On 24.03.2023 00:15, Andrew Cooper wrote:
> On 04/02/2022 12:54 pm, Jan Beulich wrote:
>> On 03.02.2022 14:38, Andrew Cooper wrote:
>>> --- a/xen/arch/x86/domain.c
>>> +++ b/xen/arch/x86/domain.c
>>> @@ -1693,11 +1693,8 @@ static void load_segments(struct vcpu *n)
>>> put_guest(uregs->fs, esp - 5) |
>>> put_guest(uregs->es, esp - 6) |
>>> put_guest(uregs->ds, esp - 7) )
>>> - {
>>> - gprintk(XENLOG_ERR,
>>> - "error while creating compat failsafe
>>> callback frame\n");
>>> - domain_crash(n->domain);
>>> - }
>>> + domain_crash(n->domain,
>>> + "Error creating compat failsafe
>>> callback frame\n");
>>>
>>> if ( n->arch.pv.vgc_flags & VGCF_failsafe_disables_events )
>>> vcpu_info(n, evtchn_upcall_mask) = 1;
>>> @@ -1732,11 +1729,8 @@ static void load_segments(struct vcpu *n)
>>> put_guest(uregs->ds, rsp - 9) |
>>> put_guest(regs->r11, rsp - 10) |
>>> put_guest(regs->rcx, rsp - 11) )
>>> - {
>>> - gprintk(XENLOG_ERR,
>>> - "error while creating failsafe callback frame\n");
>>> - domain_crash(n->domain);
>>> - }
>>> + domain_crash(n->domain,
>>> + "Error creating failsafe callback frame\n");
>>
>> I assume it wasn't really intended to hide potentially relevant
>> information
>> (the subject vCPU) by this change, which - by way of gprintk() - did get
>> logged before (since we already have n == current at this point)?
>
> The information is not lost. __domain_crash() prints current too,
> albeit in a long-winded way.
Oh, right - n == current guarantees the middle path to be taken there.
Considering the other sub-thread also ended up okay-ish:
Reviewed-by: Jan Beulich <jbeul...@suse.com>
Jan
