Re: [PATCH v1 1/3] x86/tdx: Add TDX Guest event notify interrupt support

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 




On 3/27/23 9:02 PM, Huang, Kai wrote:
> On Mon, 2023-03-27 at 19:50 -0700, Sathyanarayanan Kuppuswamy wrote:
>> Hi Kai,
>>
>> On 3/27/23 7:38 PM, Huang, Kai wrote:
>>>> +/* Reserve an IRQ from x86_vector_domain for TD event notification */
>>>> +static int __init tdx_event_irq_init(void)
>>>> +{
>>>> +	struct irq_alloc_info info;
>>>> +	cpumask_t saved_cpumask;
>>>> +	struct irq_cfg *cfg;
>>>> +	int cpu, irq;
>>>> +
>>>> +	if (!cpu_feature_enabled(X86_FEATURE_TDX_GUEST))
>>>> +		return 0;
>>>> +
>>>> +	init_irq_alloc_info(&info, NULL);
>>>> +
>>>> +	/*
>>>> +	 * Event notification vector will be delivered to the CPU
>>>> +	 * in which TDVMCALL_SETUP_NOTIFY_INTR hypercall is requested.
>>>> +	 * So set the IRQ affinity to the current CPU.
>>>> +	 */
>>>> +	cpu = get_cpu();
>>>> +	cpumask_copy(&saved_cpumask, current->cpus_ptr);
>>>> +	info.mask = cpumask_of(cpu);
>>>> +	put_cpu();
>>> The 'saved_cpumask' related code is ugly.  If you move put_cpu() to the end of
>>> this function, I think you can remove all related code:
>>>
>>> 	cpu = get_cpu();
>>>
>>> 	/*
>>> 	 * Set @info->mask to local cpu to make sure a valid vector is
>>> 	 * pre-allocated when TDX event notification IRQ is allocated
>>> 	 * from x86_vector_domain.
>>> 	 */
>>> 	init_irq_alloc_info(&info, cpumask_of(cpu));
>>>
>>> 	// rest staff: request_irq(), hypercall ...
>>>
>>> 	put_cpu();
>>> 	
>>
>> init_irq_alloc_info() is a sleeping function. Since get_cpu() disables
>> preemption, we cannot call sleeping function after it. Initially, I
>> have implemented it like you have mentioned. However, I discovered the
>> following error.
> 
> Oh sorry I forgot this.  So I think we should use migrate_disable() instead:
> 
> 	migrate_disable();
> 
> 	init_irq_alloc_info(&info, cpumask_of(smp_processor_id()));
> 
> 	...
> 
> 	migrate_enable();
> 
> Or, should we just use early_initcall() so that only BSP is running?  IMHO it's
> OK to always allocate the vector from BSP.
> 
> Anyway, either way is fine to me.

Final version looks like below. 

static int __init tdx_event_irq_init(void)
{
        struct irq_alloc_info info;
        struct irq_cfg *cfg;
        int irq;

        if (!cpu_feature_enabled(X86_FEATURE_TDX_GUEST))
                return 0;

        init_irq_alloc_info(&info, NULL);

        /*
         * Event notification vector will be delivered to the CPU
         * in which TDVMCALL_SETUP_NOTIFY_INTR hypercall is requested.
         * So set the IRQ affinity to the current CPU.
         */
        info.mask = cpumask_of(0);

        irq = irq_domain_alloc_irqs(x86_vector_domain, 1, cpu_to_node(0), &info);
        if (irq <= 0) {
                pr_err("Event notification IRQ allocation failed %d\n", irq);
                return -EIO;
        }

        irq_set_handler(irq, handle_edge_irq);

        /* Since the IRQ affinity is set, it cannot be balanced */
        if (request_irq(irq, tdx_event_irq_handler, IRQF_NOBALANCING,
                        "tdx_event_irq", NULL)) {
                pr_err("Event notification IRQ request failed\n");
                goto err_free_domain_irqs;
        }

        cfg = irq_cfg(irq);

        /*
         * Since tdx_event_irq_init() is triggered via early_initcall(),
         * it will called before secondary CPUs bringup. Since there is
         * only one CPU, it complies with the requirement of executing
         * the TDVMCALL_SETUP_NOTIFY_INTR hypercall on the same CPU where
         * the IRQ vector is allocated.
         *
         * Register callback vector address with VMM. More details
         * about the ABI can be found in TDX Guest-Host-Communication
         * Interface (GHCI), sec titled
         * "TDG.VP.VMCALL<SetupEventNotifyInterrupt>".
         */
        if (_tdx_hypercall(TDVMCALL_SETUP_NOTIFY_INTR, cfg->vector, 0, 0, 0)) {
                pr_err("Event notification hypercall failed\n");
                goto err_free_irqs;
        }

        tdx_event_irq = irq;

        return 0;

err_free_irqs:
        free_irq(irq, NULL);
err_free_domain_irqs:
        irq_domain_free_irqs(irq, 1);

        return -EIO;
}
early_initcall(tdx_event_irq_init)


> 
>>
>> [    2.400755] BUG: sleeping function called from invalid context at kernel/locking/mutex.c:580
>> [    2.404664] in_atomic(): 1, irqs_disabled(): 0, non_block: 0, pid: 1, name: swapper/0
>> [    2.408671] preempt_count: 1, expected: 0
>> [    2.412650] RCU nest depth: 0, expected: 0
>> [    2.412666] no locks held by swapper/0/1.
>> [    2.416650] Preemption disabled at:
>> [    2.416650] [<ffffffff83b8089f>] tdx_arch_init+0x38/0x117
>> [    2.420670] CPU: 0 PID: 1 Comm: swapper/0 Not tainted 6.1.0-rc4-00117-g672ca073d9f9-dirty #2527
>> [    2.424650] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 0.0.0 02/06/2015
>> [    2.424650] Call Trace:
>> [    2.424650]  <TASK>
>> [    2.424650]  dump_stack_lvl+0x6a/0x86
>> [    2.424650]  __might_resched.cold+0xf4/0x12f
>> [    2.424650]  __mutex_lock+0x50/0x810
>> [    2.424650]  ? lock_is_held_type+0xd8/0x130
>> [    2.424650]  ? __irq_alloc_descs+0xcf/0x310
>> [    2.424650]  ? find_held_lock+0x2b/0x80
>> [    2.424650]  ? __irq_alloc_descs+0xcf/0x310
>> [    2.424650]  __irq_alloc_descs+0xcf/0x310
>> [    2.424650]  irq_domain_alloc_descs.part.0+0x49/0xa0
>> [    2.424650]  __irq_domain_alloc_irqs+0x2a0/0x4f0
>> [    2.424650]  ? next_arg+0x129/0x1f0
>> [    2.424650]  ? tdx_guest_init+0x5b/0x5b
>> [    2.424650]  tdx_arch_init+0x8e/0x117
>> [    2.424650]  do_one_initcall+0x137/0x2ec
>> [    2.424650]  ? rcu_read_lock_sched_held+0x36/0x60
>> [    2.424650]  kernel_init_freeable+0x1e3/0x241
>> [    2.424650]  ? rest_init+0x1a0/0x1a0
>> [    2.424650]  kernel_init+0x17/0x170
>> [    2.424650]  ? rest_init+0x1a0/0x1a0
>> [    2.424650]  ret_from_fork+0x1f/0x30
>> [    2.424650]  </TASK>
> 

-- 
Sathyanarayanan Kuppuswamy
Linux Kernel Developer



[Index of Archives]     [Linux Wireless]     [Linux Kernel]     [ATH6KL]     [Linux Bluetooth]     [Linux Netdev]     [Kernel Newbies]     [Share Photos]     [IDE]     [Security]     [Git]     [Netfilter]     [Bugtraq]     [Yosemite News]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Linux RAID]     [Linux ATA RAID]     [Samba]     [Device Mapper]

  Powered by Linux