spapr: Implement H_TPM_COMM for accessing host TPM device
cover spapr: Implement H_TPM_COMM for accessing host TPM device
These patches are also available at:
https://github.com/mdroth/qemu/commits/spapr-tpm-hcall-v0
This patchset implements the H_TPM_COMM hypercall, which provides a way for an Ultravisor to pass raw TPM commands on to a host's TPM device, either directly or through a TPM Resource Manager (needed to support multiple guests).
Secure Guests running on an Ultravisor have a symmetric key that is encrypted using a public key that is bound to a trusted host's TPM hardware. This hypercall provides a means to decrypt the symmetric key on behalf of a Secure Guest using the host's TPM hardware.
More details are provided in the spec summary introduced in patch 1.
docs/specs/ppc-spapr-uv-hcalls.txt | 74 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ hw/ppc/Makefile.objs | 1 + hw/ppc/spapr.c | 27 +++++++++++++++++++++++++++ hw/ppc/spapr_hcall_tpm.c | 135 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ hw/ppc/trace-events | 4 ++++ include/hw/ppc/spapr.h | 7 ++++++- 6 files changed, 247 insertions(+), 1 deletion(-)
PATCH 1/2 docs/specs: initial spec summary for Ultravisor-related hcalls
For now this only covers hcalls relating to TPM communication since it's the only one particularly important from a QEMU perspective atm, but others can be added here where it makes sense.
The full specification for all hcalls/ucalls will eventually be made available in the public/OpenPower version of the PAPR specification.
--- docs/specs/ppc-spapr-uv-hcalls.txt | 74 ++++++++++++++++++++++++++++++ 1 file changed, 74 insertions(+) create mode 100644 docs/specs/ppc-spapr-uv-hcalls.txt
docs/specs/ppc-spapr-uv-hcalls.txt new file +74 -0
@@ -0,0 +1,74 @@ +On PPC64 systems supporting Protected Execution Facility (PEF), system +memory can be placed in a secured region where only an "ultravisor" +running in firmware can provide to access it. pseries guests on such +systems can communicate with the ultravisor (via ultracalls) to switch to a +secure VM mode (SVM) where the guest's memory is relocated to this secured +region, making its memory inaccessible to normal processes/guests running on +the host. + +The various ultracalls/hypercalls relating to SVM mode are currently +only documented internally, but are planned for direct inclusion into the +public OpenPOWER version of the PAPR specification (LoPAPR/LoPAR). An internal +ACR has been filed to reserve a hypercall number range specific to this +use-case to avoid any future conflicts with the internally-maintained PAPR +specification. This document summarizes some of these details as they relate +to QEMU. + +== hypercalls needed by the ultravisor == + +Switching to SVM mode involves a number of hcalls issued by the ultravisor +to the hypervisor to orchestrate the movement of guest memory to secure +memory and various other aspects SVM mode. The below documents the hcalls +relevant to QEMU. + +- H_TPM_COMM (0xef10) + + For TPM_COMM_OP_EXECUTE operation: + Send a request to a TPM and receive a response, opening a new TPM session + if one has not already been opened. + + For TPM_COMM_OP_CLOSE_SESSION operation: + Close the existing TPM session, if any. + + Arguments: + + r3 : H_TPM_COMM (0xef10) + r4 : TPM operation, one of: + TPM_COMM_OP_EXECUTE (0x1) + TPM_COMM_OP_CLOSE_SESSION (0x2) + r5 : in_buffer, guest physical address of buffer containing the request + - Caller may use the same address for both request and response + r6 : in_size, size of the in buffer, must + - Must be less than or equal to 4KB + r7 : out_buffer, guest physical address of buffer to store the response + - Caller may use the same address for both request and response + r8 : out_size, size of the out buffer + - Must be at least 4KB, as this is the maximum request/response size + supported by most TPM implementations, including the TPM Resource + Manager in the linux kernel. + + Return values: + + r3 : H_Success request processed successfully + H_PARAMETER invalid TPM operation + H_P2 in_buffer is invalid + H_P3 in_size is invalid + H_P4 out_buffer is invalid + H_P5 out_size is invalid + H_RESOURCE TPM is unavailable + r4 : For TPM_COMM_OP_EXECUTE, the size of the response will be stored here + upon success. + + Use-case/notes: + + SVM filesystems are encrypted using a symmetric key. This key is then + wrapped/encrypted using the public key of a trusted system which has the + private key stored in the system's TPM. An Ultravisor will use this + hcall to unwrap/unseal the symmetric key using the system's TPM device + or a TPM Resource Manager associated with the device. + + The Ultravisor sets up a separate session key with the TPM in advance + during host system boot. All sensitive in and out values will be + encrypted using the session key. Though the hypervisor will see the 'in' + and 'out' buffers in raw form, any sensitive contents will generally be + encrypted using this session key.
Signature · 1 line
2.17.1
Re: [Qemu-devel] [PATCH 1/2] docs/specs: initial spec summary for Ultravisor-related hcalls
Quoted 8 lines · Michael Roth
For now this only covers hcalls relating to TPM communication since it's the only one particularly important from a QEMU perspective atm, but others can be added here where it makes sense.
The full specification for all hcalls/ucalls will eventually be made available in the public/OpenPower version of the PAPR specification.
Signed-off-by Michael Roth <[email protected]>
Thanks for adding this documentation. Is there a PAPR extension proposal which covers this, which we could cite as the source?
Quoted 85 lines
--- docs/specs/ppc-spapr-uv-hcalls.txt | 74 ++++++++++++++++++++++++++++++ 1 file changed, 74 insertions(+) create mode 100644 docs/specs/ppc-spapr-uv-hcalls.txtdocs/specs/ppc-spapr-uv-hcalls.txt new file +74
-0@@ -0,0 +1,74 @@ +On PPC64 systems supporting Protected Execution Facility (PEF), system +memory can be placed in a secured region where only an "ultravisor" +running in firmware can provide to access it. pseries guests on such +systems can communicate with the ultravisor (via ultracalls) to switch to a +secure VM mode (SVM) where the guest's memory is relocated to this secured +region, making its memory inaccessible to normal processes/guests running on +the host. + +The various ultracalls/hypercalls relating to SVM mode are currently +only documented internally, but are planned for direct inclusion into the +public OpenPOWER version of the PAPR specification (LoPAPR/LoPAR). An internal +ACR has been filed to reserve a hypercall number range specific to this +use-case to avoid any future conflicts with the internally-maintained PAPR +specification. This document summarizes some of these details as they relate +to QEMU. + +== hypercalls needed by the ultravisor == + +Switching to SVM mode involves a number of hcalls issued by the ultravisor +to the hypervisor to orchestrate the movement of guest memory to secure +memory and various other aspects SVM mode. The below documents the hcalls +relevant to QEMU. + +- H_TPM_COMM (0xef10) + + For TPM_COMM_OP_EXECUTE operation: + Send a request to a TPM and receive a response, opening a new TPM session + if one has not already been opened. + + For TPM_COMM_OP_CLOSE_SESSION operation: + Close the existing TPM session, if any. + + Arguments: + + r3 : H_TPM_COMM (0xef10) + r4 : TPM operation, one of: + TPM_COMM_OP_EXECUTE (0x1) + TPM_COMM_OP_CLOSE_SESSION (0x2) + r5 : in_buffer, guest physical address of buffer containing the request + - Caller may use the same address for both request and response + r6 : in_size, size of the in buffer, must + - Must be less than or equal to 4KB + r7 : out_buffer, guest physical address of buffer to store the response + - Caller may use the same address for both request and response + r8 : out_size, size of the out buffer + - Must be at least 4KB, as this is the maximum request/response size + supported by most TPM implementations, including the TPM Resource + Manager in the linux kernel. + + Return values: + + r3 : H_Success request processed successfully + H_PARAMETER invalid TPM operation + H_P2 in_buffer is invalid + H_P3 in_size is invalid + H_P4 out_buffer is invalid + H_P5 out_size is invalid + H_RESOURCE TPM is unavailable + r4 : For TPM_COMM_OP_EXECUTE, the size of the response will be stored here + upon success. + + Use-case/notes: + + SVM filesystems are encrypted using a symmetric key. This key is then + wrapped/encrypted using the public key of a trusted system which has the + private key stored in the system's TPM. An Ultravisor will use this + hcall to unwrap/unseal the symmetric key using the system's TPM device + or a TPM Resource Manager associated with the device. + + The Ultravisor sets up a separate session key with the TPM in advance + during host system boot. All sensitive in and out values will be + encrypted using the session key. Though the hypervisor will see the 'in' + and 'out' buffers in raw form, any sensitive contents will generally be + encrypted using this session key.
Signature · 4 lines
David Gibson | I'll have my music baroque, and my code david AT gibson.dropbear.id.au | minimalist, thank you. NOT _the_ _other_ | _way_ _around_! http://www.ozlabs.org/~dgibson
Re: [Qemu-devel] [PATCH 1/2] docs/specs: initial spec summary for Ultravisor-related hcalls
Quoting David Gibson (2019-07-12 01:40:27)
Quoted 5 lines · Michael Roth
Quoted 4 lines
For now this only covers hcalls relating to TPM communication since it's the only one particularly important from a QEMU perspective atm, but others can be added here where it makes sense.
Quoted 7 lines
Quoted 4 lines
The full specification for all hcalls/ucalls will eventually be made available in the public/OpenPower version of the PAPR specification.
Signed-off-by Michael Roth <[email protected]>Thanks for adding this documentation. Is there a PAPR extension proposal which covers this, which we could cite as the source?
We have an internal document/repo that serves as a catch-all for the Ultravisor related spec changes. We could make that available publically via github and cite it here until it hits the full spec. Would that work?
Quoted 92 lines
Quoted 86 lines
--- docs/specs/ppc-spapr-uv-hcalls.txt | 74 ++++++++++++++++++++++++++++++ 1 file changed, 74 insertions(+) create mode 100644 docs/specs/ppc-spapr-uv-hcalls.txtdocs/specs/ppc-spapr-uv-hcalls.txt new file +74
-0@@ -0,0 +1,74 @@ +On PPC64 systems supporting Protected Execution Facility (PEF), system +memory can be placed in a secured region where only an "ultravisor" +running in firmware can provide to access it. pseries guests on such +systems can communicate with the ultravisor (via ultracalls) to switch to a +secure VM mode (SVM) where the guest's memory is relocated to this secured +region, making its memory inaccessible to normal processes/guests running on +the host. + +The various ultracalls/hypercalls relating to SVM mode are currently +only documented internally, but are planned for direct inclusion into the +public OpenPOWER version of the PAPR specification (LoPAPR/LoPAR). An internal +ACR has been filed to reserve a hypercall number range specific to this +use-case to avoid any future conflicts with the internally-maintained PAPR +specification. This document summarizes some of these details as they relate +to QEMU. + +== hypercalls needed by the ultravisor == + +Switching to SVM mode involves a number of hcalls issued by the ultravisor +to the hypervisor to orchestrate the movement of guest memory to secure +memory and various other aspects SVM mode. The below documents the hcalls +relevant to QEMU. + +- H_TPM_COMM (0xef10) + + For TPM_COMM_OP_EXECUTE operation: + Send a request to a TPM and receive a response, opening a new TPM session + if one has not already been opened. + + For TPM_COMM_OP_CLOSE_SESSION operation: + Close the existing TPM session, if any. + + Arguments: + + r3 : H_TPM_COMM (0xef10) + r4 : TPM operation, one of: + TPM_COMM_OP_EXECUTE (0x1) + TPM_COMM_OP_CLOSE_SESSION (0x2) + r5 : in_buffer, guest physical address of buffer containing the request + - Caller may use the same address for both request and response + r6 : in_size, size of the in buffer, must + - Must be less than or equal to 4KB + r7 : out_buffer, guest physical address of buffer to store the response + - Caller may use the same address for both request and response + r8 : out_size, size of the out buffer + - Must be at least 4KB, as this is the maximum request/response size + supported by most TPM implementations, including the TPM Resource + Manager in the linux kernel. + + Return values: + + r3 : H_Success request processed successfully + H_PARAMETER invalid TPM operation + H_P2 in_buffer is invalid + H_P3 in_size is invalid + H_P4 out_buffer is invalid + H_P5 out_size is invalid + H_RESOURCE TPM is unavailable + r4 : For TPM_COMM_OP_EXECUTE, the size of the response will be stored here + upon success. + + Use-case/notes: + + SVM filesystems are encrypted using a symmetric key. This key is then + wrapped/encrypted using the public key of a trusted system which has the + private key stored in the system's TPM. An Ultravisor will use this + hcall to unwrap/unseal the symmetric key using the system's TPM device + or a TPM Resource Manager associated with the device. + + The Ultravisor sets up a separate session key with the TPM in advance + during host system boot. All sensitive in and out values will be + encrypted using the session key. Though the hypervisor will see the 'in' + and 'out' buffers in raw form, any sensitive contents will generally be + encrypted using this session key.Signature · 4 lines
David Gibson | I'll have my music baroque, and my code david AT gibson.dropbear.id.au | minimalist, thank you. NOT _the_ _other_ | _way_ _around_! http://www.ozlabs.org/~dgibson
Re: [Qemu-devel] [PATCH 1/2] docs/specs: initial spec summary for Ultravisor-related hcalls
Quoted 9 lines · Michael Roth
Quoting David Gibson (2019-07-12 01:40:27)
Quoted 8 lines · Michael Roth
Quoted 7 lines
For now this only covers hcalls relating to TPM communication since it's the only one particularly important from a QEMU perspective atm, but others can be added here where it makes sense.
The full specification for all hcalls/ucalls will eventually be made available in the public/OpenPower version of the PAPR specification.
Quoted 8 lines
Quoted 4 lines
Quoted 1 line
Signed-off-by Michael Roth <[email protected]>Thanks for adding this documentation. Is there a PAPR extension proposal which covers this, which we could cite as the source?
We have an internal document/repo that serves as a catch-all for the Ultravisor related spec changes. We could make that available publically via github and cite it here until it hits the full spec. Would that work?
Yes, that sounds good.
Signature · 4 lines
David Gibson | I'll have my music baroque, and my code david AT gibson.dropbear.id.au | minimalist, thank you. NOT _the_ _other_ | _way_ _around_! http://www.ozlabs.org/~dgibson
Re: [Qemu-devel] [PATCH 1/2] docs/specs: initial spec summary for Ultravisor-related hcalls
Quoting David Gibson (2019-07-14 21:25:55)
Quoted 12 lines · Michael Roth
Quoted 11 lines
Quoting David Gibson (2019-07-12 01:40:27)
Quoted 10 lines · Michael Roth
Quoted 8 lines
For now this only covers hcalls relating to TPM communication since it's the only one particularly important from a QEMU perspective atm, but others can be added here where it makes sense.
The full specification for all hcalls/ucalls will eventually be made available in the public/OpenPower version of the PAPR specification.
Signed-off-by Michael Roth <[email protected]>
Quoted 8 lines
Quoted 6 lines
Quoted 2 lines
Thanks for adding this documentation. Is there a PAPR extension proposal which covers this, which we could cite as the source?
We have an internal document/repo that serves as a catch-all for the Ultravisor related spec changes. We could make that available publically via github and cite it here until it hits the full spec. Would that work?
Yes, that sounds good.
Ok, we're working on getting that posted externally. If it's not up in time for next submission I will send a follow-up patch to add a reference.
Quoted 6 lines
Signature · 5 lines
-- David Gibson | I'll have my music baroque, and my code david AT gibson.dropbear.id.au | minimalist, thank you. NOT _the_ _other_ | _way_ _around_! http://www.ozlabs.org/~dgibson
PATCH 2/2 spapr: initial implementation for H_TPM_COMM hcall
In reply to Michael RothThis implements the H_TPM_COMM hypercall, which is used by an Ultravisor to pass TPM commands directly to the host's TPM device, or a TPM Resource Manager associated with the device.
This also introduces a new pseries machine option which is used to configure what TPM device to pass commands to, for example:
-machine pseries,...,tpm-device-file=/dev/tmprm0
By default, no tpm-device-file is defined and hcalls will return H_RESOURCE.
The full specification for this hypercall can be found in docs/specs/ppc-spapr-uv-hcalls.txt
--- hw/ppc/Makefile.objs | 1 + hw/ppc/spapr.c | 27 ++++++++ hw/ppc/spapr_hcall_tpm.c | 135 +++++++++++++++++++++++++++++++++++++++ hw/ppc/trace-events | 4 ++ include/hw/ppc/spapr.h | 7 +- 5 files changed, 173 insertions(+), 1 deletion(-) create mode 100644 hw/ppc/spapr_hcall_tpm.c
hw/ppc/Makefile.objs +1 -0
@@ -5,6 +5,7 @@ obj-$(CONFIG_PSERIES) += spapr.o spapr_caps.o spapr_vio.o spapr_events.o obj-$(CONFIG_PSERIES) += spapr_hcall.o spapr_iommu.o spapr_rtas.o obj-$(CONFIG_PSERIES) += spapr_pci.o spapr_rtc.o spapr_drc.o obj-$(CONFIG_PSERIES) += spapr_cpu_core.o spapr_ovec.o spapr_irq.o +obj-$(CONFIG_PSERIES) += spapr_hcall_tpm.o obj-$(CONFIG_SPAPR_RNG) += spapr_rng.o # IBM PowerNV obj-$(CONFIG_POWERNV) += pnv.o pnv_xscom.o pnv_core.o pnv_lpc.o pnv_psi.o pnv_occ.o pnv_bmc.o
hw/ppc/spapr.c +27 -0
@@ -1776,6 +1776,10 @@ static void spapr_machine_reset(MachineState *machine)*/ object_child_foreach_recursive (object_get_root (), spapr_reset_drcs,NULL ); +if (spapr->tpm_device_file) { +spapr_hcall_tpm_reset (); + } +spapr_clear_pending_events (spapr);/*
@@ -3340,6 +3344,21 @@ static void spapr_set_host_serial(Object *obj, const char *value, Error **errp) spapr->host_serial =g_strdup (value); } +static char *spapr_get_tpm_device_file (Object *obj,Error **errp) +{ +SpaprMachineState *spapr =SPAPR_MACHINE (obj); + +return g_strdup (spapr->tpm_device_file); +} + +static void spapr_set_tpm_device_file (Object *obj,const char *value,Error **errp) +{ +SpaprMachineState *spapr =SPAPR_MACHINE (obj); + +g_free (spapr->tpm_device_file); + spapr->tpm_device_file =g_strdup (value); +} +static void spapr_instance_init (Object *obj) {SpaprMachineState *spapr =SPAPR_MACHINE (obj);
@@ -3396,6 +3415,14 @@ static void spapr_instance_init(Object *obj) &error_abort);object_property_set_description (obj,"host-serial" ,"Host serial number to advertise in guest device tree" , &error_abort); +object_property_add_str (obj,"tpm-device-file" , + spapr_get_tpm_device_file, + spapr_set_tpm_device_file, &error_abort); +object_property_set_description (obj,"tpm-device-file" , +"Specifies the path to the TPM character device file to use" +" for TPM communication via hypercalls (usually a TPM" +" resource manager)" , + &error_abort); }static void spapr_machine_finalizefn (Object *obj)
hw/ppc/spapr_hcall_tpm.c new file +135 -0
135 lines not on this page · raw message
hw/ppc/trace-events +4 -0
@@ -25,6 +25,10 @@ spapr_update_dt(unsigned cb) "New blob %u bytes" spapr_update_dt_failed_size(unsigned cbold, unsigned cbnew, unsigned magic) "Old blob %u bytes, new blob %u bytes, magic 0x%x" spapr_update_dt_failed_check(unsigned cbold, unsigned cbnew, unsigned magic) "Old blob %u bytes, new blob %u bytes, magic 0x%x" +# spapr_hcall_tpm.c +spapr_h_tpm_comm(const char *device_path, uint64_t operation) "tpm_device_path=%s operation=0x%"PRIu64 +spapr_tpm_execute(uint64_t data_in, uint64_t data_in_sz, uint64_t data_out, uint64_t data_out_sz) "data_in=0x%"PRIx64", data_in_sz=%"PRIu64", data_out=0x%"PRIx64", data_out_sz=%"PRIu64 + # spapr_iommu.c spapr_iommu_put(uint64_t liobn, uint64_t ioba, uint64_t tce, uint64_t ret) "liobn=0x%"PRIx64" ioba=0x%"PRIx64" tce=0x%"PRIx64" ret=%"PRId64 spapr_iommu_get(uint64_t liobn, uint64_t ioba, uint64_t ret, uint64_t tce) "liobn=0x%"PRIx64" ioba=0x%"PRIx64" ret=%"PRId64" tce=0x%"PRIx64
include/hw/ppc/spapr.h +6 -1
@@ -198,6 +198,7 @@ struct SpaprMachineState {SpaprXive *xive;SpaprIrq *irq;qemu_irq *qirqs; +char *tpm_device_file;bool cmd_line_caps[SPAPR_CAP_NUM ];SpaprCapabilities def, eff, mig;
@@ -490,8 +491,9 @@ struct SpaprMachineState {#define H_INT_ESB 0x3C8#define H_INT_SYNC 0x3CC#define H_INT_RESET 0x3D0 +#define H_TPM_COMM 0xEF10 -#define MAX_HCALL_OPCODE H_INT_RESET +#define MAX_HCALL_OPCODE H_TPM_COMM/* The hcalls above are standardized in PAPR and implemented by pHyp * as well.
@@ -864,6 +866,9 @@ int spapr_caps_post_migration(SpaprMachineState *spapr);void spapr_check_pagesize (SpaprMachineState *spapr,hwaddr pagesize,Error **errp); + +void spapr_hcall_tpm_reset (void ); +/* * XIVE definitions */
Signature · 1 line
2.17.1
Re: [Qemu-devel] [PATCH 2/2] spapr: initial implementation for H_TPM_COMM hcall
Quoted 8 lines · Michael Roth
This implements the
H_TPM_COMMhypercall, which is used by an Ultravisor to pass TPM commands directly to the host's TPM device, or a TPM Resource Manager associated with the device.This also introduces a new pseries machine option which is used to configure what TPM device to pass commands to, for example:
-machine pseries,...,tpm-device-file=/dev/tmprm0
Bolting this into yet another machine parameter seems kind of ugly. Wouldn't it make more sense to treat this as an virtual device (say "spapr-vtpm"). Adding that device would enable the hcall, and would have properties for the back end host device.
Quoted 2 lines
By default, no tpm-device-file is defined and hcalls will return
H_RESOURCE.
Wouldn't H_FUNCTION make more sense?
Quoted 102 lines
The full specification for this hypercall can be found in
docs/specs/ppc-spapr-uv-hcalls.txtSigned-off-by Michael Roth <[email protected]--- hw/ppc/Makefile.objs | 1 + hw/ppc/spapr.c | 27 ++++++++ hw/ppc/spapr_hcall_tpm.c | 135 +++++++++++++++++++++++++++++++++++++++ hw/ppc/trace-events | 4 ++ include/hw/ppc/spapr.h | 7 +- 5 files changed, 173 insertions(+), 1 deletion(-) create mode 100644 hw/ppc/spapr_hcall_tpm.chw/ppc/Makefile.objs +1
-0@@ -5,6 +5,7 @@ obj-$(CONFIG_PSERIES) += spapr.o spapr_caps.o spapr_vio.o spapr_events.o obj-$(CONFIG_PSERIES) += spapr_hcall.o spapr_iommu.o spapr_rtas.o obj-$(CONFIG_PSERIES) += spapr_pci.o spapr_rtc.o spapr_drc.o obj-$(CONFIG_PSERIES) += spapr_cpu_core.o spapr_ovec.o spapr_irq.o +obj-$(CONFIG_PSERIES) += spapr_hcall_tpm.o obj-$(CONFIG_SPAPR_RNG) += spapr_rng.o # IBM PowerNV obj-$(CONFIG_POWERNV) += pnv.o pnv_xscom.o pnv_core.o pnv_lpc.o pnv_psi.o pnv_occ.o pnv_bmc.ohw/ppc/spapr.c +27
-0@@ -1776,6 +1776,10 @@ static void spapr_machine_reset(MachineState *machine)*/ object_child_foreach_recursive (object_get_root (), spapr_reset_drcs,NULL ); +if (spapr->tpm_device_file) { +spapr_hcall_tpm_reset (); + } +spapr_clear_pending_events (spapr);/* @@ -3340,6 +3344,21 @@ static void spapr_set_host_serial(Object *obj, const char *value, Error **errp) spapr->host_serial =g_strdup (value); } +static char *spapr_get_tpm_device_file (Object *obj,Error **errp) +{ +SpaprMachineState *spapr =SPAPR_MACHINE (obj); + +return g_strdup (spapr->tpm_device_file); +} + +static void spapr_set_tpm_device_file (Object *obj,const char *value,Error **errp) +{ +SpaprMachineState *spapr =SPAPR_MACHINE (obj); + +g_free (spapr->tpm_device_file); + spapr->tpm_device_file =g_strdup (value); +} +static void spapr_instance_init (Object *obj) {SpaprMachineState *spapr =SPAPR_MACHINE (obj);@@ -3396,6 +3415,14 @@ static void spapr_instance_init(Object *obj) &error_abort);object_property_set_description (obj,"host-serial" ,"Host serial number to advertise in guest device tree" , &error_abort); +object_property_add_str (obj,"tpm-device-file" , + spapr_get_tpm_device_file, + spapr_set_tpm_device_file, &error_abort); +object_property_set_description (obj,"tpm-device-file" , +"Specifies the path to the TPM character device file to use" +" for TPM communication via hypercalls (usually a TPM" +" resource manager)" , + &error_abort); }static void spapr_machine_finalizefn (Object *obj)hw/ppc/spapr_hcall_tpm.c new file +18
-0@@ -0,0 +1,135 @@ +/* +* SPAPR TPM Hypercall +* +* Copyright IBM Corp. 2019 +* +* Authors: +* Michael Roth <[email protected]> +* +* This work is licensed under the terms of the GNU GPL, version 2 or later. +* See the COPYING file in the top-level directory. +*/ + +#include "qemu/osdep.h" +#include "qemu-common.h" +#include "qapi/error.h" +#include "qemu/error-report.h" +#include "cpu.h" +#include "hw/ppc/spapr.h"
hw/ppc/spapr_hcall_tpm.c line 28
Quoted 10 lines
hw/ppc/spapr_hcall_tpm.c new file +10
-0+#include "trace.h" + +#define TPM_SPAPR_BUFSIZE 4096 + +enum { +TPM_COMM_OP_EXECUTE =1 , +TPM_COMM_OP_CLOSE_SESSION =2 , +}; + +static int tpm_devfd =-1 ;
A global? Really? You can do better.
Quoted 155 lines
hw/ppc/spapr_hcall_tpm.c new file +107
-0+ +static ssize_t tpm_execute (SpaprMachineState *spapr,target_ulong *args) +{ +uint64_t data_in =ppc64_phys_to_real (args[1 ]); +target_ulong data_in_size = args[2 ]; +uint64_t data_out =ppc64_phys_to_real (args[3 ]); +target_ulong data_out_size = args[4 ]; +uint8_t buf_in[TPM_SPAPR_BUFSIZE ]; +uint8_t buf_out[TPM_SPAPR_BUFSIZE ]; +ssize_t ret; + +trace_spapr_tpm_execute (data_in, data_in_size, data_out, data_out_size); + +if (data_in_size >TPM_SPAPR_BUFSIZE ) { +error_report ("invalid TPM input buffer size: " TARGET_FMT_lu"\n" , + data_in_size); +return H_P3 ; + } + +if (data_out_size <TPM_SPAPR_BUFSIZE ) { +error_report ("invalid TPM output buffer size: " TARGET_FMT_lu"\n" , + data_out_size); +return H_P5 ; + } + +if (tpm_devfd ==-1 ) { + tpm_devfd =open (spapr->tpm_device_file,O_RDWR ); +if (tpm_devfd ==-1 ) { +error_report ("failed to open TPM device %s: %d" , + spapr->tpm_device_file, errno); +return H_RESOURCE ; + } + } + +cpu_physical_memory_read (data_in, buf_in, data_in_size); + +do { + ret =write (tpm_devfd, buf_in, data_in_size); +if (ret >0 ) { + data_in_size -= ret; + } + }while ((ret >=0 && data_in_size >0 ) || (ret ==-1 && errno ==EINTR )); + +if (ret ==-1 ) { +error_report ("failed to write to TPM device %s: %d" , + spapr->tpm_device_file, errno); +return H_RESOURCE ; + } + +do { + ret =read (tpm_devfd, buf_out, data_out_size); + }while (ret ==0 || (ret ==-1 && errno ==EINTR )); + +if (ret ==-1 ) { +error_report ("failed to read from TPM device %s: %d" , + spapr->tpm_device_file, errno); +return H_RESOURCE ; + } + +cpu_physical_memory_write (data_out, buf_out, ret); + args[0 ] = ret; + +return H_SUCCESS ; +} + +static target_ulong h_tpm_comm (PowerPCCPU *cpu, +SpaprMachineState *spapr, +target_ulong opcode, +target_ulong *args) +{ +target_ulong op = args[0 ]; + +trace_spapr_h_tpm_comm (spapr->tpm_device_file ?:"null" , op); + +if (!spapr->tpm_device_file) { +error_report ("TPM device not specified" ); +return H_RESOURCE ; + } + +switch (op) { +case TPM_COMM_OP_EXECUTE : +return tpm_execute (spapr, args); +case TPM_COMM_OP_CLOSE_SESSION : +if (tpm_devfd !=-1 ) { +close (tpm_devfd); + tpm_devfd =-1 ; + } +return H_SUCCESS ; +default : +return H_PARAMETER ; + } +} + +void spapr_hcall_tpm_reset (void ) +{ +if (tpm_devfd !=-1 ) { +close (tpm_devfd); + tpm_devfd =-1 ; + } +} + +static void hypercall_register_types (void ) +{ +spapr_register_hypercall (H_TPM_COMM , h_tpm_comm); +} + +type_init (hypercall_register_types)hw/ppc/trace-events +4
-0@@ -25,6 +25,10 @@ spapr_update_dt(unsigned cb) "New blob %u bytes" spapr_update_dt_failed_size(unsigned cbold, unsigned cbnew, unsigned magic) "Old blob %u bytes, new blob %u bytes, magic 0x%x" spapr_update_dt_failed_check(unsigned cbold, unsigned cbnew, unsigned magic) "Old blob %u bytes, new blob %u bytes, magic 0x%x" +# spapr_hcall_tpm.c +spapr_h_tpm_comm(const char *device_path, uint64_t operation) "tpm_device_path=%s operation=0x%"PRIu64 +spapr_tpm_execute(uint64_t data_in, uint64_t data_in_sz, uint64_t data_out, uint64_t data_out_sz) "data_in=0x%"PRIx64", data_in_sz=%"PRIu64", data_out=0x%"PRIx64", data_out_sz=%"PRIu64 + # spapr_iommu.c spapr_iommu_put(uint64_t liobn, uint64_t ioba, uint64_t tce, uint64_t ret) "liobn=0x%"PRIx64" ioba=0x%"PRIx64" tce=0x%"PRIx64" ret=%"PRId64 spapr_iommu_get(uint64_t liobn, uint64_t ioba, uint64_t ret, uint64_t tce) "liobn=0x%"PRIx64" ioba=0x%"PRIx64" ret=%"PRId64" tce=0x%"PRIx64include/hw/ppc/spapr.h +6
-1@@ -198,6 +198,7 @@ struct SpaprMachineState {SpaprXive *xive;SpaprIrq *irq;qemu_irq *qirqs; +char *tpm_device_file;bool cmd_line_caps[SPAPR_CAP_NUM ];SpaprCapabilities def, eff, mig;@@ -490,8 +491,9 @@ struct SpaprMachineState {#define H_INT_ESB 0x3C8#define H_INT_SYNC 0x3CC#define H_INT_RESET 0x3D0 +#define H_TPM_COMM 0xEF10 -#define MAX_HCALL_OPCODE H_INT_RESET +#define MAX_HCALL_OPCODE H_TPM_COMM/* The hcalls above are standardized in PAPR and implemented by pHyp * as well. @@ -864,6 +866,9 @@ int spapr_caps_post_migration(SpaprMachineState *spapr);void spapr_check_pagesize (SpaprMachineState *spapr,hwaddr pagesize,Error **errp); + +void spapr_hcall_tpm_reset (void ); +/* * XIVE definitions */
Signature · 4 lines
David Gibson | I'll have my music baroque, and my code david AT gibson.dropbear.id.au | minimalist, thank you. NOT _the_ _other_ | _way_ _around_! http://www.ozlabs.org/~dgibson
Re: [Qemu-devel] [PATCH 2/2] spapr: initial implementation for H_TPM_COMM hcall
Quoting David Gibson (2019-07-12 01:46:19)
Quoted 8 lines · Michael Roth
Quoted 7 lines
This implements the
H_TPM_COMMhypercall, which is used by an Ultravisor to pass TPM commands directly to the host's TPM device, or a TPM Resource Manager associated with the device.This also introduces a new pseries machine option which is used to configure what TPM device to pass commands to, for example:
Quoted 6 lines
Quoted 1 line
-machine pseries,...,tpm-device-file=/dev/tmprm0
Bolting this into yet another machine parameter seems kind of ugly. Wouldn't it make more sense to treat this as an virtual device (say "spapr-vtpm"). Adding that device would enable the hcall, and would have properties for the back end host device.
That does sound nicer.
Originally I had SpaprMachineClass implement the TYPE_TPM_IF interface so we could define a TPM backend via -tpmdev passthrough,path=..., but after some discussion with the TPM maintainer it didn't quite work for the main use-case of passing through a TPM Resource Manager since it isn't suitable for full vTPM front-ends (since multiple guests can interfere with each other's operations when running the full gamut of TPM functionality).
I hadn't consider a stand-alone -device implementation though. It's not a proper VIO or PCI device so there's no proper bus to attach it to. I guess we would just make it a direct child of SpaprMachineState (sort of like SpaprDrcClass), then we could define it via something like -object spapr-tpm-proxy,path=....
I'll go ahead and give that a shot, assuming it seems reasonable to you.
Quoted 5 lines
Quoted 3 lines
By default, no tpm-device-file is defined and hcalls will return
H_RESOURCE.Wouldn't
H_FUNCTIONmake more sense?
Yes, for this case it probably would.
Thanks for the suggestions!
Quoted 277 lines
Quoted 103 lines
The full specification for this hypercall can be found in
docs/specs/ppc-spapr-uv-hcalls.txtSigned-off-by Michael Roth <[email protected]--- hw/ppc/Makefile.objs | 1 + hw/ppc/spapr.c | 27 ++++++++ hw/ppc/spapr_hcall_tpm.c | 135 +++++++++++++++++++++++++++++++++++++++ hw/ppc/trace-events | 4 ++ include/hw/ppc/spapr.h | 7 +- 5 files changed, 173 insertions(+), 1 deletion(-) create mode 100644 hw/ppc/spapr_hcall_tpm.chw/ppc/Makefile.objs +1
-0@@ -5,6 +5,7 @@ obj-$(CONFIG_PSERIES) += spapr.o spapr_caps.o spapr_vio.o spapr_events.o obj-$(CONFIG_PSERIES) += spapr_hcall.o spapr_iommu.o spapr_rtas.o obj-$(CONFIG_PSERIES) += spapr_pci.o spapr_rtc.o spapr_drc.o obj-$(CONFIG_PSERIES) += spapr_cpu_core.o spapr_ovec.o spapr_irq.o +obj-$(CONFIG_PSERIES) += spapr_hcall_tpm.o obj-$(CONFIG_SPAPR_RNG) += spapr_rng.o # IBM PowerNV obj-$(CONFIG_POWERNV) += pnv.o pnv_xscom.o pnv_core.o pnv_lpc.o pnv_psi.o pnv_occ.o pnv_bmc.ohw/ppc/spapr.c +27
-0@@ -1776,6 +1776,10 @@ static void spapr_machine_reset(MachineState *machine)*/ object_child_foreach_recursive (object_get_root (), spapr_reset_drcs,NULL ); +if (spapr->tpm_device_file) { +spapr_hcall_tpm_reset (); + } +spapr_clear_pending_events (spapr);/* @@ -3340,6 +3344,21 @@ static void spapr_set_host_serial(Object *obj, const char *value, Error **errp) spapr->host_serial =g_strdup (value); } +static char *spapr_get_tpm_device_file (Object *obj,Error **errp) +{ +SpaprMachineState *spapr =SPAPR_MACHINE (obj); + +return g_strdup (spapr->tpm_device_file); +} + +static void spapr_set_tpm_device_file (Object *obj,const char *value,Error **errp) +{ +SpaprMachineState *spapr =SPAPR_MACHINE (obj); + +g_free (spapr->tpm_device_file); + spapr->tpm_device_file =g_strdup (value); +} +static void spapr_instance_init (Object *obj) {SpaprMachineState *spapr =SPAPR_MACHINE (obj);@@ -3396,6 +3415,14 @@ static void spapr_instance_init(Object *obj) &error_abort);object_property_set_description (obj,"host-serial" ,"Host serial number to advertise in guest device tree" , &error_abort); +object_property_add_str (obj,"tpm-device-file" , + spapr_get_tpm_device_file, + spapr_set_tpm_device_file, &error_abort); +object_property_set_description (obj,"tpm-device-file" , +"Specifies the path to the TPM character device file to use" +" for TPM communication via hypercalls (usually a TPM" +" resource manager)" , + &error_abort); }static void spapr_machine_finalizefn (Object *obj)hw/ppc/spapr_hcall_tpm.c new file +18
-0@@ -0,0 +1,135 @@ +/* +* SPAPR TPM Hypercall +* +* Copyright IBM Corp. 2019 +* +* Authors: +* Michael Roth <[email protected]> +* +* This work is licensed under the terms of the GNU GPL, version 2 or later. +* See the COPYING file in the top-level directory. +*/ + +#include "qemu/osdep.h" +#include "qemu-common.h" +#include "qapi/error.h" +#include "qemu/error-report.h" +#include "cpu.h" +#include "hw/ppc/spapr.h" hw/ppc/spapr_hcall_tpm.c line 28
Quoted 10 lines
hw/ppc/spapr_hcall_tpm.c new file +10
-0+#include "trace.h" + +#define TPM_SPAPR_BUFSIZE 4096 + +enum { +TPM_COMM_OP_EXECUTE =1 , +TPM_COMM_OP_CLOSE_SESSION =2 , +}; + +static int tpm_devfd =-1 ;A global? Really? You can do better.
Quoted 155 lines
hw/ppc/spapr_hcall_tpm.c new file +107
-0+ +static ssize_t tpm_execute (SpaprMachineState *spapr,target_ulong *args) +{ +uint64_t data_in =ppc64_phys_to_real (args[1 ]); +target_ulong data_in_size = args[2 ]; +uint64_t data_out =ppc64_phys_to_real (args[3 ]); +target_ulong data_out_size = args[4 ]; +uint8_t buf_in[TPM_SPAPR_BUFSIZE ]; +uint8_t buf_out[TPM_SPAPR_BUFSIZE ]; +ssize_t ret; + +trace_spapr_tpm_execute (data_in, data_in_size, data_out, data_out_size); + +if (data_in_size >TPM_SPAPR_BUFSIZE ) { +error_report ("invalid TPM input buffer size: " TARGET_FMT_lu"\n" , + data_in_size); +return H_P3 ; + } + +if (data_out_size <TPM_SPAPR_BUFSIZE ) { +error_report ("invalid TPM output buffer size: " TARGET_FMT_lu"\n" , + data_out_size); +return H_P5 ; + } + +if (tpm_devfd ==-1 ) { + tpm_devfd =open (spapr->tpm_device_file,O_RDWR ); +if (tpm_devfd ==-1 ) { +error_report ("failed to open TPM device %s: %d" , + spapr->tpm_device_file, errno); +return H_RESOURCE ; + } + } + +cpu_physical_memory_read (data_in, buf_in, data_in_size); + +do { + ret =write (tpm_devfd, buf_in, data_in_size); +if (ret >0 ) { + data_in_size -= ret; + } + }while ((ret >=0 && data_in_size >0 ) || (ret ==-1 && errno ==EINTR )); + +if (ret ==-1 ) { +error_report ("failed to write to TPM device %s: %d" , + spapr->tpm_device_file, errno); +return H_RESOURCE ; + } + +do { + ret =read (tpm_devfd, buf_out, data_out_size); + }while (ret ==0 || (ret ==-1 && errno ==EINTR )); + +if (ret ==-1 ) { +error_report ("failed to read from TPM device %s: %d" , + spapr->tpm_device_file, errno); +return H_RESOURCE ; + } + +cpu_physical_memory_write (data_out, buf_out, ret); + args[0 ] = ret; + +return H_SUCCESS ; +} + +static target_ulong h_tpm_comm (PowerPCCPU *cpu, +SpaprMachineState *spapr, +target_ulong opcode, +target_ulong *args) +{ +target_ulong op = args[0 ]; + +trace_spapr_h_tpm_comm (spapr->tpm_device_file ?:"null" , op); + +if (!spapr->tpm_device_file) { +error_report ("TPM device not specified" ); +return H_RESOURCE ; + } + +switch (op) { +case TPM_COMM_OP_EXECUTE : +return tpm_execute (spapr, args); +case TPM_COMM_OP_CLOSE_SESSION : +if (tpm_devfd !=-1 ) { +close (tpm_devfd); + tpm_devfd =-1 ; + } +return H_SUCCESS ; +default : +return H_PARAMETER ; + } +} + +void spapr_hcall_tpm_reset (void ) +{ +if (tpm_devfd !=-1 ) { +close (tpm_devfd); + tpm_devfd =-1 ; + } +} + +static void hypercall_register_types (void ) +{ +spapr_register_hypercall (H_TPM_COMM , h_tpm_comm); +} + +type_init (hypercall_register_types)hw/ppc/trace-events +4
-0@@ -25,6 +25,10 @@ spapr_update_dt(unsigned cb) "New blob %u bytes" spapr_update_dt_failed_size(unsigned cbold, unsigned cbnew, unsigned magic) "Old blob %u bytes, new blob %u bytes, magic 0x%x" spapr_update_dt_failed_check(unsigned cbold, unsigned cbnew, unsigned magic) "Old blob %u bytes, new blob %u bytes, magic 0x%x" +# spapr_hcall_tpm.c +spapr_h_tpm_comm(const char *device_path, uint64_t operation) "tpm_device_path=%s operation=0x%"PRIu64 +spapr_tpm_execute(uint64_t data_in, uint64_t data_in_sz, uint64_t data_out, uint64_t data_out_sz) "data_in=0x%"PRIx64", data_in_sz=%"PRIu64", data_out=0x%"PRIx64", data_out_sz=%"PRIu64 + # spapr_iommu.c spapr_iommu_put(uint64_t liobn, uint64_t ioba, uint64_t tce, uint64_t ret) "liobn=0x%"PRIx64" ioba=0x%"PRIx64" tce=0x%"PRIx64" ret=%"PRId64 spapr_iommu_get(uint64_t liobn, uint64_t ioba, uint64_t ret, uint64_t tce) "liobn=0x%"PRIx64" ioba=0x%"PRIx64" ret=%"PRId64" tce=0x%"PRIx64include/hw/ppc/spapr.h +6
-1@@ -198,6 +198,7 @@ struct SpaprMachineState {SpaprXive *xive;SpaprIrq *irq;qemu_irq *qirqs; +char *tpm_device_file;bool cmd_line_caps[SPAPR_CAP_NUM ];SpaprCapabilities def, eff, mig;@@ -490,8 +491,9 @@ struct SpaprMachineState {#define H_INT_ESB 0x3C8#define H_INT_SYNC 0x3CC#define H_INT_RESET 0x3D0 +#define H_TPM_COMM 0xEF10 -#define MAX_HCALL_OPCODE H_INT_RESET +#define MAX_HCALL_OPCODE H_TPM_COMM/* The hcalls above are standardized in PAPR and implemented by pHyp * as well. @@ -864,6 +866,9 @@ int spapr_caps_post_migration(SpaprMachineState *spapr);void spapr_check_pagesize (SpaprMachineState *spapr,hwaddr pagesize,Error **errp); + +void spapr_hcall_tpm_reset (void ); +/* * XIVE definitions */ Signature · 4 lines
David Gibson | I'll have my music baroque, and my code david AT gibson.dropbear.id.au | minimalist, thank you. NOT _the_ _other_ | _way_ _around_! http://www.ozlabs.org/~dgibson
Re: [Qemu-devel] [PATCH 2/2] spapr: initial implementation for H_TPM_COMM hcall
Quoted 25 lines · Michael Roth
Quoting David Gibson (2019-07-12 01:46:19)
Quoted 8 lines · Michael Roth
Quoted 7 lines
This implements the
H_TPM_COMMhypercall, which is used by an Ultravisor to pass TPM commands directly to the host's TPM device, or a TPM Resource Manager associated with the device.This also introduces a new pseries machine option which is used to configure what TPM device to pass commands to, for example:
Quoted 6 lines
Quoted 1 line
-machine pseries,...,tpm-device-file=/dev/tmprm0
Bolting this into yet another machine parameter seems kind of ugly. Wouldn't it make more sense to treat this as an virtual device (say "spapr-vtpm"). Adding that device would enable the hcall, and would have properties for the back end host device.
That does sound nicer.
Originally I had SpaprMachineClass implement the
TYPE_TPM_IFinterface so we could define a TPM backend via -tpmdev passthrough,path=..., but after some discussion with the TPM maintainer it didn't quite work for the main use-case of passing through a TPM Resource Manager since it isn't suitable for full vTPM front-ends (since multiple guests can interfere with each other's operations when running the full gamut of TPM functionality).
Quoted 5 lines
I hadn't consider a stand-alone -device implementation though. It's not a proper VIO or PCI device so there's no proper bus to attach it to. I guess we would just make it a direct child of SpaprMachineState (sort of like SpaprDrcClass), then we could define it via something like -object spapr-tpm-proxy,path=....
It should be -device not -object, but otherwise that looks ok.
How does the TPM appear in the device tree?
Quoted 285 lines
I'll go ahead and give that a shot, assuming it seems reasonable to you.
Quoted 5 lines
Quoted 3 lines
By default, no tpm-device-file is defined and hcalls will return
H_RESOURCE.Wouldn't
H_FUNCTIONmake more sense?Yes, for this case it probably would.
Thanks for the suggestions!
Quoted 272 lines
Quoted 103 lines
The full specification for this hypercall can be found in
docs/specs/ppc-spapr-uv-hcalls.txtSigned-off-by Michael Roth <[email protected]--- hw/ppc/Makefile.objs | 1 + hw/ppc/spapr.c | 27 ++++++++ hw/ppc/spapr_hcall_tpm.c | 135 +++++++++++++++++++++++++++++++++++++++ hw/ppc/trace-events | 4 ++ include/hw/ppc/spapr.h | 7 +- 5 files changed, 173 insertions(+), 1 deletion(-) create mode 100644 hw/ppc/spapr_hcall_tpm.chw/ppc/Makefile.objs +1
-0@@ -5,6 +5,7 @@ obj-$(CONFIG_PSERIES) += spapr.o spapr_caps.o spapr_vio.o spapr_events.o obj-$(CONFIG_PSERIES) += spapr_hcall.o spapr_iommu.o spapr_rtas.o obj-$(CONFIG_PSERIES) += spapr_pci.o spapr_rtc.o spapr_drc.o obj-$(CONFIG_PSERIES) += spapr_cpu_core.o spapr_ovec.o spapr_irq.o +obj-$(CONFIG_PSERIES) += spapr_hcall_tpm.o obj-$(CONFIG_SPAPR_RNG) += spapr_rng.o # IBM PowerNV obj-$(CONFIG_POWERNV) += pnv.o pnv_xscom.o pnv_core.o pnv_lpc.o pnv_psi.o pnv_occ.o pnv_bmc.ohw/ppc/spapr.c +27
-0@@ -1776,6 +1776,10 @@ static void spapr_machine_reset(MachineState *machine)*/ object_child_foreach_recursive (object_get_root (), spapr_reset_drcs,NULL ); +if (spapr->tpm_device_file) { +spapr_hcall_tpm_reset (); + } +spapr_clear_pending_events (spapr);/* @@ -3340,6 +3344,21 @@ static void spapr_set_host_serial(Object *obj, const char *value, Error **errp) spapr->host_serial =g_strdup (value); } +static char *spapr_get_tpm_device_file (Object *obj,Error **errp) +{ +SpaprMachineState *spapr =SPAPR_MACHINE (obj); + +return g_strdup (spapr->tpm_device_file); +} + +static void spapr_set_tpm_device_file (Object *obj,const char *value,Error **errp) +{ +SpaprMachineState *spapr =SPAPR_MACHINE (obj); + +g_free (spapr->tpm_device_file); + spapr->tpm_device_file =g_strdup (value); +} +static void spapr_instance_init (Object *obj) {SpaprMachineState *spapr =SPAPR_MACHINE (obj);@@ -3396,6 +3415,14 @@ static void spapr_instance_init(Object *obj) &error_abort);object_property_set_description (obj,"host-serial" ,"Host serial number to advertise in guest device tree" , &error_abort); +object_property_add_str (obj,"tpm-device-file" , + spapr_get_tpm_device_file, + spapr_set_tpm_device_file, &error_abort); +object_property_set_description (obj,"tpm-device-file" , +"Specifies the path to the TPM character device file to use" +" for TPM communication via hypercalls (usually a TPM" +" resource manager)" , + &error_abort); }static void spapr_machine_finalizefn (Object *obj)hw/ppc/spapr_hcall_tpm.c new file +18
-0@@ -0,0 +1,135 @@ +/* +* SPAPR TPM Hypercall +* +* Copyright IBM Corp. 2019 +* +* Authors: +* Michael Roth <[email protected]> +* +* This work is licensed under the terms of the GNU GPL, version 2 or later. +* See the COPYING file in the top-level directory. +*/ + +#include "qemu/osdep.h" +#include "qemu-common.h" +#include "qapi/error.h" +#include "qemu/error-report.h" +#include "cpu.h" +#include "hw/ppc/spapr.h" hw/ppc/spapr_hcall_tpm.c line 28
Quoted 10 lines
hw/ppc/spapr_hcall_tpm.c new file +10
-0+#include "trace.h" + +#define TPM_SPAPR_BUFSIZE 4096 + +enum { +TPM_COMM_OP_EXECUTE =1 , +TPM_COMM_OP_CLOSE_SESSION =2 , +}; + +static int tpm_devfd =-1 ;A global? Really? You can do better.
Quoted 155 lines
hw/ppc/spapr_hcall_tpm.c new file +107
-0+ +static ssize_t tpm_execute (SpaprMachineState *spapr,target_ulong *args) +{ +uint64_t data_in =ppc64_phys_to_real (args[1 ]); +target_ulong data_in_size = args[2 ]; +uint64_t data_out =ppc64_phys_to_real (args[3 ]); +target_ulong data_out_size = args[4 ]; +uint8_t buf_in[TPM_SPAPR_BUFSIZE ]; +uint8_t buf_out[TPM_SPAPR_BUFSIZE ]; +ssize_t ret; + +trace_spapr_tpm_execute (data_in, data_in_size, data_out, data_out_size); + +if (data_in_size >TPM_SPAPR_BUFSIZE ) { +error_report ("invalid TPM input buffer size: " TARGET_FMT_lu"\n" , + data_in_size); +return H_P3 ; + } + +if (data_out_size <TPM_SPAPR_BUFSIZE ) { +error_report ("invalid TPM output buffer size: " TARGET_FMT_lu"\n" , + data_out_size); +return H_P5 ; + } + +if (tpm_devfd ==-1 ) { + tpm_devfd =open (spapr->tpm_device_file,O_RDWR ); +if (tpm_devfd ==-1 ) { +error_report ("failed to open TPM device %s: %d" , + spapr->tpm_device_file, errno); +return H_RESOURCE ; + } + } + +cpu_physical_memory_read (data_in, buf_in, data_in_size); + +do { + ret =write (tpm_devfd, buf_in, data_in_size); +if (ret >0 ) { + data_in_size -= ret; + } + }while ((ret >=0 && data_in_size >0 ) || (ret ==-1 && errno ==EINTR )); + +if (ret ==-1 ) { +error_report ("failed to write to TPM device %s: %d" , + spapr->tpm_device_file, errno); +return H_RESOURCE ; + } + +do { + ret =read (tpm_devfd, buf_out, data_out_size); + }while (ret ==0 || (ret ==-1 && errno ==EINTR )); + +if (ret ==-1 ) { +error_report ("failed to read from TPM device %s: %d" , + spapr->tpm_device_file, errno); +return H_RESOURCE ; + } + +cpu_physical_memory_write (data_out, buf_out, ret); + args[0 ] = ret; + +return H_SUCCESS ; +} + +static target_ulong h_tpm_comm (PowerPCCPU *cpu, +SpaprMachineState *spapr, +target_ulong opcode, +target_ulong *args) +{ +target_ulong op = args[0 ]; + +trace_spapr_h_tpm_comm (spapr->tpm_device_file ?:"null" , op); + +if (!spapr->tpm_device_file) { +error_report ("TPM device not specified" ); +return H_RESOURCE ; + } + +switch (op) { +case TPM_COMM_OP_EXECUTE : +return tpm_execute (spapr, args); +case TPM_COMM_OP_CLOSE_SESSION : +if (tpm_devfd !=-1 ) { +close (tpm_devfd); + tpm_devfd =-1 ; + } +return H_SUCCESS ; +default : +return H_PARAMETER ; + } +} + +void spapr_hcall_tpm_reset (void ) +{ +if (tpm_devfd !=-1 ) { +close (tpm_devfd); + tpm_devfd =-1 ; + } +} + +static void hypercall_register_types (void ) +{ +spapr_register_hypercall (H_TPM_COMM , h_tpm_comm); +} + +type_init (hypercall_register_types)hw/ppc/trace-events +4
-0@@ -25,6 +25,10 @@ spapr_update_dt(unsigned cb) "New blob %u bytes" spapr_update_dt_failed_size(unsigned cbold, unsigned cbnew, unsigned magic) "Old blob %u bytes, new blob %u bytes, magic 0x%x" spapr_update_dt_failed_check(unsigned cbold, unsigned cbnew, unsigned magic) "Old blob %u bytes, new blob %u bytes, magic 0x%x" +# spapr_hcall_tpm.c +spapr_h_tpm_comm(const char *device_path, uint64_t operation) "tpm_device_path=%s operation=0x%"PRIu64 +spapr_tpm_execute(uint64_t data_in, uint64_t data_in_sz, uint64_t data_out, uint64_t data_out_sz) "data_in=0x%"PRIx64", data_in_sz=%"PRIu64", data_out=0x%"PRIx64", data_out_sz=%"PRIu64 + # spapr_iommu.c spapr_iommu_put(uint64_t liobn, uint64_t ioba, uint64_t tce, uint64_t ret) "liobn=0x%"PRIx64" ioba=0x%"PRIx64" tce=0x%"PRIx64" ret=%"PRId64 spapr_iommu_get(uint64_t liobn, uint64_t ioba, uint64_t ret, uint64_t tce) "liobn=0x%"PRIx64" ioba=0x%"PRIx64" ret=%"PRId64" tce=0x%"PRIx64include/hw/ppc/spapr.h +6
-1@@ -198,6 +198,7 @@ struct SpaprMachineState {SpaprXive *xive;SpaprIrq *irq;qemu_irq *qirqs; +char *tpm_device_file;bool cmd_line_caps[SPAPR_CAP_NUM ];SpaprCapabilities def, eff, mig;@@ -490,8 +491,9 @@ struct SpaprMachineState {#define H_INT_ESB 0x3C8#define H_INT_SYNC 0x3CC#define H_INT_RESET 0x3D0 +#define H_TPM_COMM 0xEF10 -#define MAX_HCALL_OPCODE H_INT_RESET +#define MAX_HCALL_OPCODE H_TPM_COMM/* The hcalls above are standardized in PAPR and implemented by pHyp * as well. @@ -864,6 +866,9 @@ int spapr_caps_post_migration(SpaprMachineState *spapr);void spapr_check_pagesize (SpaprMachineState *spapr,hwaddr pagesize,Error **errp); + +void spapr_hcall_tpm_reset (void ); +/* * XIVE definitions */
Signature · 4 lines
David Gibson | I'll have my music baroque, and my code david AT gibson.dropbear.id.au | minimalist, thank you. NOT _the_ _other_ | _way_ _around_! http://www.ozlabs.org/~dgibson
Re: [Qemu-devel] [PATCH 2/2] spapr: initial implementation for H_TPM_COMM hcall
Quoting David Gibson (2019-07-14 21:25:24)
Quoted 26 lines · Michael Roth
Quoted 25 lines
Quoting David Gibson (2019-07-12 01:46:19)
Quoted 8 lines · Michael Roth
Quoted 7 lines
This implements the
H_TPM_COMMhypercall, which is used by an Ultravisor to pass TPM commands directly to the host's TPM device, or a TPM Resource Manager associated with the device.This also introduces a new pseries machine option which is used to configure what TPM device to pass commands to, for example:
Quoted 6 lines
Quoted 1 line
-machine pseries,...,tpm-device-file=/dev/tmprm0
Bolting this into yet another machine parameter seems kind of ugly. Wouldn't it make more sense to treat this as an virtual device (say "spapr-vtpm"). Adding that device would enable the hcall, and would have properties for the back end host device.
That does sound nicer.
Originally I had SpaprMachineClass implement the
TYPE_TPM_IFinterface so we could define a TPM backend via -tpmdev passthrough,path=..., but after some discussion with the TPM maintainer it didn't quite work for the main use-case of passing through a TPM Resource Manager since it isn't suitable for full vTPM front-ends (since multiple guests can interfere with each other's operations when running the full gamut of TPM functionality).
Quoted 7 lines
Quoted 5 lines
I hadn't consider a stand-alone -device implementation though. It's not a proper VIO or PCI device so there's no proper bus to attach it to. I guess we would just make it a direct child of SpaprMachineState (sort of like SpaprDrcClass), then we could define it via something like -object spapr-tpm-proxy,path=....
It should be -device not -object, but otherwise that looks ok.
Ok, for some reason I thought -device needed either a specific bus or needed to be a SysBusDevice to attach to main-system-bus, but maybe that was just for qdev-managed reset handling. I've re-worked the series to allow configuration via:
-device spapr-tpm-proxy,host_path=/dev/tpmrmX
Quoted 2 lines
How does the TPM appear in the device tree?
Nothing in the guest, on the host it appears as:
./xscom@603fc00000000/i2cm@a2000/i2c-bus@0/tpm@57 ./xscom@603fc00000000/i2cm@a2000/i2c-bus@0/tpm@57/link-id ./xscom@603fc00000000/i2cm@a2000/i2c-bus@0/tpm@57/linux,sml-size ./xscom@603fc00000000/i2cm@a2000/i2c-bus@0/tpm@57/label ./xscom@603fc00000000/i2cm@a2000/i2c-bus@0/tpm@57/compatible ./xscom@603fc00000000/i2cm@a2000/i2c-bus@0/tpm@57/status ./xscom@603fc00000000/i2cm@a2000/i2c-bus@0/tpm@57/reg ./xscom@603fc00000000/i2cm@a2000/i2c-bus@0/tpm@57/phandle ./xscom@603fc00000000/i2cm@a2000/i2c-bus@0/tpm@57/linux,sml-base ./xscom@603fc00000000/i2cm@a2000/i2c-bus@0/tpm@57/name
Quoted 292 lines
Quoted 286 lines
I'll go ahead and give that a shot, assuming it seems reasonable to you.
Quoted 5 lines
Quoted 3 lines
By default, no tpm-device-file is defined and hcalls will return
H_RESOURCE.Wouldn't
H_FUNCTIONmake more sense?Yes, for this case it probably would.
Thanks for the suggestions!
Quoted 272 lines
Quoted 103 lines
The full specification for this hypercall can be found in
docs/specs/ppc-spapr-uv-hcalls.txtSigned-off-by Michael Roth <[email protected]--- hw/ppc/Makefile.objs | 1 + hw/ppc/spapr.c | 27 ++++++++ hw/ppc/spapr_hcall_tpm.c | 135 +++++++++++++++++++++++++++++++++++++++ hw/ppc/trace-events | 4 ++ include/hw/ppc/spapr.h | 7 +- 5 files changed, 173 insertions(+), 1 deletion(-) create mode 100644 hw/ppc/spapr_hcall_tpm.chw/ppc/Makefile.objs +1
-0@@ -5,6 +5,7 @@ obj-$(CONFIG_PSERIES) += spapr.o spapr_caps.o spapr_vio.o spapr_events.o obj-$(CONFIG_PSERIES) += spapr_hcall.o spapr_iommu.o spapr_rtas.o obj-$(CONFIG_PSERIES) += spapr_pci.o spapr_rtc.o spapr_drc.o obj-$(CONFIG_PSERIES) += spapr_cpu_core.o spapr_ovec.o spapr_irq.o +obj-$(CONFIG_PSERIES) += spapr_hcall_tpm.o obj-$(CONFIG_SPAPR_RNG) += spapr_rng.o # IBM PowerNV obj-$(CONFIG_POWERNV) += pnv.o pnv_xscom.o pnv_core.o pnv_lpc.o pnv_psi.o pnv_occ.o pnv_bmc.ohw/ppc/spapr.c +27
-0@@ -1776,6 +1776,10 @@ static void spapr_machine_reset(MachineState *machine)*/ object_child_foreach_recursive (object_get_root (), spapr_reset_drcs,NULL ); +if (spapr->tpm_device_file) { +spapr_hcall_tpm_reset (); + } +spapr_clear_pending_events (spapr);/* @@ -3340,6 +3344,21 @@ static void spapr_set_host_serial(Object *obj, const char *value, Error **errp) spapr->host_serial =g_strdup (value); } +static char *spapr_get_tpm_device_file (Object *obj,Error **errp) +{ +SpaprMachineState *spapr =SPAPR_MACHINE (obj); + +return g_strdup (spapr->tpm_device_file); +} + +static void spapr_set_tpm_device_file (Object *obj,const char *value,Error **errp) +{ +SpaprMachineState *spapr =SPAPR_MACHINE (obj); + +g_free (spapr->tpm_device_file); + spapr->tpm_device_file =g_strdup (value); +} +static void spapr_instance_init (Object *obj) {SpaprMachineState *spapr =SPAPR_MACHINE (obj);@@ -3396,6 +3415,14 @@ static void spapr_instance_init(Object *obj) &error_abort);object_property_set_description (obj,"host-serial" ,"Host serial number to advertise in guest device tree" , &error_abort); +object_property_add_str (obj,"tpm-device-file" , + spapr_get_tpm_device_file, + spapr_set_tpm_device_file, &error_abort); +object_property_set_description (obj,"tpm-device-file" , +"Specifies the path to the TPM character device file to use" +" for TPM communication via hypercalls (usually a TPM" +" resource manager)" , + &error_abort); }static void spapr_machine_finalizefn (Object *obj)hw/ppc/spapr_hcall_tpm.c new file +18
-0@@ -0,0 +1,135 @@ +/* +* SPAPR TPM Hypercall +* +* Copyright IBM Corp. 2019 +* +* Authors: +* Michael Roth <[email protected]> +* +* This work is licensed under the terms of the GNU GPL, version 2 or later. +* See the COPYING file in the top-level directory. +*/ + +#include "qemu/osdep.h" +#include "qemu-common.h" +#include "qapi/error.h" +#include "qemu/error-report.h" +#include "cpu.h" +#include "hw/ppc/spapr.h" hw/ppc/spapr_hcall_tpm.c line 28
Quoted 10 lines
hw/ppc/spapr_hcall_tpm.c new file +10
-0+#include "trace.h" + +#define TPM_SPAPR_BUFSIZE 4096 + +enum { +TPM_COMM_OP_EXECUTE =1 , +TPM_COMM_OP_CLOSE_SESSION =2 , +}; + +static int tpm_devfd =-1 ;A global? Really? You can do better.
Quoted 155 lines
hw/ppc/spapr_hcall_tpm.c new file +107
-0+ +static ssize_t tpm_execute (SpaprMachineState *spapr,target_ulong *args) +{ +uint64_t data_in =ppc64_phys_to_real (args[1 ]); +target_ulong data_in_size = args[2 ]; +uint64_t data_out =ppc64_phys_to_real (args[3 ]); +target_ulong data_out_size = args[4 ]; +uint8_t buf_in[TPM_SPAPR_BUFSIZE ]; +uint8_t buf_out[TPM_SPAPR_BUFSIZE ]; +ssize_t ret; + +trace_spapr_tpm_execute (data_in, data_in_size, data_out, data_out_size); + +if (data_in_size >TPM_SPAPR_BUFSIZE ) { +error_report ("invalid TPM input buffer size: " TARGET_FMT_lu"\n" , + data_in_size); +return H_P3 ; + } + +if (data_out_size <TPM_SPAPR_BUFSIZE ) { +error_report ("invalid TPM output buffer size: " TARGET_FMT_lu"\n" , + data_out_size); +return H_P5 ; + } + +if (tpm_devfd ==-1 ) { + tpm_devfd =open (spapr->tpm_device_file,O_RDWR ); +if (tpm_devfd ==-1 ) { +error_report ("failed to open TPM device %s: %d" , + spapr->tpm_device_file, errno); +return H_RESOURCE ; + } + } + +cpu_physical_memory_read (data_in, buf_in, data_in_size); + +do { + ret =write (tpm_devfd, buf_in, data_in_size); +if (ret >0 ) { + data_in_size -= ret; + } + }while ((ret >=0 && data_in_size >0 ) || (ret ==-1 && errno ==EINTR )); + +if (ret ==-1 ) { +error_report ("failed to write to TPM device %s: %d" , + spapr->tpm_device_file, errno); +return H_RESOURCE ; + } + +do { + ret =read (tpm_devfd, buf_out, data_out_size); + }while (ret ==0 || (ret ==-1 && errno ==EINTR )); + +if (ret ==-1 ) { +error_report ("failed to read from TPM device %s: %d" , + spapr->tpm_device_file, errno); +return H_RESOURCE ; + } + +cpu_physical_memory_write (data_out, buf_out, ret); + args[0 ] = ret; + +return H_SUCCESS ; +} + +static target_ulong h_tpm_comm (PowerPCCPU *cpu, +SpaprMachineState *spapr, +target_ulong opcode, +target_ulong *args) +{ +target_ulong op = args[0 ]; + +trace_spapr_h_tpm_comm (spapr->tpm_device_file ?:"null" , op); + +if (!spapr->tpm_device_file) { +error_report ("TPM device not specified" ); +return H_RESOURCE ; + } + +switch (op) { +case TPM_COMM_OP_EXECUTE : +return tpm_execute (spapr, args); +case TPM_COMM_OP_CLOSE_SESSION : +if (tpm_devfd !=-1 ) { +close (tpm_devfd); + tpm_devfd =-1 ; + } +return H_SUCCESS ; +default : +return H_PARAMETER ; + } +} + +void spapr_hcall_tpm_reset (void ) +{ +if (tpm_devfd !=-1 ) { +close (tpm_devfd); + tpm_devfd =-1 ; + } +} + +static void hypercall_register_types (void ) +{ +spapr_register_hypercall (H_TPM_COMM , h_tpm_comm); +} + +type_init (hypercall_register_types)hw/ppc/trace-events +4
-0@@ -25,6 +25,10 @@ spapr_update_dt(unsigned cb) "New blob %u bytes" spapr_update_dt_failed_size(unsigned cbold, unsigned cbnew, unsigned magic) "Old blob %u bytes, new blob %u bytes, magic 0x%x" spapr_update_dt_failed_check(unsigned cbold, unsigned cbnew, unsigned magic) "Old blob %u bytes, new blob %u bytes, magic 0x%x" +# spapr_hcall_tpm.c +spapr_h_tpm_comm(const char *device_path, uint64_t operation) "tpm_device_path=%s operation=0x%"PRIu64 +spapr_tpm_execute(uint64_t data_in, uint64_t data_in_sz, uint64_t data_out, uint64_t data_out_sz) "data_in=0x%"PRIx64", data_in_sz=%"PRIu64", data_out=0x%"PRIx64", data_out_sz=%"PRIu64 + # spapr_iommu.c spapr_iommu_put(uint64_t liobn, uint64_t ioba, uint64_t tce, uint64_t ret) "liobn=0x%"PRIx64" ioba=0x%"PRIx64" tce=0x%"PRIx64" ret=%"PRId64 spapr_iommu_get(uint64_t liobn, uint64_t ioba, uint64_t ret, uint64_t tce) "liobn=0x%"PRIx64" ioba=0x%"PRIx64" ret=%"PRId64" tce=0x%"PRIx64include/hw/ppc/spapr.h +6
-1@@ -198,6 +198,7 @@ struct SpaprMachineState {SpaprXive *xive;SpaprIrq *irq;qemu_irq *qirqs; +char *tpm_device_file;bool cmd_line_caps[SPAPR_CAP_NUM ];SpaprCapabilities def, eff, mig;@@ -490,8 +491,9 @@ struct SpaprMachineState {#define H_INT_ESB 0x3C8#define H_INT_SYNC 0x3CC#define H_INT_RESET 0x3D0 +#define H_TPM_COMM 0xEF10 -#define MAX_HCALL_OPCODE H_INT_RESET +#define MAX_HCALL_OPCODE H_TPM_COMM/* The hcalls above are standardized in PAPR and implemented by pHyp * as well. @@ -864,6 +866,9 @@ int spapr_caps_post_migration(SpaprMachineState *spapr);void spapr_check_pagesize (SpaprMachineState *spapr,hwaddr pagesize,Error **errp); + +void spapr_hcall_tpm_reset (void ); +/* * XIVE definitions */ Signature · 4 lines
David Gibson | I'll have my music baroque, and my code david AT gibson.dropbear.id.au | minimalist, thank you. NOT _the_ _other_ | _way_ _around_! http://www.ozlabs.org/~dgibson
Re: [Qemu-devel] [PATCH 2/2] spapr: initial implementation for H_TPM_COMM hcall
Quoted 33 lines · Michael Roth
Quoting David Gibson (2019-07-14 21:25:24)
Quoted 32 lines · Michael Roth
Quoted 30 lines
Quoting David Gibson (2019-07-12 01:46:19)
Quoted 8 lines · Michael Roth
Quoted 7 lines
This implements the
H_TPM_COMMhypercall, which is used by an Ultravisor to pass TPM commands directly to the host's TPM device, or a TPM Resource Manager associated with the device.This also introduces a new pseries machine option which is used to configure what TPM device to pass commands to, for example:
Quoted 6 lines
Quoted 1 line
-machine pseries,...,tpm-device-file=/dev/tmprm0
Bolting this into yet another machine parameter seems kind of ugly. Wouldn't it make more sense to treat this as an virtual device (say "spapr-vtpm"). Adding that device would enable the hcall, and would have properties for the back end host device.
That does sound nicer.
Originally I had SpaprMachineClass implement the
TYPE_TPM_IFinterface so we could define a TPM backend via -tpmdev passthrough,path=..., but after some discussion with the TPM maintainer it didn't quite work for the main use-case of passing through a TPM Resource Manager since it isn't suitable for full vTPM front-ends (since multiple guests can interfere with each other's operations when running the full gamut of TPM functionality).I hadn't consider a stand-alone -device implementation though. It's not a proper VIO or PCI device so there's no proper bus to attach it to. I guess we would just make it a direct child of SpaprMachineState (sort of like SpaprDrcClass), then we could define it via something like -object spapr-tpm-proxy,path=....
Quoted 8 lines
Quoted 1 line
It should be -device not -object, but otherwise that looks ok.
Ok, for some reason I thought -device needed either a specific bus or needed to be a SysBusDevice to attach to main-system-bus, but maybe that was just for qdev-managed reset handling. I've re-worked the series to allow configuration via:
-device spapr-tpm-proxy,host_path=/dev/tpmrmX
That looks good.
Quoted 3 lines
Quoted 1 line
How does the TPM appear in the device tree?
Nothing in the guest, on the host it appears as:
Hrm. That seems unwise. I mean, I guess its being treated as a hypervisor facility rather than a device per se, but what if we ever need to advertise more metadata about it.
Quoted 299 lines
./xscom@603fc00000000/i2cm@a2000/i2c-bus@0/tpm@57 ./xscom@603fc00000000/i2cm@a2000/i2c-bus@0/tpm@57/link-id ./xscom@603fc00000000/i2cm@a2000/i2c-bus@0/tpm@57/linux,sml-size ./xscom@603fc00000000/i2cm@a2000/i2c-bus@0/tpm@57/label ./xscom@603fc00000000/i2cm@a2000/i2c-bus@0/tpm@57/compatible ./xscom@603fc00000000/i2cm@a2000/i2c-bus@0/tpm@57/status ./xscom@603fc00000000/i2cm@a2000/i2c-bus@0/tpm@57/reg ./xscom@603fc00000000/i2cm@a2000/i2c-bus@0/tpm@57/phandle ./xscom@603fc00000000/i2cm@a2000/i2c-bus@0/tpm@57/linux,sml-base ./xscom@603fc00000000/i2cm@a2000/i2c-bus@0/tpm@57/name
Quoted 287 lines
Quoted 286 lines
I'll go ahead and give that a shot, assuming it seems reasonable to you.
Quoted 5 lines
Quoted 3 lines
By default, no tpm-device-file is defined and hcalls will return
H_RESOURCE.Wouldn't
H_FUNCTIONmake more sense?Yes, for this case it probably would.
Thanks for the suggestions!
Quoted 272 lines
Quoted 103 lines
The full specification for this hypercall can be found in
docs/specs/ppc-spapr-uv-hcalls.txtSigned-off-by Michael Roth <[email protected]--- hw/ppc/Makefile.objs | 1 + hw/ppc/spapr.c | 27 ++++++++ hw/ppc/spapr_hcall_tpm.c | 135 +++++++++++++++++++++++++++++++++++++++ hw/ppc/trace-events | 4 ++ include/hw/ppc/spapr.h | 7 +- 5 files changed, 173 insertions(+), 1 deletion(-) create mode 100644 hw/ppc/spapr_hcall_tpm.chw/ppc/Makefile.objs +1
-0@@ -5,6 +5,7 @@ obj-$(CONFIG_PSERIES) += spapr.o spapr_caps.o spapr_vio.o spapr_events.o obj-$(CONFIG_PSERIES) += spapr_hcall.o spapr_iommu.o spapr_rtas.o obj-$(CONFIG_PSERIES) += spapr_pci.o spapr_rtc.o spapr_drc.o obj-$(CONFIG_PSERIES) += spapr_cpu_core.o spapr_ovec.o spapr_irq.o +obj-$(CONFIG_PSERIES) += spapr_hcall_tpm.o obj-$(CONFIG_SPAPR_RNG) += spapr_rng.o # IBM PowerNV obj-$(CONFIG_POWERNV) += pnv.o pnv_xscom.o pnv_core.o pnv_lpc.o pnv_psi.o pnv_occ.o pnv_bmc.ohw/ppc/spapr.c +27
-0@@ -1776,6 +1776,10 @@ static void spapr_machine_reset(MachineState *machine)*/ object_child_foreach_recursive (object_get_root (), spapr_reset_drcs,NULL ); +if (spapr->tpm_device_file) { +spapr_hcall_tpm_reset (); + } +spapr_clear_pending_events (spapr);/* @@ -3340,6 +3344,21 @@ static void spapr_set_host_serial(Object *obj, const char *value, Error **errp) spapr->host_serial =g_strdup (value); } +static char *spapr_get_tpm_device_file (Object *obj,Error **errp) +{ +SpaprMachineState *spapr =SPAPR_MACHINE (obj); + +return g_strdup (spapr->tpm_device_file); +} + +static void spapr_set_tpm_device_file (Object *obj,const char *value,Error **errp) +{ +SpaprMachineState *spapr =SPAPR_MACHINE (obj); + +g_free (spapr->tpm_device_file); + spapr->tpm_device_file =g_strdup (value); +} +static void spapr_instance_init (Object *obj) {SpaprMachineState *spapr =SPAPR_MACHINE (obj);@@ -3396,6 +3415,14 @@ static void spapr_instance_init(Object *obj) &error_abort);object_property_set_description (obj,"host-serial" ,"Host serial number to advertise in guest device tree" , &error_abort); +object_property_add_str (obj,"tpm-device-file" , + spapr_get_tpm_device_file, + spapr_set_tpm_device_file, &error_abort); +object_property_set_description (obj,"tpm-device-file" , +"Specifies the path to the TPM character device file to use" +" for TPM communication via hypercalls (usually a TPM" +" resource manager)" , + &error_abort); }static void spapr_machine_finalizefn (Object *obj)hw/ppc/spapr_hcall_tpm.c new file +18
-0@@ -0,0 +1,135 @@ +/* +* SPAPR TPM Hypercall +* +* Copyright IBM Corp. 2019 +* +* Authors: +* Michael Roth <[email protected]> +* +* This work is licensed under the terms of the GNU GPL, version 2 or later. +* See the COPYING file in the top-level directory. +*/ + +#include "qemu/osdep.h" +#include "qemu-common.h" +#include "qapi/error.h" +#include "qemu/error-report.h" +#include "cpu.h" +#include "hw/ppc/spapr.h" hw/ppc/spapr_hcall_tpm.c line 28
Quoted 10 lines
hw/ppc/spapr_hcall_tpm.c new file +10
-0+#include "trace.h" + +#define TPM_SPAPR_BUFSIZE 4096 + +enum { +TPM_COMM_OP_EXECUTE =1 , +TPM_COMM_OP_CLOSE_SESSION =2 , +}; + +static int tpm_devfd =-1 ;A global? Really? You can do better.
Quoted 155 lines
hw/ppc/spapr_hcall_tpm.c new file +107
-0+ +static ssize_t tpm_execute (SpaprMachineState *spapr,target_ulong *args) +{ +uint64_t data_in =ppc64_phys_to_real (args[1 ]); +target_ulong data_in_size = args[2 ]; +uint64_t data_out =ppc64_phys_to_real (args[3 ]); +target_ulong data_out_size = args[4 ]; +uint8_t buf_in[TPM_SPAPR_BUFSIZE ]; +uint8_t buf_out[TPM_SPAPR_BUFSIZE ]; +ssize_t ret; + +trace_spapr_tpm_execute (data_in, data_in_size, data_out, data_out_size); + +if (data_in_size >TPM_SPAPR_BUFSIZE ) { +error_report ("invalid TPM input buffer size: " TARGET_FMT_lu"\n" , + data_in_size); +return H_P3 ; + } + +if (data_out_size <TPM_SPAPR_BUFSIZE ) { +error_report ("invalid TPM output buffer size: " TARGET_FMT_lu"\n" , + data_out_size); +return H_P5 ; + } + +if (tpm_devfd ==-1 ) { + tpm_devfd =open (spapr->tpm_device_file,O_RDWR ); +if (tpm_devfd ==-1 ) { +error_report ("failed to open TPM device %s: %d" , + spapr->tpm_device_file, errno); +return H_RESOURCE ; + } + } + +cpu_physical_memory_read (data_in, buf_in, data_in_size); + +do { + ret =write (tpm_devfd, buf_in, data_in_size); +if (ret >0 ) { + data_in_size -= ret; + } + }while ((ret >=0 && data_in_size >0 ) || (ret ==-1 && errno ==EINTR )); + +if (ret ==-1 ) { +error_report ("failed to write to TPM device %s: %d" , + spapr->tpm_device_file, errno); +return H_RESOURCE ; + } + +do { + ret =read (tpm_devfd, buf_out, data_out_size); + }while (ret ==0 || (ret ==-1 && errno ==EINTR )); + +if (ret ==-1 ) { +error_report ("failed to read from TPM device %s: %d" , + spapr->tpm_device_file, errno); +return H_RESOURCE ; + } + +cpu_physical_memory_write (data_out, buf_out, ret); + args[0 ] = ret; + +return H_SUCCESS ; +} + +static target_ulong h_tpm_comm (PowerPCCPU *cpu, +SpaprMachineState *spapr, +target_ulong opcode, +target_ulong *args) +{ +target_ulong op = args[0 ]; + +trace_spapr_h_tpm_comm (spapr->tpm_device_file ?:"null" , op); + +if (!spapr->tpm_device_file) { +error_report ("TPM device not specified" ); +return H_RESOURCE ; + } + +switch (op) { +case TPM_COMM_OP_EXECUTE : +return tpm_execute (spapr, args); +case TPM_COMM_OP_CLOSE_SESSION : +if (tpm_devfd !=-1 ) { +close (tpm_devfd); + tpm_devfd =-1 ; + } +return H_SUCCESS ; +default : +return H_PARAMETER ; + } +} + +void spapr_hcall_tpm_reset (void ) +{ +if (tpm_devfd !=-1 ) { +close (tpm_devfd); + tpm_devfd =-1 ; + } +} + +static void hypercall_register_types (void ) +{ +spapr_register_hypercall (H_TPM_COMM , h_tpm_comm); +} + +type_init (hypercall_register_types)hw/ppc/trace-events +4
-0@@ -25,6 +25,10 @@ spapr_update_dt(unsigned cb) "New blob %u bytes" spapr_update_dt_failed_size(unsigned cbold, unsigned cbnew, unsigned magic) "Old blob %u bytes, new blob %u bytes, magic 0x%x" spapr_update_dt_failed_check(unsigned cbold, unsigned cbnew, unsigned magic) "Old blob %u bytes, new blob %u bytes, magic 0x%x" +# spapr_hcall_tpm.c +spapr_h_tpm_comm(const char *device_path, uint64_t operation) "tpm_device_path=%s operation=0x%"PRIu64 +spapr_tpm_execute(uint64_t data_in, uint64_t data_in_sz, uint64_t data_out, uint64_t data_out_sz) "data_in=0x%"PRIx64", data_in_sz=%"PRIu64", data_out=0x%"PRIx64", data_out_sz=%"PRIu64 + # spapr_iommu.c spapr_iommu_put(uint64_t liobn, uint64_t ioba, uint64_t tce, uint64_t ret) "liobn=0x%"PRIx64" ioba=0x%"PRIx64" tce=0x%"PRIx64" ret=%"PRId64 spapr_iommu_get(uint64_t liobn, uint64_t ioba, uint64_t ret, uint64_t tce) "liobn=0x%"PRIx64" ioba=0x%"PRIx64" ret=%"PRId64" tce=0x%"PRIx64include/hw/ppc/spapr.h +6
-1@@ -198,6 +198,7 @@ struct SpaprMachineState {SpaprXive *xive;SpaprIrq *irq;qemu_irq *qirqs; +char *tpm_device_file;bool cmd_line_caps[SPAPR_CAP_NUM ];SpaprCapabilities def, eff, mig;@@ -490,8 +491,9 @@ struct SpaprMachineState {#define H_INT_ESB 0x3C8#define H_INT_SYNC 0x3CC#define H_INT_RESET 0x3D0 +#define H_TPM_COMM 0xEF10 -#define MAX_HCALL_OPCODE H_INT_RESET +#define MAX_HCALL_OPCODE H_TPM_COMM/* The hcalls above are standardized in PAPR and implemented by pHyp * as well. @@ -864,6 +866,9 @@ int spapr_caps_post_migration(SpaprMachineState *spapr);void spapr_check_pagesize (SpaprMachineState *spapr,hwaddr pagesize,Error **errp); + +void spapr_hcall_tpm_reset (void ); +/* * XIVE definitions */
Signature · 4 lines
David Gibson | I'll have my music baroque, and my code david AT gibson.dropbear.id.au | minimalist, thank you. NOT _the_ _other_ | _way_ _around_! http://www.ozlabs.org/~dgibson
Re: [Qemu-devel] [PATCH 2/2] spapr: initial implementation for H_TPM_COMM hcall
Quoting David Gibson (2019-07-16 20:29:12)
Quoted 45 lines · Michael Roth
Quoted 33 lines
Quoting David Gibson (2019-07-14 21:25:24)
Quoted 32 lines · Michael Roth
Quoted 30 lines
Quoting David Gibson (2019-07-12 01:46:19)
Quoted 8 lines · Michael Roth
Quoted 7 lines
This implements the
H_TPM_COMMhypercall, which is used by an Ultravisor to pass TPM commands directly to the host's TPM device, or a TPM Resource Manager associated with the device.This also introduces a new pseries machine option which is used to configure what TPM device to pass commands to, for example:
Quoted 6 lines
Quoted 1 line
-machine pseries,...,tpm-device-file=/dev/tmprm0
Bolting this into yet another machine parameter seems kind of ugly. Wouldn't it make more sense to treat this as an virtual device (say "spapr-vtpm"). Adding that device would enable the hcall, and would have properties for the back end host device.
That does sound nicer.
Originally I had SpaprMachineClass implement the
TYPE_TPM_IFinterface so we could define a TPM backend via -tpmdev passthrough,path=..., but after some discussion with the TPM maintainer it didn't quite work for the main use-case of passing through a TPM Resource Manager since it isn't suitable for full vTPM front-ends (since multiple guests can interfere with each other's operations when running the full gamut of TPM functionality).I hadn't consider a stand-alone -device implementation though. It's not a proper VIO or PCI device so there's no proper bus to attach it to. I guess we would just make it a direct child of SpaprMachineState (sort of like SpaprDrcClass), then we could define it via something like -object spapr-tpm-proxy,path=....
Quoted 8 lines
Quoted 1 line
It should be -device not -object, but otherwise that looks ok.
Ok, for some reason I thought -device needed either a specific bus or needed to be a SysBusDevice to attach to main-system-bus, but maybe that was just for qdev-managed reset handling. I've re-worked the series to allow configuration via:
-device spapr-tpm-proxy,host_path=/dev/tpmrmX
That looks good.
Quoted 7 lines
Quoted 3 lines
Quoted 1 line
How does the TPM appear in the device tree?
Nothing in the guest, on the host it appears as:
Hrm. That seems unwise. I mean, I guess its being treated as a hypervisor facility rather than a device per se, but what if we ever need to advertise more metadata about it.
It's a little bit awkward using a device tree in this case since it's generally the ultravisor that will be making this hcall on behalf of a guest requesting switch-over to SVM mode. The TPM device itself has a GetCapabilities command that seems like it would cover most of the metadata we might need though:
https://trustedcomputinggroup.org/wp-content/uploads/TPM-Rev-2.0-Part-3-Commands-01.38.pdf
(page 340)
and if we need to add a layer of metadata on top of that there's also the option of introducing support for an additional operation in H_TPM_COMM itself, e.g. TPM_COMM_OP_GET_CAPABILITIES. Unsupported or invalid operations have a unique H_PARAMETER return code so callers should be able to reliably probe for it in the future if they need more information.
Quoted 306 lines
Quoted 300 lines
./xscom@603fc00000000/i2cm@a2000/i2c-bus@0/tpm@57 ./xscom@603fc00000000/i2cm@a2000/i2c-bus@0/tpm@57/link-id ./xscom@603fc00000000/i2cm@a2000/i2c-bus@0/tpm@57/linux,sml-size ./xscom@603fc00000000/i2cm@a2000/i2c-bus@0/tpm@57/label ./xscom@603fc00000000/i2cm@a2000/i2c-bus@0/tpm@57/compatible ./xscom@603fc00000000/i2cm@a2000/i2c-bus@0/tpm@57/status ./xscom@603fc00000000/i2cm@a2000/i2c-bus@0/tpm@57/reg ./xscom@603fc00000000/i2cm@a2000/i2c-bus@0/tpm@57/phandle ./xscom@603fc00000000/i2cm@a2000/i2c-bus@0/tpm@57/linux,sml-base ./xscom@603fc00000000/i2cm@a2000/i2c-bus@0/tpm@57/name
Quoted 287 lines
Quoted 286 lines
I'll go ahead and give that a shot, assuming it seems reasonable to you.
Quoted 5 lines
Quoted 3 lines
By default, no tpm-device-file is defined and hcalls will return
H_RESOURCE.Wouldn't
H_FUNCTIONmake more sense?Yes, for this case it probably would.
Thanks for the suggestions!
Quoted 272 lines
Quoted 103 lines
The full specification for this hypercall can be found in
docs/specs/ppc-spapr-uv-hcalls.txtSigned-off-by Michael Roth <[email protected]--- hw/ppc/Makefile.objs | 1 + hw/ppc/spapr.c | 27 ++++++++ hw/ppc/spapr_hcall_tpm.c | 135 +++++++++++++++++++++++++++++++++++++++ hw/ppc/trace-events | 4 ++ include/hw/ppc/spapr.h | 7 +- 5 files changed, 173 insertions(+), 1 deletion(-) create mode 100644 hw/ppc/spapr_hcall_tpm.chw/ppc/Makefile.objs +1
-0@@ -5,6 +5,7 @@ obj-$(CONFIG_PSERIES) += spapr.o spapr_caps.o spapr_vio.o spapr_events.o obj-$(CONFIG_PSERIES) += spapr_hcall.o spapr_iommu.o spapr_rtas.o obj-$(CONFIG_PSERIES) += spapr_pci.o spapr_rtc.o spapr_drc.o obj-$(CONFIG_PSERIES) += spapr_cpu_core.o spapr_ovec.o spapr_irq.o +obj-$(CONFIG_PSERIES) += spapr_hcall_tpm.o obj-$(CONFIG_SPAPR_RNG) += spapr_rng.o # IBM PowerNV obj-$(CONFIG_POWERNV) += pnv.o pnv_xscom.o pnv_core.o pnv_lpc.o pnv_psi.o pnv_occ.o pnv_bmc.ohw/ppc/spapr.c +27
-0@@ -1776,6 +1776,10 @@ static void spapr_machine_reset(MachineState *machine)*/ object_child_foreach_recursive (object_get_root (), spapr_reset_drcs,NULL ); +if (spapr->tpm_device_file) { +spapr_hcall_tpm_reset (); + } +spapr_clear_pending_events (spapr);/* @@ -3340,6 +3344,21 @@ static void spapr_set_host_serial(Object *obj, const char *value, Error **errp) spapr->host_serial =g_strdup (value); } +static char *spapr_get_tpm_device_file (Object *obj,Error **errp) +{ +SpaprMachineState *spapr =SPAPR_MACHINE (obj); + +return g_strdup (spapr->tpm_device_file); +} + +static void spapr_set_tpm_device_file (Object *obj,const char *value,Error **errp) +{ +SpaprMachineState *spapr =SPAPR_MACHINE (obj); + +g_free (spapr->tpm_device_file); + spapr->tpm_device_file =g_strdup (value); +} +static void spapr_instance_init (Object *obj) {SpaprMachineState *spapr =SPAPR_MACHINE (obj);@@ -3396,6 +3415,14 @@ static void spapr_instance_init(Object *obj) &error_abort);object_property_set_description (obj,"host-serial" ,"Host serial number to advertise in guest device tree" , &error_abort); +object_property_add_str (obj,"tpm-device-file" , + spapr_get_tpm_device_file, + spapr_set_tpm_device_file, &error_abort); +object_property_set_description (obj,"tpm-device-file" , +"Specifies the path to the TPM character device file to use" +" for TPM communication via hypercalls (usually a TPM" +" resource manager)" , + &error_abort); }static void spapr_machine_finalizefn (Object *obj)hw/ppc/spapr_hcall_tpm.c new file +18
-0@@ -0,0 +1,135 @@ +/* +* SPAPR TPM Hypercall +* +* Copyright IBM Corp. 2019 +* +* Authors: +* Michael Roth <[email protected]> +* +* This work is licensed under the terms of the GNU GPL, version 2 or later. +* See the COPYING file in the top-level directory. +*/ + +#include "qemu/osdep.h" +#include "qemu-common.h" +#include "qapi/error.h" +#include "qemu/error-report.h" +#include "cpu.h" +#include "hw/ppc/spapr.h" hw/ppc/spapr_hcall_tpm.c line 28
Quoted 10 lines
hw/ppc/spapr_hcall_tpm.c new file +10
-0+#include "trace.h" + +#define TPM_SPAPR_BUFSIZE 4096 + +enum { +TPM_COMM_OP_EXECUTE =1 , +TPM_COMM_OP_CLOSE_SESSION =2 , +}; + +static int tpm_devfd =-1 ;A global? Really? You can do better.
Quoted 155 lines
hw/ppc/spapr_hcall_tpm.c new file +107
-0+ +static ssize_t tpm_execute (SpaprMachineState *spapr,target_ulong *args) +{ +uint64_t data_in =ppc64_phys_to_real (args[1 ]); +target_ulong data_in_size = args[2 ]; +uint64_t data_out =ppc64_phys_to_real (args[3 ]); +target_ulong data_out_size = args[4 ]; +uint8_t buf_in[TPM_SPAPR_BUFSIZE ]; +uint8_t buf_out[TPM_SPAPR_BUFSIZE ]; +ssize_t ret; + +trace_spapr_tpm_execute (data_in, data_in_size, data_out, data_out_size); + +if (data_in_size >TPM_SPAPR_BUFSIZE ) { +error_report ("invalid TPM input buffer size: " TARGET_FMT_lu"\n" , + data_in_size); +return H_P3 ; + } + +if (data_out_size <TPM_SPAPR_BUFSIZE ) { +error_report ("invalid TPM output buffer size: " TARGET_FMT_lu"\n" , + data_out_size); +return H_P5 ; + } + +if (tpm_devfd ==-1 ) { + tpm_devfd =open (spapr->tpm_device_file,O_RDWR ); +if (tpm_devfd ==-1 ) { +error_report ("failed to open TPM device %s: %d" , + spapr->tpm_device_file, errno); +return H_RESOURCE ; + } + } + +cpu_physical_memory_read (data_in, buf_in, data_in_size); + +do { + ret =write (tpm_devfd, buf_in, data_in_size); +if (ret >0 ) { + data_in_size -= ret; + } + }while ((ret >=0 && data_in_size >0 ) || (ret ==-1 && errno ==EINTR )); + +if (ret ==-1 ) { +error_report ("failed to write to TPM device %s: %d" , + spapr->tpm_device_file, errno); +return H_RESOURCE ; + } + +do { + ret =read (tpm_devfd, buf_out, data_out_size); + }while (ret ==0 || (ret ==-1 && errno ==EINTR )); + +if (ret ==-1 ) { +error_report ("failed to read from TPM device %s: %d" , + spapr->tpm_device_file, errno); +return H_RESOURCE ; + } + +cpu_physical_memory_write (data_out, buf_out, ret); + args[0 ] = ret; + +return H_SUCCESS ; +} + +static target_ulong h_tpm_comm (PowerPCCPU *cpu, +SpaprMachineState *spapr, +target_ulong opcode, +target_ulong *args) +{ +target_ulong op = args[0 ]; + +trace_spapr_h_tpm_comm (spapr->tpm_device_file ?:"null" , op); + +if (!spapr->tpm_device_file) { +error_report ("TPM device not specified" ); +return H_RESOURCE ; + } + +switch (op) { +case TPM_COMM_OP_EXECUTE : +return tpm_execute (spapr, args); +case TPM_COMM_OP_CLOSE_SESSION : +if (tpm_devfd !=-1 ) { +close (tpm_devfd); + tpm_devfd =-1 ; + } +return H_SUCCESS ; +default : +return H_PARAMETER ; + } +} + +void spapr_hcall_tpm_reset (void ) +{ +if (tpm_devfd !=-1 ) { +close (tpm_devfd); + tpm_devfd =-1 ; + } +} + +static void hypercall_register_types (void ) +{ +spapr_register_hypercall (H_TPM_COMM , h_tpm_comm); +} + +type_init (hypercall_register_types)hw/ppc/trace-events +4
-0@@ -25,6 +25,10 @@ spapr_update_dt(unsigned cb) "New blob %u bytes" spapr_update_dt_failed_size(unsigned cbold, unsigned cbnew, unsigned magic) "Old blob %u bytes, new blob %u bytes, magic 0x%x" spapr_update_dt_failed_check(unsigned cbold, unsigned cbnew, unsigned magic) "Old blob %u bytes, new blob %u bytes, magic 0x%x" +# spapr_hcall_tpm.c +spapr_h_tpm_comm(const char *device_path, uint64_t operation) "tpm_device_path=%s operation=0x%"PRIu64 +spapr_tpm_execute(uint64_t data_in, uint64_t data_in_sz, uint64_t data_out, uint64_t data_out_sz) "data_in=0x%"PRIx64", data_in_sz=%"PRIu64", data_out=0x%"PRIx64", data_out_sz=%"PRIu64 + # spapr_iommu.c spapr_iommu_put(uint64_t liobn, uint64_t ioba, uint64_t tce, uint64_t ret) "liobn=0x%"PRIx64" ioba=0x%"PRIx64" tce=0x%"PRIx64" ret=%"PRId64 spapr_iommu_get(uint64_t liobn, uint64_t ioba, uint64_t ret, uint64_t tce) "liobn=0x%"PRIx64" ioba=0x%"PRIx64" ret=%"PRId64" tce=0x%"PRIx64include/hw/ppc/spapr.h +6
-1@@ -198,6 +198,7 @@ struct SpaprMachineState {SpaprXive *xive;SpaprIrq *irq;qemu_irq *qirqs; +char *tpm_device_file;bool cmd_line_caps[SPAPR_CAP_NUM ];SpaprCapabilities def, eff, mig;@@ -490,8 +491,9 @@ struct SpaprMachineState {#define H_INT_ESB 0x3C8#define H_INT_SYNC 0x3CC#define H_INT_RESET 0x3D0 +#define H_TPM_COMM 0xEF10 -#define MAX_HCALL_OPCODE H_INT_RESET +#define MAX_HCALL_OPCODE H_TPM_COMM/* The hcalls above are standardized in PAPR and implemented by pHyp * as well. @@ -864,6 +866,9 @@ int spapr_caps_post_migration(SpaprMachineState *spapr);void spapr_check_pagesize (SpaprMachineState *spapr,hwaddr pagesize,Error **errp); + +void spapr_hcall_tpm_reset (void ); +/* * XIVE definitions */ Signature · 4 lines
David Gibson | I'll have my music baroque, and my code david AT gibson.dropbear.id.au | minimalist, thank you. NOT _the_ _other_ | _way_ _around_! http://www.ozlabs.org/~dgibson
Re: [Qemu-devel] [PATCH 2/2] spapr: initial implementation for H_TPM_COMM hcall
Quoted 63 lines · Michael Roth
Quoting David Gibson (2019-07-16 20:29:12)
Quoted 45 lines · Michael Roth
Quoted 33 lines
Quoting David Gibson (2019-07-14 21:25:24)
Quoted 32 lines · Michael Roth
Quoted 30 lines
Quoting David Gibson (2019-07-12 01:46:19)
Quoted 8 lines · Michael Roth
Quoted 7 lines
This implements the
H_TPM_COMMhypercall, which is used by an Ultravisor to pass TPM commands directly to the host's TPM device, or a TPM Resource Manager associated with the device.This also introduces a new pseries machine option which is used to configure what TPM device to pass commands to, for example:
Quoted 6 lines
Quoted 1 line
-machine pseries,...,tpm-device-file=/dev/tmprm0
Bolting this into yet another machine parameter seems kind of ugly. Wouldn't it make more sense to treat this as an virtual device (say "spapr-vtpm"). Adding that device would enable the hcall, and would have properties for the back end host device.
That does sound nicer.
Originally I had SpaprMachineClass implement the
TYPE_TPM_IFinterface so we could define a TPM backend via -tpmdev passthrough,path=..., but after some discussion with the TPM maintainer it didn't quite work for the main use-case of passing through a TPM Resource Manager since it isn't suitable for full vTPM front-ends (since multiple guests can interfere with each other's operations when running the full gamut of TPM functionality).I hadn't consider a stand-alone -device implementation though. It's not a proper VIO or PCI device so there's no proper bus to attach it to. I guess we would just make it a direct child of SpaprMachineState (sort of like SpaprDrcClass), then we could define it via something like -object spapr-tpm-proxy,path=....
Quoted 8 lines
Quoted 1 line
It should be -device not -object, but otherwise that looks ok.
Ok, for some reason I thought -device needed either a specific bus or needed to be a SysBusDevice to attach to main-system-bus, but maybe that was just for qdev-managed reset handling. I've re-worked the series to allow configuration via:
-device spapr-tpm-proxy,host_path=/dev/tpmrmX
That looks good.
Quoted 7 lines
Quoted 3 lines
Quoted 1 line
How does the TPM appear in the device tree?
Nothing in the guest, on the host it appears as:
Hrm. That seems unwise. I mean, I guess its being treated as a hypervisor facility rather than a device per se, but what if we ever need to advertise more metadata about it.
It's a little bit awkward using a device tree in this case since it's generally the ultravisor that will be making this hcall on behalf of a guest requesting switch-over to SVM mode. The TPM device itself has a GetCapabilities command that seems like it would cover most of the metadata we might need though:
https://trustedcomputinggroup.org/wp-content/uploads/TPM-Rev-2.0-Part-3-Commands-01.38.pdf
(page 340)
Quoted 6 lines
and if we need to add a layer of metadata on top of that there's also the option of introducing support for an additional operation in
H_TPM_COMMitself, e.g.TPM_COMM_OP_GET_CAPABILITIES. Unsupported or invalid operations have a uniqueH_PARAMETERreturn code so callers should be able to reliably probe for it in the future if they need more information.
Yeah, ok.
Signature · 4 lines
David Gibson | I'll have my music baroque, and my code david AT gibson.dropbear.id.au | minimalist, thank you. NOT _the_ _other_ | _way_ _around_! http://www.ozlabs.org/~dgibson
Quoted 9 lines · Michael Roth
These patches are also available at:
https://github.com/mdroth/qemu/commits/spapr-tpm-hcall-v0
This patchset implements the
H_TPM_COMMhypercall, which provides a way for an Ultravisor to pass raw TPM commands on to a host's TPM device, either directly or through a TPM Resource Manager (needed to support multiple guests).
Quoted 6 lines
Secure Guests running on an Ultravisor have a symmetric key that is encrypted using a public key that is bound to a trusted host's TPM hardware. This hypercall provides a means to decrypt the symmetric key on behalf of a Secure Guest using the host's TPM hardware.
More details are provided in the spec summary introduced in patch 1.
This is obviously 4.2 material, other comments on the individual patches.
Quoted 10 lines
docs/specs/ppc-spapr-uv-hcalls.txt | 74 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ hw/ppc/Makefile.objs | 1 + hw/ppc/spapr.c | 27 +++++++++++++++++++++++++++ hw/ppc/spapr_hcall_tpm.c | 135 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ hw/ppc/trace-events | 4 ++++ include/hw/ppc/spapr.h | 7 ++++++- 6 files changed, 247 insertions(+), 1 deletion(-)
Signature · 4 lines
David Gibson | I'll have my music baroque, and my code david AT gibson.dropbear.id.au | minimalist, thank you. NOT _the_ _other_ | _way_ _around_! http://www.ozlabs.org/~dgibson
Patchew URL: https://patchew.org/QEMU/[email protected]/
Patchew URL: https://patchew.org/QEMU/[email protected]/
Hi,
This series seems to have some coding style problems. See output below for more information:
Message-id: [email protected]
Type: series
Subject: [Qemu-devel] [RFC PATCH 0/2] spapr: Implement H_TPM_COMM for accessing host TPM device
=== TEST SCRIPT BEGIN ===
#!/bin/bash
git rev-parse base > /dev/null || exit 0
git config --local diff.renamelimit 0
git config --local diff.renames True
git config --local diff.algorithm histogram./scripts/checkpatch.pl --mailback base..
=== TEST SCRIPT END ===
Switched to a new branch 'test'12b8055a19 spapr: initial implementation for H_TPM_COMM hcall 47c8841564 docs/specs: initial spec summary for Ultravisor-related hcalls
=== OUTPUT BEGIN ===
1/2 Checking commit 47c884156452 (docs/specs: initial spec summary for Ultravisor-related hcalls)
WARNING: added, moved or deleted file(s), does MAINTAINERS need updating?
#18:
new file mode 100644
total: 0 errors, 1 warnings, 74 lines checked
Patch 1/2 has style problems, please review. If any of these errors are false positives report them to the maintainer, see
CHECKPATCH in MAINTAINERS.
2/2 Checking commit 12b8055a1905 (spapr: initial implementation for H_TPM_COMM hcall)
WARNING: line over 80 characters
#63: FILE: hw/ppc/spapr.c:3354:
+static void spapr_set_tpm_device_file(Object *obj, const char *value, Error **errp)
WARNING: added, moved or deleted file(s), does MAINTAINERS need updating? #90:
new file mode 100644
ERROR: Error messages should not contain newlines
#137: FILE: hw/ppc/spapr_hcall_tpm.c:43:
+ error_report("invalid TPM input buffer size: " TARGET_FMT_lu "\n",
ERROR: Error messages should not contain newlines
#143: FILE: hw/ppc/spapr_hcall_tpm.c:49:
+ error_report("invalid TPM output buffer size: " TARGET_FMT_lu "\n",
ERROR: switch and case should be at the same indent #202: FILE: hw/ppc/spapr_hcall_tpm.c:108:
+ switch (op) {
+ case TPM_COMM_OP_EXECUTE:
[...]
+ case TPM_COMM_OP_CLOSE_SESSION:
[...]
+ default:
total: 3 errors, 2 warnings, 223 lines checked
Patch 2/2 has style problems, please review. If any of these errors are false positives report them to the maintainer, see CHECKPATCH in MAINTAINERS.
=== OUTPUT END ===
Test command exited with code: 1
The full log is available at
http://patchew.org/logs/[email protected]/testing.checkpatch/?type=message.
---
Email generated automatically by Patchew [https://patchew.org/].
Please send your feedback to [email protected]