Re: [PATCH v6 03/18] s390x: protvirt: Support unpack facility

[Janosch Frank]

On 3/5/20 2:51 PM, David Hildenbrand wrote:
> On 04.03.20 12:42, Janosch Frank wrote:
>> When a guest has saved a ipib of type 5 and calls diagnose308 with
> 
> s/a/an/
> 
>> subcode 10, we have to setup the protected processing environment via
>> Ultravisor calls. The calls are done by KVM and are exposed via an
>> API.
>>
>> The following steps are necessary:
>> 1. Enable protected mode for the VM (register it and its cpus with the 
>> Ultravisor)
>> 2. Forward the secure header to the Ultravisor (has all information on
>> how to decrypt the image and VM information)
>> 3. Protect image pages from the host and decrypt them
>> 4. Verify the image integrity
>>
>> Only after step 4 a protected VM is allowed to run.
>>
>> Signed-off-by: Janosch Frank <address@hidden>
>> Signed-off-by: Christian Borntraeger <address@hidden> [Changes
>> to machine]
>> ---
>>  hw/s390x/Makefile.objs              |   1 +
>>  hw/s390x/ipl.c                      |  33 +++++++++
>>  hw/s390x/ipl.h                      |   2 +
>>  hw/s390x/pv.c                       | 106 ++++++++++++++++++++++++++++
>>  hw/s390x/pv.h                       |  34 +++++++++
>>  hw/s390x/s390-virtio-ccw.c          |  91 ++++++++++++++++++++++++
>>  include/hw/s390x/s390-virtio-ccw.h  |   1 +
>>  target/s390x/cpu.c                  |   4 ++
>>  target/s390x/cpu.h                  |   1 +
>>  target/s390x/cpu_features_def.inc.h |   1 +
>>  10 files changed, 274 insertions(+)
>>  create mode 100644 hw/s390x/pv.c
>>  create mode 100644 hw/s390x/pv.h
> 
> [...]
> 
>>  
>>  #define KERN_IMAGE_START                0x010000UL
>>  #define LINUX_MAGIC_ADDR                0x010008UL
>> @@ -676,6 +677,38 @@ static void s390_ipl_prepare_qipl(S390CPU *cpu)
>>      cpu_physical_memory_unmap(addr, len, 1, len);
>>  }
>>  
>> +int s390_ipl_prepare_pv_header(void)
>> +{
>> +    S390IPLState *ipl = get_ipl_device();
>> +    IPLBlockPV *ipib_pv = &ipl->iplb_pv.pv;
>> +    void *hdr = g_malloc(ipib_pv->pv_header_len);
>> +    int rc;
>> +
>> +    cpu_physical_memory_read(ipib_pv->pv_header_addr, hdr,
>> +                             ipib_pv->pv_header_len);
>> +    rc = s390_pv_set_sec_parms((uint64_t)hdr,
>> +                          ipib_pv->pv_header_len);
>> +    g_free(hdr);
>> +    return rc;
>> +}
>> +
>> +int s390_ipl_pv_unpack(void)
>> +{
>> +    int i, rc = 0;
>> +    S390IPLState *ipl = get_ipl_device();
>> +    IPLBlockPV *ipib_pv = &ipl->iplb_pv.pv;
>> +
>> +    for (i = 0; i < ipib_pv->num_comp; i++) {
>> +        rc = s390_pv_unpack(ipib_pv->components[i].addr,
>> +                            TARGET_PAGE_ALIGN(ipib_pv->components[i].size),
>> +                            ipib_pv->components[i].tweak_pref);
>> +        if (rc) {
>> +            break;
>> +        }
> 
> You can check for " && !rc" in the loop condition

Not sure if that would make it more readable...

> 
>> +    }
>> +    return rc;
>> +}
>> +
>>  void s390_ipl_prepare_cpu(S390CPU *cpu)
>>  {
>>      S390IPLState *ipl = get_ipl_device();
>> diff --git a/hw/s390x/ipl.h b/hw/s390x/ipl.h
>> index 04be63cee1..ad8090a02c 100644
>> --- a/hw/s390x/ipl.h
>> +++ b/hw/s390x/ipl.h
>> @@ -105,6 +105,8 @@ typedef union IplParameterBlock IplParameterBlock;
>>  int s390_ipl_set_loadparm(uint8_t *loadparm);
>>  int s390_ipl_pv_check_components(IplParameterBlock *iplb);
>>  void s390_ipl_update_diag308(IplParameterBlock *iplb);
>> +int s390_ipl_prepare_pv_header(void);
>> +int s390_ipl_pv_unpack(void);
>>  void s390_ipl_prepare_cpu(S390CPU *cpu);
>>  IplParameterBlock *s390_ipl_get_iplb(void);
>>  IplParameterBlock *s390_ipl_get_iplb_secure(void);
>> diff --git a/hw/s390x/pv.c b/hw/s390x/pv.c
>> new file mode 100644
>> index 0000000000..50b68b6c34
>> --- /dev/null
>> +++ b/hw/s390x/pv.c
>> @@ -0,0 +1,106 @@
>> +/*
>> + * Secure execution functions
>> + *
>> + * Copyright IBM Corp. 2020
>> + * Author(s):
>> + *  Janosch Frank <address@hidden>
>> + *
>> + * This work is licensed under the terms of the GNU GPL, version 2 or (at
>> + * your option) any later version. See the COPYING file in the top-level
>> + * directory.
>> + */
>> +#include "qemu/osdep.h"
>> +#include <sys/ioctl.h>
>> +
>> +#include <linux/kvm.h>
>> +
>> +#include "qemu/error-report.h"
>> +#include "sysemu/kvm.h"
>> +#include "pv.h"
>> +
>> +const char *cmd_names[] = {
>> +    "VM_ENABLE",
>> +    "VM_DISABLE",
>> +    "VM_SET_SEC_PARAMS",
>> +    "VM_UNPACK",
>> +    "VM_VERIFY",
>> +    "VM_PREP_RESET",
>> +    "VM_UNSHARE_ALL",
>> +    NULL
> 
> Is the NULL really needed? This will be somewhat error prone when we add
> new PV commands. Not sure if guarding this by an access function (chack
> against ARRAY_SIZE() makes sense).
> 
>> diff --git a/hw/s390x/s390-virtio-ccw.c b/hw/s390x/s390-virtio-ccw.c
>> index a89cf4c129..dd39890f89 100644
>> --- a/hw/s390x/s390-virtio-ccw.c
>> +++ b/hw/s390x/s390-virtio-ccw.c
>> @@ -41,6 +41,8 @@
>>  #include "hw/qdev-properties.h"
>>  #include "hw/s390x/tod.h"
>>  #include "sysemu/sysemu.h"
>> +#include "hw/s390x/pv.h"
>> +#include <linux/kvm.h>
>>  
>>  S390CPU *s390_cpu_addr2state(uint16_t cpu_addr)
>>  {
>> @@ -238,9 +240,11 @@ static void s390_create_sclpconsole(const char *type, 
>> Chardev *chardev)
>>  static void ccw_init(MachineState *machine)
>>  {
>>      int ret;
>> +    S390CcwMachineState *ms = S390_CCW_MACHINE(machine);
>>      VirtualCssBus *css_bus;
>>      DeviceState *dev;
>>  
>> +    ms->pv = false;
> 
> Should not be necessary, default is false.

ok

> 
>>      s390_sclp_init();
>>      /* init memory + setup max page size. Required for the CPU model */
>>      s390_memory_init(machine->ram);
>> @@ -316,10 +320,75 @@ static inline void s390_do_cpu_ipl(CPUState *cs, 
>> run_on_cpu_data arg)
>>      s390_cpu_set_state(S390_CPU_STATE_OPERATING, cpu);
>>  }
>>  
>> +static void s390_machine_unprotect(S390CcwMachineState *ms)
>> +{
>> +    CPUState *t;
>> +
>> +    if (!ms->pv)
>> +        return;
> 
> How can this ever happen? g_assert(ms->pv) ?

Currently not, that's only used in the follow up patches with the ballon
and migration blocker

> 
> Also, i don't see this function getting called from anywhere else except
> when s390_machine_protect() fails. That looks wrong. This has to be
> called when going out of PV mode.

Yes, but that's in the diag308 1-4 patch.

> 
> 
>> +    s390_pv_vm_disable();
>> +    CPU_FOREACH(t) {
>> +        S390_CPU(t)->env.pv = false;
>> +    }
>> +    ms->pv = false;
>> +}
>> +
>> +static int s390_machine_protect(S390CcwMachineState *ms)
>> +{
>> +    CPUState *t;
>> +    int rc;
>> +
> 
> g_assert(!ms->pv) ?

Ok

> 
>> +    /* Create SE VM */
>> +    rc = s390_pv_vm_enable();
>> +    if (rc) {
>> +        return rc;
>> +    }
>> +
>> +    CPU_FOREACH(t) {
>> +        S390_CPU(t)->env.pv = true;
>> +    }
>> +    ms->pv = true;
>> +
>> +    /* Set SE header and unpack */
>> +    rc = s390_ipl_prepare_pv_header();
>> +    if (rc) {
>> +        goto out_err;
>> +    }
>> +
>> +    /* Decrypt image */
>> +    rc = s390_ipl_pv_unpack();
>> +    if (rc) {
>> +        goto out_err;
>> +    }
>> +
>> +    /* Verify integrity */
>> +    rc = s390_pv_verify();
>> +    if (rc) {
>> +        goto out_err;
>> +    }
>> +    return rc;
>> +
>> +out_err:
>> +    s390_machine_unprotect(ms);
>> +    return rc;
>> +}
>> +
>> +#define DIAG_308_RC_INVAL_FOR_PV    0x0a02
>> +static void s390_machine_inject_pv_error(CPUState *cs)
>> +{
>> +    int r1 = (cs->kvm_run->s390_sieic.ipa & 0x00f0) >> 4;
>> +    CPUS390XState *env = &S390_CPU(cs)->env;
>> +
>> +    /* Report that we are unable to enter protected mode */
>> +    env->regs[r1 + 1] = DIAG_308_RC_INVAL_FOR_PV;
>> +}
>> +
> 
> [...]
>>      switch (reset_type) {
>>      case S390_RESET_EXTERNAL:
>>      case S390_RESET_REIPL:
>> @@ -353,6 +424,26 @@ static void s390_machine_reset(MachineState *machine)
>>          }
>>          subsystem_reset();
>>          run_on_cpu(cs, s390_do_cpu_initial_reset, RUN_ON_CPU_NULL);
>> +        run_on_cpu(cs, s390_do_cpu_load_normal, RUN_ON_CPU_NULL);
> 
> This does look unrelated and wrong?

Indeed, that looks dodgy

> 
>> +        break;
>> +    case S390_RESET_PV: /* Subcode 10 */
>> +        subsystem_reset();
>> +        s390_crypto_reset();
>> +
>> +        CPU_FOREACH(t) {
>> +            if (t == cs) {
>> +                continue;
>> +            }
>> +            run_on_cpu(t, s390_do_cpu_full_reset, RUN_ON_CPU_NULL);
>> +        }
>> +        run_on_cpu(cs, s390_do_cpu_reset, RUN_ON_CPU_NULL);
>> +
>> +        if (s390_machine_protect(ms)) {
>> +            s390_machine_inject_pv_error(cs);
> 
> Ah, so it's not an actual exception. BUT: All other guest cpus were
> reset, can the guest deal with that?

Well, all other CPUs should be stopped for diag308, no?
Also it's done by the bootloader and not a OS which just stops its cpus
and goes into protected mode.

> 
> (run_on_cpu(cs, s390_do_cpu_reset, RUN_ON_CPU_NULL) should go after the
> s390_machine_protect() I assume - the change you had in the other patch)

That's not a good idea, I want to reset before we automatically call the
UV routines on a reset.

> 
>> +            s390_cpu_set_state(S390_CPU_STATE_OPERATING, cpu);
>> +            return;
>> +        }
>> +
>>          run_on_cpu(cs, s390_do_cpu_load_normal, RUN_ON_CPU_NULL);
>>          break;
>>      default:
> 
>

signature.asc
Description: OpenPGP digital signature