qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: "Philippe Mathieu-Daudé" <philmd@linaro.org>
To: Jiaxun Yang <jiaxun.yang@flygoat.com>, qemu-devel@nongnu.org
Cc: Huacai Chen <chenhuacai@kernel.org>, Song Gao <gaosong@loongson.cn>
Subject: Re: [PATCH v3 3/4] hw/mips/loongson3_virt: Wire up loongson_ipi device
Date: Tue, 18 Jun 2024 13:24:01 +0200	[thread overview]
Message-ID: <972034d6-23b3-415a-b401-b8bc1cc515c9@linaro.org> (raw)
In-Reply-To: <3561837a-895b-4e5f-bc40-bdf101cb38cb@linaro.org>

On 18/6/24 13:17, Philippe Mathieu-Daudé wrote:
> On 5/6/24 04:15, Jiaxun Yang wrote:
>> Wire up loongson_ipi device for loongson3_virt machine, so we
>> can have SMP support for TCG backend as well.
>>
>> Signed-off-by: Jiaxun Yang <jiaxun.yang@flygoat.com>
>> ---
>>   hw/mips/Kconfig           |  1 +
>>   hw/mips/loongson3_bootp.c |  2 --
>>   hw/mips/loongson3_bootp.h |  3 +++
>>   hw/mips/loongson3_virt.c  | 39 +++++++++++++++++++++++++++++++++++++--
>>   4 files changed, 41 insertions(+), 4 deletions(-)
> 
> 
>> @@ -527,6 +531,19 @@ static void mips_loongson3_virt_init(MachineState 
>> *machine)
>>       create_unimplemented_device("mmio fallback 0", 0x10000000, 256 * 
>> MiB);
>>       create_unimplemented_device("mmio fallback 1", 0x30000000, 256 * 
>> MiB);
>> +    memory_region_init(iocsr, OBJECT(machine), "loongson3.iocsr", 
>> UINT32_MAX);
>> +
>> +    /* IPI controller is in kernel for KVM */
>> +    if (!kvm_enabled()) {
> 
> Generically one could come with another hypervisor support, so better
> to check for what you are expecting. You could see some uses of:
> 
>    if (tcg) ... else if (kvm) ... else abort().
> 
>> +        ipi = qdev_new(TYPE_LOONGSON_IPI);
>> +        qdev_prop_set_uint32(ipi, "num-cpu", machine->smp.cpus);
>> +        sysbus_realize_and_unref(SYS_BUS_DEVICE(ipi), &error_fatal);
>> +        memory_region_add_subregion(iocsr, SMP_IPI_MAILBOX,
>> +            sysbus_mmio_get_region(SYS_BUS_DEVICE(ipi), 0));
>> +        memory_region_add_subregion(iocsr, MAIL_SEND_ADDR,
>> +            sysbus_mmio_get_region(SYS_BUS_DEVICE(ipi), 1));
>> +    }
>> +
>>       liointc = qdev_new("loongson.liointc");
>>       sysbus_realize_and_unref(SYS_BUS_DEVICE(liointc), &error_fatal);
>> @@ -543,6 +560,8 @@ static void mips_loongson3_virt_init(MachineState 
>> *machine)
>>       clock_set_hz(cpuclk, DEF_LOONGSON3_FREQ);
>>       for (i = 0; i < machine->smp.cpus; i++) {
>> +        int node = i / LOONGSON3_CORE_PER_NODE;
>> +        int core = i % LOONGSON3_CORE_PER_NODE;
>>           int ip;
>>           /* init CPUs */
>> @@ -553,12 +572,28 @@ static void 
>> mips_loongson3_virt_init(MachineState *machine)
>>           cpu_mips_clock_init(cpu);
>>           qemu_register_reset(main_cpu_reset, cpu);
>> -        if (i >= 4) {
>> +        if (ipi) {
>> +            hwaddr base = ((hwaddr)node << 44) + 
>> virt_memmap[VIRT_IPI].base;
>> +            base += core * 0x100;
>> +            qdev_connect_gpio_out(ipi, i, cpu->env.irq[6]);
>> +            sysbus_mmio_map(SYS_BUS_DEVICE(ipi), i + 2, base);
>> +        }
>> +
>> +        if (ase_lcsr_available(&MIPS_CPU(cpu)->env)) {
>> +            MemoryRegion *core_iocsr = g_new(MemoryRegion, 1);

Please store this in LoongsonMachineState (new patch):

   MemoryRegion core_iocsr[LOONGSON_MAX_VCPUS];

>> +            g_autofree char *name = 
>> g_strdup_printf("loongson3.core%d_iocsr", i);
>> +            memory_region_init_alias(core_iocsr, OBJECT(cpu), name,
>> +                                     iocsr, 0, UINT32_MAX);
>> +            memory_region_add_subregion(&MIPS_CPU(cpu)->env.iocsr.mr,
>> +                                        0, core_iocsr);
>> +        }
>> +
>> +        if (node > 0) {
>>               continue; /* Only node-0 can be connected to LIOINTC */
>>           }
> 
> Pre-existing, but the logic appears clearer as:
> 
>            if (node == 0) {
>>           for (ip = 0; ip < 4 ; ip++) {
>> -            int pin = i * 4 + ip;
>> +            int pin = core * LOONGSON3_CORE_PER_NODE + ip;
>>               sysbus_connect_irq(SYS_BUS_DEVICE(liointc),
>>                                  pin, cpu->env.irq[ip + 2]);
>>           }
>>
> 
> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
> 



  reply	other threads:[~2024-06-18 11:24 UTC|newest]

Thread overview: 27+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-06-05  2:15 [PATCH v3 0/4] hw/mips/loongson3_virt: Implement IPI support Jiaxun Yang
2024-06-05  2:15 ` [PATCH v3 1/4] hw/intc: Remove loongarch_ipi.c Jiaxun Yang
2024-06-05  3:32   ` gaosong
2024-06-18 11:09   ` Philippe Mathieu-Daudé
2024-06-26  4:11   ` maobibo
2024-06-26  7:40     ` Jiaxun Yang
2024-06-26  7:51       ` maobibo
2024-06-26 10:03         ` Jiaxun Yang
2024-06-26 12:10     ` Philippe Mathieu-Daudé
2024-06-27  2:44       ` gaosong
2024-06-27 13:02         ` Philippe Mathieu-Daudé
2024-07-01  1:35           ` maobibo
2024-07-01  6:44             ` maobibo
2024-07-01  7:01               ` Jiaxun Yang
2024-07-01  7:22                 ` maobibo
2024-07-01  8:29                   ` Jiaxun Yang
2024-07-01  8:51                     ` maobibo
2024-07-01  7:08             ` Jiaxun Yang
2024-07-01  7:24               ` maobibo
2024-06-05  2:15 ` [PATCH v3 2/4] hw/intc/loongson_ipi: Provide per core MMIO address spaces Jiaxun Yang
2024-06-05  2:15 ` [PATCH v3 3/4] hw/mips/loongson3_virt: Wire up loongson_ipi device Jiaxun Yang
2024-06-18 11:17   ` Philippe Mathieu-Daudé
2024-06-18 11:24     ` Philippe Mathieu-Daudé [this message]
2024-06-05  2:15 ` [PATCH v3 4/4] hw/intc/loongson_ipi: Replace ipi_getcpu with cpu_by_arch_id Jiaxun Yang
2024-06-05  3:33   ` gaosong
2024-06-12 12:09 ` [PATCH v3 0/4] hw/mips/loongson3_virt: Implement IPI support Jiaxun Yang
2024-06-18 11:49   ` Philippe Mathieu-Daudé

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=972034d6-23b3-415a-b401-b8bc1cc515c9@linaro.org \
    --to=philmd@linaro.org \
    --cc=chenhuacai@kernel.org \
    --cc=gaosong@loongson.cn \
    --cc=jiaxun.yang@flygoat.com \
    --cc=qemu-devel@nongnu.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).