On 07.12.2012, at 13:50, Cornelia Huck wrote:
> Add a new virtio transport that uses channel commands to perform
> virtio operations.
>
> Add a new machine type s390-ccw that uses this virtio-ccw transport
> and make it the default machine for s390.
>
> Signed-off-by: Cornelia Huck <cornelia.huck@xxxxxxxxxx>
> ---
> hw/s390-virtio.c | 149 ++++++--
> hw/s390x/Makefile.objs | 1 +
> hw/s390x/virtio-ccw.c | 909 +++++++++++++++++++++++++++++++++++++++++++++++++
> hw/s390x/virtio-ccw.h | 81 +++++
> trace-events | 4 +
> 5 files changed, 1124 insertions(+), 20 deletions(-)
> create mode 100644 hw/s390x/virtio-ccw.c
> create mode 100644 hw/s390x/virtio-ccw.h
>
> diff --git a/hw/s390-virtio.c b/hw/s390-virtio.c
> index 9e1afb2..f29ff74 100644
> --- a/hw/s390-virtio.c
> +++ b/hw/s390-virtio.c
> @@ -33,6 +33,8 @@
>
> #include "hw/s390-virtio-bus.h"
> #include "hw/s390x/sclp.h"
> +#include "hw/s390x/css.h"
> +#include "hw/s390x/virtio-ccw.h"
>
> //#define DEBUG_S390
>
> @@ -47,6 +49,7 @@
> #define KVM_S390_VIRTIO_NOTIFY 0
> #define KVM_S390_VIRTIO_RESET 1
> #define KVM_S390_VIRTIO_SET_STATUS 2
> +#define KVM_S390_VIRTIO_CCW_NOTIFY 3
>
> #define KERN_IMAGE_START 0x010000UL
> #define KERN_PARM_AREA 0x010480UL
> @@ -63,6 +66,7 @@
>
> static VirtIOS390Bus *s390_bus;
> static S390CPU **ipi_states;
> +VirtioCcwBus *ccw_bus;
>
> S390CPU *s390_cpu_addr2state(uint16_t cpu_addr)
> {
> @@ -76,15 +80,21 @@ S390CPU *s390_cpu_addr2state(uint16_t cpu_addr)
> int s390_virtio_hypercall(CPUS390XState *env, uint64_t mem, uint64_t hypercall)
> {
> int r = 0, i;
> + int cssid, ssid, schid, m;
> + SubchDev *sch;
>
> dprintf("KVM hypercall: %ld\n", hypercall);
> switch (hypercall) {
> case KVM_S390_VIRTIO_NOTIFY:
> if (mem > ram_size) {
> - VirtIOS390Device *dev = s390_virtio_bus_find_vring(s390_bus,
> - mem, &i);
> - if (dev) {
> - virtio_queue_notify(dev->vdev, i);
> + if (s390_bus) {
> + VirtIOS390Device *dev = s390_virtio_bus_find_vring(s390_bus,
> + mem, &i);
> + if (dev) {
> + virtio_queue_notify(dev->vdev, i);
> + } else {
> + r = -EINVAL;
> + }
We really want to factor out the DIAG handling code similar to how spapr handles its hypercalls. That way the legacy s390-virtio machine can register a VIRTIO_NOTIFY hypercall that works for it here, while the s390-virtio-ccw machine doesn't.
> } else {
> r = -EINVAL;
> }
> @@ -93,28 +103,49 @@ int s390_virtio_hypercall(CPUS390XState *env, uint64_t mem, uint64_t hypercall)
> }
> break;
> case KVM_S390_VIRTIO_RESET:
> - {
> - VirtIOS390Device *dev;
> -
> - dev = s390_virtio_bus_find_mem(s390_bus, mem);
> - virtio_reset(dev->vdev);
> - stb_phys(dev->dev_offs + VIRTIO_DEV_OFFS_STATUS, 0);
> - s390_virtio_device_sync(dev);
> - s390_virtio_reset_idx(dev);
> + if (s390_bus) {
> + VirtIOS390Device *dev;
> +
> + dev = s390_virtio_bus_find_mem(s390_bus, mem);
> + virtio_reset(dev->vdev);
> + stb_phys(dev->dev_offs + VIRTIO_DEV_OFFS_STATUS, 0);
> + s390_virtio_device_sync(dev);
> + s390_virtio_reset_idx(dev);
> + } else {
> + r = -EINVAL;
> + }
> break;
> - }
> case KVM_S390_VIRTIO_SET_STATUS:
> - {
> - VirtIOS390Device *dev;
> + if (s390_bus) {
> + VirtIOS390Device *dev;
>
> - dev = s390_virtio_bus_find_mem(s390_bus, mem);
> - if (dev) {
> - s390_virtio_device_update_status(dev);
> + dev = s390_virtio_bus_find_mem(s390_bus, mem);
> + if (dev) {
> + s390_virtio_device_update_status(dev);
> + } else {
> + r = -EINVAL;
> + }
> } else {
> r = -EINVAL;
> }
> break;
> - }
> + case KVM_S390_VIRTIO_CCW_NOTIFY:
> + if (ccw_bus) {
> + if (ioinst_disassemble_sch_ident(env->regs[2], &m, &cssid, &ssid,
> + &schid)) {
> + r = -EINVAL;
> + } else {
> + sch = css_find_subch(m, cssid, ssid, schid);
> + if (sch && css_subch_visible(sch)) {
> + virtio_queue_notify(virtio_ccw_get_vdev(sch), env->regs[3]);
> + } else {
> + r = -EINVAL;
> + }
> + }
> + } else {
> + r = -EINVAL;
> + }
> + break;
> default:
> r = -EINVAL;
> break;
> @@ -370,7 +401,6 @@ static QEMUMachine s390_machine = {
> .no_sdcard = 1,
> .use_virtcon = 1,
> .max_cpus = 255,
> - .is_default = 1,
> };
>
> static void s390_machine_init(void)
> @@ -379,3 +409,82 @@ static void s390_machine_init(void)
> }
>
> machine_init(s390_machine_init);
> +
> +static void ccw_init(QEMUMachineInitArgs *args)
Do you think it'd be possible to move this into its own .c file?
> +{
> + ram_addr_t my_ram_size = args->ram_size;
> + ram_addr_t ram_size = args->ram_size;
> + const char *cpu_model = args->cpu_model;
> + const char *kernel_filename = args->kernel_filename;
> + const char *kernel_cmdline = args->kernel_cmdline;
> + const char *initrd_filename = args->initrd_filename;
> + CPUS390XState *env = NULL;
> + MemoryRegion *sysmem = get_system_memory();
> + MemoryRegion *ram = g_new(MemoryRegion, 1);
> + int shift = 0;
> + uint8_t *storage_keys;
> + int ret;
> +
> + /* s390x ram size detection needs a 16bit multiplier + an increment. So
> + guests > 64GB can be specified in 2MB steps etc. */
> + while ((my_ram_size >> (20 + shift)) > 65535) {
> + shift++;
> + }
> + my_ram_size = my_ram_size >> (20 + shift) << (20 + shift);
> +
> + /* lets propagate the changed ram size into the global variable. */
> + ram_size = my_ram_size;
> +
> + /* get a BUS */
> + ccw_bus = virtio_ccw_bus_init();
> + s390_sclp_init();
> +
> + /* allocate RAM */
> + memory_region_init_ram(ram, "s390.ram", my_ram_size);
> + vmstate_register_ram_global(ram);
> + memory_region_add_subregion(sysmem, 0, ram);
> +
> + /* allocate storage keys */
> + storage_keys = g_malloc0(my_ram_size / TARGET_PAGE_SIZE);
> +
> + /* init CPUs */
> + env = s390_init_cpus(cpu_model, storage_keys);
> +
> + kvm_s390_enable_css_support(env);
> +
> + /*
> + * Create virtual css and set it as default so that non mcss-e
> + * enabled guests only see virtio devices.
> + */
> + ret = css_create_css_image(VIRTUAL_CSSID, true);
> + assert(ret == 0);
> +
> +
> + s390_set_up_kernel(env, kernel_filename, kernel_cmdline, initrd_filename);
> +
> + /* Create VirtIO network adapters */
> + s390_create_virtio_net((BusState *)ccw_bus, "virtio-net-ccw");
> +
> +}
> +
> +static QEMUMachine ccw_machine = {
> + .name = "s390-ccw-virtio",
> + .alias = "s390-ccw",
> + .desc = "VirtIO-ccw based S390 machine",
> + .init = ccw_init,
> + .no_cdrom = 1,
> + .no_floppy = 1,
> + .no_serial = 1,
> + .no_parallel = 1,
> + .no_sdcard = 1,
> + .use_virtcon = 1,
> + .max_cpus = 255,
> + .is_default = 1,
> +};
> +
> +static void ccw_machine_init(void)
> +{
> + qemu_register_machine(&ccw_machine);
> +}
> +
> +machine_init(ccw_machine_init);
> diff --git a/hw/s390x/Makefile.objs b/hw/s390x/Makefile.objs
> index 378b099..d408558 100644
> --- a/hw/s390x/Makefile.objs
> +++ b/hw/s390x/Makefile.objs
> @@ -5,3 +5,4 @@ obj-y += sclp.o
> obj-y += event-facility.o
> obj-y += sclpquiesce.o sclpconsole.o
> obj-y += css.o
> +obj-y += virtio-ccw.o
> diff --git a/hw/s390x/virtio-ccw.c b/hw/s390x/virtio-ccw.c
> new file mode 100644
> index 0000000..b4624ab
> --- /dev/null
> +++ b/hw/s390x/virtio-ccw.c
> @@ -0,0 +1,909 @@
> +/*
> + * virtio ccw target implementation
Please extract this into a separate patch.
> + *
> + * Copyright 2012 IBM Corp.
> + * Author(s): Cornelia Huck <cornelia.huck@xxxxxxxxxx>
> + *
> + * 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 <hw/hw.h>
> +#include "block.h"
> +#include "blockdev.h"
> +#include "sysemu.h"
> +#include "net.h"
> +#include "monitor.h"
> +#include "qemu-thread.h"
> +#include "hw/virtio.h"
> +#include "hw/virtio-serial.h"
> +#include "hw/virtio-net.h"
> +#include "hw/sysbus.h"
> +#include "bitops.h"
> +
> +#include "ioinst.h"
> +#include "css.h"
> +#include "virtio-ccw.h"
> +#include "trace.h"
> +
> +static const TypeInfo virtio_ccw_bus_info = {
> + .name = TYPE_VIRTIO_CCW_BUS,
> + .parent = TYPE_BUS,
> + .instance_size = sizeof(VirtioCcwBus),
> +};
> +
> +static const VirtIOBindings virtio_ccw_bindings;
> +
> +VirtIODevice *virtio_ccw_get_vdev(SubchDev *sch)
> +{
> + VirtIODevice *vdev = NULL;
> +
> + if (sch->driver_data) {
> + vdev = ((VirtioCcwData *)sch->driver_data)->vdev;
> + }
> + return vdev;
> +}
> +
> +static void virtio_ccw_reset_subchannels(void *opaque)
> +{
> + VirtioCcwBus *bus = opaque;
> + BusChild *kid;
> + VirtioCcwData *data;
> +
> + QTAILQ_FOREACH(kid, &bus->bus.children, sibling) {
> + data = (VirtioCcwData *)kid->child;
> + virtio_reset(data->vdev);
> + css_reset_sch(data->sch);
> + }
> + css_reset();
> +}
> +
> +VirtioCcwBus *virtio_ccw_bus_init(void)
> +{
> + VirtioCcwBus *cbus;
> + BusState *bus;
> + DeviceState *dev;
> +
> + /* Create bridge device */
> + dev = qdev_create(NULL, "virtio-ccw-bridge");
> + qdev_init_nofail(dev);
> +
> + /* Create bus on bridge device */
> + bus = qbus_create(TYPE_VIRTIO_CCW_BUS, dev, "virtio-ccw");
> + cbus = DO_UPCAST(VirtioCcwBus, bus, bus);
> +
> + /* Enable hotplugging */
> + bus->allow_hotplug = 1;
> +
> + qemu_register_reset(virtio_ccw_reset_subchannels, cbus);
> + return cbus;
> +}
> +
> +/* Communication blocks used by several channel commands. */
> +typedef struct VqInfoBlock {
> + uint64_t queue;
> + uint32_t align;
> + uint16_t index;
> + uint16_t num;
> +} QEMU_PACKED VqInfoBlock;
> +
> +typedef struct VqConfigBlock {
> + uint16_t index;
> + uint16_t num_max;
> +} QEMU_PACKED VqConfigBlock;
> +
> +typedef struct VirtioFeatDesc {
> + uint32_t features;
> + uint8_t index;
> +} QEMU_PACKED VirtioFeatDesc;
> +
> +/* Specify where the virtqueues for the subchannel are in guest memory. */
> +static int virtio_ccw_set_vqs(SubchDev *sch, uint64_t addr, uint32_t align,
> + uint16_t index, uint16_t num)
> +{
> + VirtioCcwData *data = sch->driver_data;
> +
> + if (index > VIRTIO_PCI_QUEUE_MAX) {
> + return -EINVAL;
> + }
> +
> + /* Current code in virtio.c relies on 4K alignment. */
> + if (addr && (align != 4096)) {
> + return -EINVAL;
> + }
> +
> + if (!data) {
> + return -EINVAL;
> + }
> +
> + virtio_queue_set_addr(data->vdev, index, addr);
> + if (!addr) {
> + virtio_queue_set_vector(data->vdev, index, 0);
> + } else {
> + /* Fail if we don't have a big enough queue. */
> + /* TODO: Add interface to handle vring.num changing */
> + if (virtio_queue_get_num(data->vdev, index) > num) {
> + return -EINVAL;
> + }
> + virtio_queue_set_vector(data->vdev, index, index);
> + }
> + /* tell notify handler in case of config change */
> + data->vdev->config_vector = VIRTIO_PCI_QUEUE_MAX;
> + return 0;
> +}
> +
> +static int virtio_ccw_cb(SubchDev *sch, CCW1 *ccw)
> +{
> + int ret;
> + VqInfoBlock info;
> + uint8_t status;
> + VirtioFeatDesc features;
> + void *config;
> + hwaddr indicators;
> + VqConfigBlock vq_config;
> + VirtioCcwData *data = sch->driver_data;
> + bool check_len;
> + int len;
> +
> + if (!ccw) {
> + return -EIO;
> + }
> +
> + if (!data) {
> + return -EINVAL;
> + }
> +
> + trace_virtio_ccw_interpret_ccw(sch->cssid, sch->ssid, sch->schid,
> + ccw->cmd_code);
> + check_len = !((ccw->flags & CCW_FLAG_SLI) && !(ccw->flags & CCW_FLAG_DC));
> +
> + /* Look at the command. */
> + switch (ccw->cmd_code) {
> + case CCW_CMD_SET_VQ:
> + if (check_len) {
> + if (ccw->count != sizeof(info)) {
> + ret = -EINVAL;
> + break;
> + }
> + } else if (ccw->count < sizeof(info)) {
> + /* Can't execute command. */
> + ret = -EINVAL;
> + break;
> + }
> + if (!qemu_get_ram_ptr(ccw->cda)) {
> + ret = -EFAULT;
> + } else {
> + info.queue = ldq_phys(ccw->cda);
> + info.align = ldl_phys(ccw->cda + sizeof(info.queue));
> + info.index = lduw_phys(ccw->cda + sizeof(info.queue)
> + + sizeof(info.align));
> + info.num = lduw_phys(ccw->cda + sizeof(info.queue)
> + + sizeof(info.align)
> + + sizeof(info.index));
> + ret = virtio_ccw_set_vqs(sch, info.queue, info.align, info.index,
> + info.num);
> + sch->curr_status.scsw.count = 0;
> + }
> + break;
> + case CCW_CMD_VDEV_RESET:
> + virtio_reset(data->vdev);
> + ret = 0;
> + break;
> + case CCW_CMD_READ_FEAT:
> + if (check_len) {
> + if (ccw->count != sizeof(features)) {
> + ret = -EINVAL;
> + break;
> + }
> + } else if (ccw->count < sizeof(features)) {
> + /* Can't execute command. */
> + ret = -EINVAL;
> + break;
> + }
> + if (!qemu_get_ram_ptr(ccw->cda)) {
> + ret = -EFAULT;
> + } else {
> + features.index = ldub_phys(ccw->cda + sizeof(features.features));
> + if (features.index < ARRAY_SIZE(data->host_features)) {
> + features.features = data->host_features[features.index];
> + } else {
> + /* Return zeroes if the guest supports more feature bits. */
> + features.features = 0;
> + }
> + stl_le_phys(ccw->cda, features.features);
> + sch->curr_status.scsw.count = ccw->count - sizeof(features);
> + ret = 0;
> + }
> + break;
> + case CCW_CMD_WRITE_FEAT:
> + if (check_len) {
> + if (ccw->count != sizeof(features)) {
> + ret = -EINVAL;
> + break;
> + }
> + } else if (ccw->count < sizeof(features)) {
> + /* Can't execute command. */
> + ret = -EINVAL;
> + break;
> + }
> + if (!qemu_get_ram_ptr(ccw->cda)) {
> + ret = -EFAULT;
> + } else {
> + features.index = ldub_phys(ccw->cda + sizeof(features.features));
> + features.features = ldl_le_phys(ccw->cda);
> + if (features.index < ARRAY_SIZE(data->host_features)) {
> + if (data->vdev->set_features) {
> + data->vdev->set_features(data->vdev, features.features);
> + }
> + data->vdev->guest_features = features.features;
> + } else {
> + /*
> + * If the guest supports more feature bits, assert that it
> + * passes us zeroes for those we don't support.
> + */
> + if (features.features) {
> + fprintf(stderr, "Guest bug: features[%i]=%x (expected 0)\n",
> + features.index, features.features);
> + /* XXX: do a unit check here? */
> + }
> + }
> + sch->curr_status.scsw.count = ccw->count - sizeof(features);
> + ret = 0;
> + }
> + break;
> + case CCW_CMD_READ_CONF:
> + if (check_len) {
> + if (ccw->count > data->vdev->config_len) {
> + ret = -EINVAL;
> + break;
> + }
> + }
> + len = MIN(ccw->count, data->vdev->config_len);
> + if (!qemu_get_ram_ptr(ccw->cda)) {
> + ret = -EFAULT;
> + } else {
> + data->vdev->get_config(data->vdev, data->vdev->config);
> + cpu_physical_memory_write(ccw->cda, data->vdev->config, len);
> + sch->curr_status.scsw.count = ccw->count - len;
> + ret = 0;
> + }
> + break;
> + case CCW_CMD_WRITE_CONF:
> + if (check_len) {
> + if (ccw->count > data->vdev->config_len) {
> + ret = -EINVAL;
> + break;
> + }
> + }
> + len = MIN(ccw->count, data->vdev->config_len);
> + config = qemu_get_ram_ptr(ccw->cda);
> + if (!config) {
> + ret = -EFAULT;
> + } else {
> + memcpy(data->vdev->config, config, len);
> + if (data->vdev->set_config) {
> + data->vdev->set_config(data->vdev, data->vdev->config);
> + }
> + sch->curr_status.scsw.count = ccw->count - len;
> + ret = 0;
> + }
> + break;
> + case CCW_CMD_WRITE_STATUS:
> + if (check_len) {
> + if (ccw->count != sizeof(status)) {
> + ret = -EINVAL;
> + break;
> + }
> + } else if (ccw->count < sizeof(status)) {
> + /* Can't execute command. */
> + ret = -EINVAL;
> + break;
> + }
> + if (!qemu_get_ram_ptr(ccw->cda)) {
> + ret = -EFAULT;
> + } else {
> + status = ldub_phys(ccw->cda);
> + virtio_set_status(data->vdev, status);
> + sch->curr_status.scsw.count = ccw->count - sizeof(status);
> + ret = 0;
> + }
> + break;
> + case CCW_CMD_SET_IND:
> + if (check_len) {
> + if (ccw->count != sizeof(indicators)) {
> + ret = -EINVAL;
> + break;
> + }
> + } else if (ccw->count < sizeof(indicators)) {
> + /* Can't execute command. */
> + ret = -EINVAL;
> + break;
> + }
> + indicators = ldq_phys(ccw->cda);
> + if (!indicators) {
> + ret = -EFAULT;
> + } else {
> + data->indicators = indicators;
> + sch->curr_status.scsw.count = ccw->count - sizeof(indicators);
> + ret = 0;
> + }
> + break;
> + case CCW_CMD_SET_CONF_IND:
> + if (check_len) {
> + if (ccw->count != sizeof(indicators)) {
> + ret = -EINVAL;
> + break;
> + }
> + } else if (ccw->count < sizeof(indicators)) {
> + /* Can't execute command. */
> + ret = -EINVAL;
> + break;
> + }
> + indicators = ldq_phys(ccw->cda);
> + if (!indicators) {
> + ret = -EFAULT;
> + } else {
> + data->indicators2 = indicators;
> + sch->curr_status.scsw.count = ccw->count - sizeof(indicators);
> + ret = 0;
> + }
> + break;
> + case CCW_CMD_READ_VQ_CONF:
> + if (check_len) {
> + if (ccw->count != sizeof(vq_config)) {
> + ret = -EINVAL;
> + break;
> + }
> + } else if (ccw->count < sizeof(vq_config)) {
> + /* Can't execute command. */
> + ret = -EINVAL;
> + break;
> + }
> + if (!qemu_get_ram_ptr(ccw->cda)) {
> + ret = -EFAULT;
> + } else {
> + vq_config.index = lduw_phys(ccw->cda);
> + vq_config.num_max = virtio_queue_get_num(data->vdev,
> + vq_config.index);
> + stw_phys(ccw->cda + sizeof(vq_config.index), vq_config.num_max);
> + sch->curr_status.scsw.count = ccw->count - sizeof(vq_config);
> + ret = 0;
> + }
> + break;
> + default:
> + ret = -EOPNOTSUPP;
> + break;
> + }
> + return ret;
> +}
> +
> +static int virtio_ccw_device_init(VirtioCcwData *dev, VirtIODevice *vdev)
> +{
> + unsigned int cssid = 0;
> + unsigned int ssid = 0;
> + unsigned int schid;
> + unsigned int devno;
> + bool have_devno = false;
> + bool found = false;
> + SubchDev *sch;
> + int ret;
> + int num;
> +
> + sch = g_malloc0(sizeof(SubchDev));
Any reason to not just make sch part of dev?
Alex
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
