if (ms->acpi_spcr_enabled) {
> + if (ms->acpi_spcr_enabled && serial_hd(0)) {
> spcr_setup(tables_blob, tables->linker, s);
> }
>
> diff --git a/tests/qtest/bios-tables-test.c b/tests/qtest/bios-tables-test.c
> index e489d94331..37f2b72d83 100644
> --- a/tests/qtest/bios-tables-test.c
> +++ b/tests/qtest/bios-tables-test.c
> @@ -825,10 +825,13 @@ static char *test_acpi_create_args(test_data *data, const char *params)
> /*
> * TODO: convert '-drive if=pflash' to new syntax (see e33763be7cd3)
> * when arm/virt boad starts to support it.
> + * NOTE: Explicitly add "-serial null" to keep SPCR enabled while
> + * discarding firmware serial output (avoid corrupting TAP
> + * stdout).
> */
> if (data->cd) {
> args = g_strdup_printf("-machine %s%s %s -accel tcg "
> - "-nodefaults -nographic "
> + "-nodefaults -serial null -nographic "
> "-drive if=pflash,format=raw,file=%s,readonly=on "
> "-drive if=pflash,format=raw,file=%s,snapshot=on -cdrom %s %s",
> data->machine, data->machine_param ?: "",
> @@ -836,7 +839,7 @@ static char *test_acpi_create_args(test_data *data, const char *params)
> data->uefi_fl1, data->uefi_fl2, data->cd, params ? params : "");
> } else {
> args = g_strdup_printf("-machine %s%s %s -accel tcg "
> - "-nodefaults -nographic "
> + "-nodefaults -serial null -nographic "
> "-drive if=pflash,format=raw,file=%s,readonly=on "
> "-drive if=pflash,format=raw,file=%s,snapshot=on %s",
> data->machine, data->machine_param ?: "",
> --
> 2.52.0
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH v7 1/1] acpi/virt: suppress SPCR when UART0 has no serial backend
2026-02-20 14:50 ` Michael S. Tsirkin
@ 2026-02-22 3:25 ` Li Chen
0 siblings, 0 replies; 4+ messages in thread
From: Li Chen @ 2026-02-22 3:25 UTC (permalink / raw)
To: Michael S. Tsirkin
Cc: qemu-devel, qemu-arm, qemu-riscv, Igor Mammedov, Ani Sinha,
Peter Maydell, Shannon Zhao, Song Gao, Bibo Mao, Jiaxun Yang,
Sunil V L, Palmer Dabbelt, Alistair Francis, Weiwei Li,
Daniel Henrique Barboza, Liu Zhiwei,
"Philippe Mathieu-Daudé"
Hi Michael,
---- On Fri, 20 Feb 2026 22:50:57 +0800 Michael S. Tsirkin <mst@redhat.com> wrote ---
> On Thu, Feb 19, 2026 at 09:04:02PM +0800, Li Chen wrote:
> > From: Li Chen <chenl311@chinatelecom.cn>
> >
> > virt machines always instantiate a PL011/16550 UART at slot 0 and describe
> > it in ACPI (DSDT and optional SPCR table). When the command line disables
> > the serial backend (e.g. "-serial none"), SPCR still points at UART0 and
> > the guest may pick it as the preferred console even though it is not
> > usable.
> >
> > Only generate SPCR when UART0 has a chardev backend attached. Keep the
> > UART device description in DSDT unchanged so the guest-visible UART
> > topology does not change.
> >
> > The bios-tables-test qtests use "-serial null" to keep SPCR enabled while
> > discarding firmware serial output, avoiding TAP stdout corruption.
> >
> > Suggested-by: Philippe Mathieu-Daudé <philmd@linaro.org>
> > Signed-off-by: Li Chen <chenl311@chinatelecom.cn>
> > Reviewed-by: Sunil V L <sunilvl@ventanamicro.com>
> > ---
>
> So -serial null gives a device -serial none hides it from guest.
> Sensible. And yet, out of abundance of caution, maybe keep
> the legacy behaviour for existing machine types. No?
Thanks. I see your point about being cautious with legacy behaviour.
Given you and Igor's feedback, gating SPCR on chardev backend state is not the
right approach, so I'll drop this patch.
Regards,
Li
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2026-02-22 3:26 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-02-19 13:04 [PATCH v7 1/1] acpi/virt: suppress SPCR when UART0 has no serial backend Li Chen
2026-02-19 13:42 ` Igor Mammedov
2026-02-20 14:50 ` Michael S. Tsirkin
2026-02-22 3:25 ` Li Chen
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.