On 11/28/19 3:07 PM, Thomas Huth wrote:
> On 20/11/2019 12.43, Janosch Frank wrote:
>> When a guest has saved a ipib of type 5 and call diagnose308 with
>> 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. Create a VM (register it with the Ultravisor)
>> 2. Create secure CPUs for all of our current cpus
>> 3. Forward the secure header to the Ultravisor (has all information on
>> how to decrypt the image and VM information)
>> 4. Protect image pages from the host and decrypt them
>> 5. Verify the image integrity
>>
>> Only after step 5 a protected VM is allowed to run.
>>
>> Signed-off-by: Janosch Frank <frankja@linux.ibm.com>
>> ---
> [...]
>> diff --git a/hw/s390x/ipl.c b/hw/s390x/ipl.c
>> index a077926f36..50501fcd27 100644
>> --- a/hw/s390x/ipl.c
>> +++ b/hw/s390x/ipl.c
>> @@ -33,6 +33,7 @@
>>  #include "qemu/cutils.h"
>>  #include "qemu/option.h"
>>  #include "exec/exec-all.h"
>> +#include "pv.h"
>>  
>>  #define KERN_IMAGE_START                0x010000UL
>>  #define LINUX_MAGIC_ADDR                0x010008UL
>> @@ -668,6 +669,38 @@ static void s390_ipl_prepare_qipl(S390CPU *cpu)
>>      cpu_physical_memory_unmap(addr, len, 1, len);
>>  }
>>  
>> +int s390_ipl_prepare_pv_header(void)
>> +{
>> +    int rc;
>> +    IplParameterBlock *iplb = s390_ipl_get_iplb_secure();
>> +    IPLBlockPV *ipib_pv = &iplb->pv;
>> +    void *hdr = g_malloc(ipib_pv->pv_header_len);
>> +
>> +    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;
>> +    IplParameterBlock *iplb = s390_ipl_get_iplb_secure();
>> +    IPLBlockPV *ipib_pv = &iplb->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) {
>> +            return rc;
>> +        }
>> +    }
>> +    return rc;
>> +}
> 
> For both functions, s390_ipl_prepare_pv_header() and
> s390_ipl_pv_unpack(), you're ignoring the return code at the calling
> site, so errors go completely unnoticed. I suggest to either do an
> error_report() here or at the calling site...
> 
>> diff --git a/hw/s390x/pv.c b/hw/s390x/pv.c
>> new file mode 100644
>> index 0000000000..0218070322
>> --- /dev/null
>> +++ b/hw/s390x/pv.c
>> @@ -0,0 +1,118 @@
>> +/*
>> + * Secure execution functions
>> + *
>> + * Copyright IBM Corp. 2019
>> + * Author(s):
>> + *  Janosch Frank <frankja@linux.ibm.com>
>> + *
>> + * 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"
>> +
>> +static int s390_pv_cmd(uint32_t cmd, void *data)
>> +{
>> +    int rc;
>> +    struct kvm_pv_cmd pv_cmd = {
>> +        .cmd = cmd,
>> +        .data = (uint64_t)data,
>> +    };
>> +
>> +    rc = kvm_vm_ioctl(kvm_state, KVM_S390_PV_COMMAND, &pv_cmd);
>> +    if (rc) {
>> +        error_report("KVM PV command failed cmd: %d rc: %d", cmd, rc);
>> +        exit(1);
>> +    }
>> +    return rc;
>> +}
>> +
>> +static int s390_pv_cmd_vcpu(CPUState *cs, uint32_t cmd, void *data)
>> +{
>> +    int rc;
>> +    struct kvm_pv_cmd pv_cmd = {
>> +        .cmd = cmd,
>> +        .data = (uint64_t)data,
>> +    };
>> +
>> +    rc = kvm_vcpu_ioctl(cs, KVM_S390_PV_COMMAND_VCPU, &pv_cmd);
>> +    if (rc) {
>> +        error_report("KVM PV VCPU command failed cmd: %d rc: %d", cmd, rc);
>> +        exit(1);
>> +    }
>> +    return rc;
>> +}
> 
> ... or even better, since these functions exit on error anyway, make
> these functions and all the others "void" instead?
> 
>  Thomas
> 

You're not the first stating this idea :-)
We might get a return code for diag308 that signals that we failed to
start the secure guest and then I might need them again.

If we don't get a rc, I'll definitely make all functions void, if we do
I'll make nearly all void:)