Skip to contentSkip to conversation

hw/tpm: TIS on sysbus: Remove unsupport ppi command line option

PATCH In review Stefan Berger 3 participants 7 messages Jul 13, 2023 – Jul 14, 2023

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 hw/tpm: TIS on sysbus: Remove unsupport ppi command line option

The ppi command line option for the TIS device on sysbus never worked and caused an immediate segfault. Remove support for it since it also needs support in the firmware and needs testing inside the VM.

Reproducer with the ppi=on option passed:

qemu-system-aarch64 \
-machine virt,gic-version=3 \
-m 4G \
-nographic -no-acpi \
-chardev socket,id=chrtpm,path=/tmp/mytpm1/swtpm-sock \
-tpmdev emulator,id=tpm0,chardev=chrtpm \
-device tpm-tis-device,tpmdev=tpm0,ppi=on
[...]
Segmentation fault (core dumped)

---
 hw/tpm/tpm_tis_sysbus.c | 1 -
 1 file changed, 1 deletion(-)
hw/tpm/tpm_tis_sysbus.c +0 -1
index 45e63efd63..6724b3d4f6 100644
@@ -93,7 +93,6 @@ static void tpm_tis_sysbus_reset(DeviceState *dev)
 static Property tpm_tis_sysbus_properties[] = {
     DEFINE_PROP_UINT32("irq", TPMStateSysBus, state.irq_num, TPM_TIS_IRQ),
     DEFINE_PROP_TPMBE("tpmdev", TPMStateSysBus, state.be_driver),
-    DEFINE_PROP_BOOL("ppi", TPMStateSysBus, state.ppi_enabled, false),
     DEFINE_PROP_END_OF_LIST(),
 };
 
Signature · 1 line
2.41.0
Reviewed-by on v1 1/1
#1 lore raw

Hi Stefan,

Quoted 13 lines · Stefan Berger

The ppi command line option for the TIS device on sysbus never worked and caused an immediate segfault. Remove support for it since it also needs support in the firmware and needs testing inside the VM.

Reproducer with the ppi=on option passed:

qemu-system-aarch64 \
   -machine virt,gic-version=3 \
   -m 4G  \
   -nographic -no-acpi \
   -chardev socket,id=chrtpm,path=/tmp/mytpm1/swtpm-sock \
   -tpmdev emulator,id=tpm0,chardev=chrtpm \
   -device tpm-tis-device,tpmdev=tpm0,ppi=on
Quoted 4 lines

[...]
Segmentation fault (core dumped)

Thanks!

Eric

Quoted 16 lines
---
 hw/tpm/tpm_tis_sysbus.c | 1 -
 1 file changed, 1 deletion(-)
hw/tpm/tpm_tis_sysbus.c +0 -1
index 45e63efd63..6724b3d4f6 100644
@@ -93,7 +93,6 @@ static void tpm_tis_sysbus_reset(DeviceState *dev)
 static Property tpm_tis_sysbus_properties[] = {
     DEFINE_PROP_UINT32("irq", TPMStateSysBus, state.irq_num, TPM_TIS_IRQ),
     DEFINE_PROP_TPMBE("tpmdev", TPMStateSysBus, state.be_driver),
-    DEFINE_PROP_BOOL("ppi", TPMStateSysBus, state.ppi_enabled, false),
     DEFINE_PROP_END_OF_LIST(),
 };
 
#2 lore raw
In reply to Stefan Berger
Quoted 14 lines · Stefan Berger

The ppi command line option for the TIS device on sysbus never worked and caused an immediate segfault. Remove support for it since it also needs support in the firmware and needs testing inside the VM.

Reproducer with the ppi=on option passed:

qemu-system-aarch64 \
   -machine virt,gic-version=3 \
   -m 4G  \
   -nographic -no-acpi \
   -chardev socket,id=chrtpm,path=/tmp/mytpm1/swtpm-sock \
   -tpmdev emulator,id=tpm0,chardev=chrtpm \
   -device tpm-tis-device,tpmdev=tpm0,ppi=on
Quoted 4 lines

[...]
Segmentation fault (core dumped)

Do you need to add a VMSTATE_UNUSED_TEST in case a future QEMU version introduces a new field in the same position which will cause an issue when restoring from an older version?

#3 lore raw
Quoted 19 lines · Joelle van Dyne
Quoted 17 lines · Stefan Berger

The ppi command line option for the TIS device on sysbus never worked and caused an immediate segfault. Remove support for it since it also needs support in the firmware and needs testing inside the VM.

Reproducer with the ppi=on option passed:

qemu-system-aarch64 \
   -machine virt,gic-version=3 \
   -m 4G  \
   -nographic -no-acpi \
   -chardev socket,id=chrtpm,path=/tmp/mytpm1/swtpm-sock \
   -tpmdev emulator,id=tpm0,chardev=chrtpm \
   -device tpm-tis-device,tpmdev=tpm0,ppi=on

[...]
Segmentation fault (core dumped)

Quoted 5 lines
Quoted 1 line

Do you need to add a VMSTATE_UNUSED_TEST in case a future QEMU version introduces a new field in the same position which will cause an issue when restoring from an older version?

Actually, ignore that last message. I misread the patch thinking it's on the VMState.

#4 lore raw
In reply to Joelle van Dyne
Quoted 18 lines · Joelle van Dyne
Quoted 17 lines · Stefan Berger

The ppi command line option for the TIS device on sysbus never worked and caused an immediate segfault. Remove support for it since it also needs support in the firmware and needs testing inside the VM.

Reproducer with the ppi=on option passed:

qemu-system-aarch64 \
    -machine virt,gic-version=3 \
    -m 4G  \
    -nographic -no-acpi \
    -chardev socket,id=chrtpm,path=/tmp/mytpm1/swtpm-sock \
    -tpmdev emulator,id=tpm0,chardev=chrtpm \
    -device tpm-tis-device,tpmdev=tpm0,ppi=on

[...]
Segmentation fault (core dumped)

Quoted 5 lines
Quoted 1 line

Do you need to add a VMSTATE_UNUSED_TEST in case a future QEMU version introduces a new field in the same position which will cause an issue when restoring from an older version?

Hm, you got a point there. We will have to error-out in case someone sets ppi=on instead since the expectation that PPI would work is simply not there. v2 coming soon.

Stefan

#5 lore raw

Hi Stefan,

Quoted 24 lines · Stefan Berger
Quoted 21 lines · Joelle van Dyne
Quoted 18 lines · Stefan Berger

The ppi command line option for the TIS device on sysbus never worked and caused an immediate segfault. Remove support for it since it also needs support in the firmware and needs testing inside the VM.

Reproducer with the ppi=on option passed:

qemu-system-aarch64 \
    -machine virt,gic-version=3 \
    -m 4G  \
    -nographic -no-acpi \
    -chardev socket,id=chrtpm,path=/tmp/mytpm1/swtpm-sock \
    -tpmdev emulator,id=tpm0,chardev=chrtpm \
    -device tpm-tis-device,tpmdev=tpm0,ppi=on [...]
Segmentation fault (core dumped)

Quoted 7 lines
Quoted 3 lines

Do you need to add a VMSTATE_UNUSED_TEST in case a future QEMU version introduces a new field in the same position which will cause an issue when restoring from an older version?

Hm, you got a point there. We will have to error-out in case someone sets ppi=on instead since the expectation that PPI would work is simply not there. v2 coming soon.

as Joelle pointed it out ppi_enabled is not part of vmstate_tpm_tis_sysbus fields. And since it has never worked I suspect we cannot have any existing VM enabling it. So I don't get the issue with this 1st version?

Thanks

Eric

Quoted 3 lines

    Stefan

#6 lore raw
Quoted 30 lines · Eric Auger

Hi Stefan,

Quoted 28 lines · Stefan Berger
Quoted 24 lines · Joelle van Dyne
Quoted 17 lines · Stefan Berger

The ppi command line option for the TIS device on sysbus never worked and caused an immediate segfault. Remove support for it since it also needs support in the firmware and needs testing inside the VM.

Reproducer with the ppi=on option passed:

qemu-system-aarch64 \
    -machine virt,gic-version=3 \
    -m 4G  \
    -nographic -no-acpi \
    -chardev socket,id=chrtpm,path=/tmp/mytpm1/swtpm-sock \
    -tpmdev emulator,id=tpm0,chardev=chrtpm \
    -device tpm-tis-device,tpmdev=tpm0,ppi=on
[...]
Segmentation fault (core dumped)

Quoted 1 line

Do you need to add a VMSTATE_UNUSED_TEST in case a future QEMU version introduces a new field in the same position which will cause an issue when restoring from an older version?

Quoted 7 lines
Quoted 3 lines

Hm, you got a point there. We will have to error-out in case someone sets ppi=on instead since the expectation that PPI would work is simply not there. v2 coming soon.

as Joelle pointed it out ppi_enabled is not part of vmstate_tpm_tis_sysbus fields. And since it has never worked I suspect we cannot have any existing VM enabling it. So I don't get the issue with this 1st version?

You are right. I repeated my test with restoring state of a VM taken before the removal of this field and it restored it. So that other patch is good and I am withdrawing this patch here.

Stefan

Quoted 8 lines

Thanks

Eric

Quoted 3 lines

    Stefan