Re: [Qemu-devel] [PATCH v6 5/6] Inter-VM shared memory PCI device

[Cam Macdonell]

On Sat, Jun 5, 2010 at 3:44 AM, Blue Swirl <address@hidden> wrote:
> On Fri, Jun 4, 2010 at 9:45 PM, Cam Macdonell <address@hidden> wrote:
>> Support an inter-vm shared memory device that maps a shared-memory object as 
>> a
>> PCI device in the guest.  This patch also supports interrupts between guest 
>> by
>> communicating over a unix domain socket.  This patch applies to the qemu-kvm
>> repository.
>>
>>    -device ivshmem,size=<size in format accepted by -m>[,shm=<shm name>]
>>
>> Interrupts are supported between multiple VMs by using a shared memory server
>> by using a chardev socket.
>>
>>    -device ivshmem,size=<size in format accepted by -m>[,shm=<shm name>]
>>           [,chardev=<id>][,msi=on][,irqfd=on][,vectors=n][,role=peer|master]
>>    -chardev socket,path=<path>,id=<id>
>>
>> (shared memory server is qemu.git/contrib/ivshmem-server)
>>
>> Sample programs and init scripts are in a git repo here:
>>
>>    www.gitorious.org/nahanni
>> ---
>>  Makefile.target |    3 +
>>  hw/ivshmem.c    |  852 
>> +++++++++++++++++++++++++++++++++++++++++++++++++++++++
>>  qemu-char.c     |    6 +
>>  qemu-char.h     |    3 +
>>  qemu-doc.texi   |   43 +++
>>  5 files changed, 907 insertions(+), 0 deletions(-)
>>  create mode 100644 hw/ivshmem.c
>>
>> diff --git a/Makefile.target b/Makefile.target
>> index c4ba592..4888308 100644
>> --- a/Makefile.target
>> +++ b/Makefile.target
>> @@ -202,6 +202,9 @@ obj-$(CONFIG_USB_OHCI) += usb-ohci.o
>>  obj-y += rtl8139.o
>>  obj-y += e1000.o
>>
>> +# Inter-VM PCI shared memory
>> +obj-y += ivshmem.o
>> +
>
> Can this be compiled once, simply by moving this to Makefile.objs
> instead of Makefile.target? Also, because the code seems to be KVM
> specific, it can't be compiled unconditionally but depending on at
> least CONFIG_KVM and maybe CONFIG_EVENTFD.

The device uses eventfds for signalling, but never creates the
eventfds itself as they are passed from a server using SCM_RIGHTS.
So, it does not depend on the eventfd API.  Its dependence on
irqfd/ioeventfd (the only KVM specific bits) are optional via the
command-line.

A runtime check for KVM is done for the reasons Avi mentioned.

>
> Why is this KVM specific BTW, Posix SHM is available on many
> platforms? What would happen if kvm_set_foobar functions were not
> called when KVM is not being used? Is host eventfd support essential?
>
>>  # Hardware support
>>  obj-i386-y += vga.o
>>  obj-i386-y += mc146818rtc.o i8259.o pc.o
>> diff --git a/hw/ivshmem.c b/hw/ivshmem.c
>> new file mode 100644
>> index 0000000..9057612
>> --- /dev/null
>> +++ b/hw/ivshmem.c
>> @@ -0,0 +1,852 @@
>> +/*
>> + * Inter-VM Shared Memory PCI device.
>> + *
>> + * Author:
>> + *      Cam Macdonell <address@hidden>
>> + *
>> + * Based On: cirrus_vga.c
>> + *          Copyright (c) 2004 Fabrice Bellard
>> + *          Copyright (c) 2004 Makoto Suzuki (suzu)
>> + *
>> + *      and rtl8139.c
>> + *          Copyright (c) 2006 Igor Kovalenko
>> + *
>> + * This code is licensed under the GNU GPL v2.
>> + */
>> +#include <sys/mman.h>
>> +#include <sys/types.h>
>> +#include <sys/socket.h>
>> +#include <sys/io.h>
>> +#include <sys/ioctl.h>
>> +#include "hw.h"
>> +#include "console.h"
>> +#include "pc.h"
>> +#include "pci.h"
>> +#include "sysemu.h"
>> +
>> +#include "msix.h"
>> +#include "qemu-kvm.h"
>> +#include "libkvm.h"
>> +
>> +#include <sys/eventfd.h>
>> +#include <sys/mman.h>
>> +#include <sys/socket.h>
>> +#include <sys/ioctl.h>
>> +
>> +#define IVSHMEM_IRQFD   0
>> +#define IVSHMEM_MSI     1
>> +
>> +//#define DEBUG_IVSHMEM
>> +#ifdef DEBUG_IVSHMEM
>> +#define IVSHMEM_DPRINTF(fmt, args...)        \
>> +    do {printf("IVSHMEM: " fmt, ##args); } while (0)
>
> Please use __VA_ARGS__.
>
>> +#else
>> +#define IVSHMEM_DPRINTF(fmt, args...)
>> +#endif
>> +
>> +typedef struct Peer {
>> +    int nb_eventfds;
>> +    int *eventfds;
>> +} Peer;
>> +
>> +typedef struct EventfdEntry {
>> +    PCIDevice *pdev;
>> +    int vector;
>> +} EventfdEntry;
>> +
>> +typedef struct IVShmemState {
>> +    PCIDevice dev;
>> +    uint32_t intrmask;
>> +    uint32_t intrstatus;
>> +    uint32_t doorbell;
>> +
>> +    CharDriverState ** eventfd_chr;
>
> I'd remove the space between '**' and 'eventfd_chr', it's used inconsistently.
>
>> +    CharDriverState * server_chr;
>> +    int ivshmem_mmio_io_addr;
>> +
>> +    pcibus_t mmio_addr;
>> +    pcibus_t shm_pci_addr;
>> +    uint64_t ivshmem_offset;
>> +    uint64_t ivshmem_size; /* size of shared memory region */
>> +    int shm_fd; /* shared memory file descriptor */
>> +
>> +    Peer *peers;
>> +    int nb_peers; /* how many guests we have space for */
>> +    int max_peer; /* maximum numbered peer */
>> +
>> +    int vm_id;
>> +    uint32_t vectors;
>> +    uint32_t features;
>> +    EventfdEntry *eventfd_table;
>> +
>> +    char * shmobj;
>> +    char * sizearg;
>> +    char * role;
>> +} IVShmemState;
>> +
>> +/* registers for the Inter-VM shared memory device */
>> +enum ivshmem_registers {
>> +    IntrMask = 0,
>> +    IntrStatus = 4,
>> +    IVPosition = 8,
>> +    Doorbell = 12,
>> +};
>
> IIRC these should be uppercase.

I worked from rtl8139 which doesn't have them uppercase.  But doing a
quick search, I can see most devices do, I'll fix that.

>
>> +
>> +static inline uint32_t ivshmem_has_feature(IVShmemState *ivs, int feature) {
>> +    return (ivs->features & (1 << feature));
>> +}
>
> Since this is the first version, do we need any features at this
> point, can't we expect that all features are available now? Why does
> the user need to specify the features?

Some features require host support such as irqfd/ioeventfds.  So
having features allows them to be disabled on the command-line (e.g.,
irqfd=off).

>
> To avoid a negative shift, I'd make 'feature' unsigned.
>
>> +
>> +static inline bool is_power_of_two(uint64_t x) {
>> +    return (x & (x - 1)) == 0;
>> +}
>> +
>> +static void ivshmem_map(PCIDevice *pci_dev, int region_num,
>> +                    pcibus_t addr, pcibus_t size, int type)
>> +{
>> +    IVShmemState *s = DO_UPCAST(IVShmemState, dev, pci_dev);
>> +
>> +    s->shm_pci_addr = addr;
>> +
>> +    if (s->ivshmem_offset > 0) {
>> +        cpu_register_physical_memory(s->shm_pci_addr, s->ivshmem_size,
>> +                                                            
>> s->ivshmem_offset);
>> +        if (s->role && strncmp(s->role, "peer", 4) == 0) {
>> +            IVSHMEM_DPRINTF("marking pages no migrate\n");
>> +            cpu_mark_pages_no_migrate(s->ivshmem_offset, s->ivshmem_size);
>> +        }
>> +    }
>> +
>> +    IVSHMEM_DPRINTF("guest pci addr = %u, guest h/w addr = %u, size = %u\n",
>> +                (uint32_t)addr, (uint32_t)s->ivshmem_offset, 
>> (uint32_t)size);
>
> Please use FMT_PCIBUS for addr and size and PRIu64 for s->ivshmem_offset.
>
>> +
>> +}
>> +
>> +/* accessing registers - based on rtl8139 */
>> +static void ivshmem_update_irq(IVShmemState *s, int val)
>> +{
>> +    int isr;
>> +    isr = (s->intrstatus & s->intrmask) & 0xffffffff;
>> +
>> +    /* don't print ISR resets */
>> +    if (isr) {
>> +        IVSHMEM_DPRINTF("Set IRQ to %d (%04x %04x)\n",
>> +           isr ? 1 : 0, s->intrstatus, s->intrmask);
>> +    }
>> +
>> +    qemu_set_irq(s->dev.irq[0], (isr != 0));
>> +}
>> +
>> +static void ivshmem_IntrMask_write(IVShmemState *s, uint32_t val)
>> +{
>> +    IVSHMEM_DPRINTF("IntrMask write(w) val = 0x%04x\n", val);
>> +
>> +    s->intrmask = val;
>> +
>> +    ivshmem_update_irq(s, val);
>> +}
>> +
>> +static uint32_t ivshmem_IntrMask_read(IVShmemState *s)
>> +{
>> +    uint32_t ret = s->intrmask;
>> +
>> +    IVSHMEM_DPRINTF("intrmask read(w) val = 0x%04x\n", ret);
>> +
>> +    return ret;
>> +}
>> +
>> +static void ivshmem_IntrStatus_write(IVShmemState *s, uint32_t val)
>> +{
>> +    IVSHMEM_DPRINTF("IntrStatus write(w) val = 0x%04x\n", val);
>> +
>> +    s->intrstatus = val;
>> +
>> +    ivshmem_update_irq(s, val);
>> +    return;
>> +}
>> +
>> +static uint32_t ivshmem_IntrStatus_read(IVShmemState *s)
>> +{
>> +    uint32_t ret = s->intrstatus;
>> +
>> +    /* reading ISR clears all interrupts */
>> +    s->intrstatus = 0;
>> +
>> +    ivshmem_update_irq(s, 0);
>> +
>> +    return ret;
>> +}
>> +
>> +static void ivshmem_io_writew(void *opaque, uint8_t addr, uint32_t val)
>> +{
>> +
>> +    IVSHMEM_DPRINTF("We shouldn't be writing words\n");
>> +}
>> +
>> +static void ivshmem_io_writel(void *opaque, uint8_t addr, uint32_t val)
>> +{
>> +    IVShmemState *s = opaque;
>> +
>> +    u_int64_t write_one = 1;
>
> Please use uintNN_t instead of u_intNN_t.
>
>> +    u_int16_t dest = val >> 16;
>> +    u_int16_t vector = val & 0xff;
>> +
>> +    addr &= 0xfc;
>
> I'd add a debug printf here, likewise for exit of ivshmem_io_readl().
> When you do the merge (see below), the correct printf format for the
> addresses will be TARGET_FMT_plx.
>
>> +
>> +    switch (addr)
>> +    {
>> +        case IntrMask:
>> +            ivshmem_IntrMask_write(s, val);
>> +            break;
>> +
>> +        case IntrStatus:
>> +            ivshmem_IntrStatus_write(s, val);
>> +            break;
>> +
>> +        case Doorbell:
>> +            /* check that dest VM ID is reasonable */
>> +            if ((dest < 0) || (dest > s->max_peer)) {
>> +                IVSHMEM_DPRINTF("Invalid destination VM ID (%d)\n", dest);
>> +                break;
>> +            }
>> +
>> +            /* check doorbell range */
>> +            if ((vector >= 0) && (vector < s->peers[dest].nb_eventfds)) {
>> +                IVSHMEM_DPRINTF("Writing %ld to VM %d on vector %d\n",
>> +                                                    write_one, dest, 
>> vector);
>
> PRId64 for write_one, %ld is not enough on ILP32.
>
>> +                if (write(s->peers[dest].eventfds[vector],
>> +                                                    &(write_one), 8) != 8) {
>> +                    IVSHMEM_DPRINTF("error writing to eventfd\n");
>> +                }
>> +            }
>> +            break;
>> +        default:
>> +            IVSHMEM_DPRINTF("Invalid VM Doorbell VM %d\n", dest);
>> +    }
>> +}
>> +
>> +static void ivshmem_io_writeb(void *opaque, uint8_t addr, uint32_t val)
>> +{
>> +    IVSHMEM_DPRINTF("We shouldn't be writing bytes\n");
>> +}
>> +
>> +static uint32_t ivshmem_io_readw(void *opaque, uint8_t addr)
>> +{
>> +
>> +    IVSHMEM_DPRINTF("We shouldn't be reading words\n");
>> +    return 0;
>> +}
>> +
>> +static uint32_t ivshmem_io_readl(void *opaque, uint8_t addr)
>> +{
>> +
>> +    IVShmemState *s = opaque;
>> +    uint32_t ret;
>> +
>> +    switch (addr)
>> +    {
>> +        case IntrMask:
>> +            ret = ivshmem_IntrMask_read(s);
>> +            break;
>> +
>> +        case IntrStatus:
>> +            ret = ivshmem_IntrStatus_read(s);
>> +            break;
>> +
>> +        case IVPosition:
>> +            /* return my VM ID if the memory is mapped */
>> +            if (s->shm_fd > 0) {
>> +                ret = s->vm_id;
>> +            } else {
>> +                ret = -1;
>> +            }
>> +            break;
>> +
>> +        default:
>> +            IVSHMEM_DPRINTF("why are we reading 0x%x\n", addr);
>> +            ret = 0;
>> +    }
>> +
>> +    return ret;
>> +}
>> +
>> +static uint32_t ivshmem_io_readb(void *opaque, uint8_t addr)
>> +{
>> +    IVSHMEM_DPRINTF("We shouldn't be reading bytes\n");
>> +
>> +    return 0;
>> +}
>> +
>> +static void ivshmem_mmio_writeb(void *opaque,
>> +                                target_phys_addr_t addr, uint32_t val)
>> +{
>> +    ivshmem_io_writeb(opaque, addr & 0xFF, val);
>> +}
>
> This function and others below only performs a cast and useless
> masking (the address passed is these days an offset from start of the
> area). Please merge these to ivshmem_io_readl() etc.

Ok, these are artifacts from basing my patch on rtl8139.c.  I'll remove them.

>
>> +
>> +static void ivshmem_mmio_writew(void *opaque,
>> +                                target_phys_addr_t addr, uint32_t val)
>> +{
>> +    ivshmem_io_writew(opaque, addr & 0xFF, val);
>> +}
>> +
>> +static void ivshmem_mmio_writel(void *opaque,
>> +                                target_phys_addr_t addr, uint32_t val)
>> +{
>> +    ivshmem_io_writel(opaque, addr & 0xFF, val);
>> +}
>> +
>> +static uint32_t ivshmem_mmio_readb(void *opaque, target_phys_addr_t addr)
>> +{
>> +    return ivshmem_io_readb(opaque, addr & 0xFF);
>> +}
>> +
>> +static uint32_t ivshmem_mmio_readw(void *opaque, target_phys_addr_t addr)
>> +{
>> +    uint32_t val = ivshmem_io_readw(opaque, addr & 0xFF);
>> +    return val;
>> +}
>> +
>> +static uint32_t ivshmem_mmio_readl(void *opaque, target_phys_addr_t addr)
>> +{
>> +    uint32_t val = ivshmem_io_readl(opaque, addr & 0xFF);
>> +    return val;
>> +}
>> +
>> +static CPUReadMemoryFunc *ivshmem_mmio_read[3] = {
>
> Please add 'const'.
>
>> +    ivshmem_mmio_readb,
>> +    ivshmem_mmio_readw,
>> +    ivshmem_mmio_readl,
>> +};
>> +
>> +static CPUWriteMemoryFunc *ivshmem_mmio_write[3] = {
>> +    ivshmem_mmio_writeb,
>> +    ivshmem_mmio_writew,
>> +    ivshmem_mmio_writel,
>> +};
>> +
>> +static void ivshmem_receive(void *opaque, const uint8_t *buf, int size)
>> +{
>> +    IVShmemState *s = opaque;
>> +
>> +    ivshmem_IntrStatus_write(s, *buf);
>> +
>> +    IVSHMEM_DPRINTF("ivshmem_receive 0x%02x\n", *buf);
>> +}
>> +
>> +static int ivshmem_can_receive(void * opaque)
>> +{
>> +    return 8;
>> +}
>> +
>> +static void ivshmem_event(void *opaque, int event)
>> +{
>> +    IVSHMEM_DPRINTF("ivshmem_event %d\n", event);
>> +}
>> +
>> +static void fake_irqfd(void *opaque, const uint8_t *buf, int size) {
>> +
>> +    EventfdEntry *entry = opaque;
>> +    PCIDevice *pdev = entry->pdev;
>> +
>> +    IVSHMEM_DPRINTF("fake irqfd on vector %p %d\n", pdev, entry->vector);
>> +    msix_notify(pdev, entry->vector);
>> +}
>> +
>> +static CharDriverState* create_eventfd_chr_device(void * opaque, int 
>> eventfd,
>> +                                                                    int 
>> vector)
>> +{
>> +    /* create a event character device based on the passed eventfd */
>> +    IVShmemState *s = opaque;
>> +    CharDriverState * chr;
>> +
>> +    chr = qemu_chr_open_eventfd(eventfd);
>> +
>> +    if (chr == NULL) {
>> +        IVSHMEM_DPRINTF("creating eventfd for eventfd %d failed\n", 
>> eventfd);
>
> This should not be a DPRINTF.
>
>> +        exit(-1);
>> +    }
>> +
>> +    /* if MSI is supported we need multiple interrupts */
>> +    if (ivshmem_has_feature(s, IVSHMEM_MSI)) {
>> +        s->eventfd_table[vector].pdev = &s->dev;
>> +        s->eventfd_table[vector].vector = vector;
>> +
>> +        qemu_chr_add_handlers(chr, ivshmem_can_receive, fake_irqfd,
>> +                      ivshmem_event, &s->eventfd_table[vector]);
>> +    } else {
>> +        qemu_chr_add_handlers(chr, ivshmem_can_receive, ivshmem_receive,
>> +                      ivshmem_event, s);
>> +    }
>> +
>> +    return chr;
>> +
>> +}
>> +
>> +static int check_shm_size(IVShmemState *s, int fd) {
>> +    /* check that the guest isn't going to try and map more memory than the
>> +     * the object has allocated return -1 to indicate error */
>> +
>> +    struct stat buf;
>> +
>> +    fstat(fd, &buf);
>> +
>> +    if (s->ivshmem_size > buf.st_size) {
>> +        fprintf(stderr, "IVSHMEM ERROR: Requested memory size greater");
>> +        fprintf(stderr, " than shared object size (%ld > %ld)\n",
>> +                                          s->ivshmem_size, buf.st_size);
>
> Please use PRIx64 for s->ivshmem_size, this will cause a warning on ILP32.
>
>> +        return -1;
>> +    } else {
>> +        return 0;
>> +    }
>> +}
>> +
>> +/* create the shared memory BAR when we are not using the server, so we can
>> + * create the BAR and map the memory immediately */
>> +static void create_shared_memory_BAR(IVShmemState *s, int fd) {
>> +
>> +    void * ptr;
>> +
>> +    s->shm_fd = fd;
>> +
>> +    ptr = mmap(0, s->ivshmem_size, PROT_READ|PROT_WRITE, MAP_SHARED, fd, 0);
>> +
>> +    s->ivshmem_offset = qemu_ram_map(s->ivshmem_size, ptr);
>
> qemu_ram_map() does not exist in HEAD.
>

Ok, so qemu_ram_map() and kvm_set_irq() are in the KVM HEAD.  I had my
own version of both functions, but removed them when Marcelo's were
merged into KVM.

>> +
>> +    /* region for shared memory */
>> +    pci_register_bar(&s->dev, 2, s->ivshmem_size,
>> +                                    PCI_BASE_ADDRESS_SPACE_MEMORY, 
>> ivshmem_map);
>> +}
>> +
>> +static void close_guest_eventfds(IVShmemState *s, int posn)
>> +{
>> +    int i, guest_curr_max;
>> +
>> +    guest_curr_max = s->peers[posn].nb_eventfds;
>> +
>> +    for (i = 0; i < guest_curr_max; i++)
>> +        close(s->peers[posn].eventfds[i]);
>
> CODING_STYLE.
>
>> +
>> +    qemu_free(s->peers[posn].eventfds);
>> +    s->peers[posn].nb_eventfds = 0;
>> +}
>> +
>> +static void setup_ioeventfds(IVShmemState *s) {
>> +
>> +    int i, j;
>> +
>> +    for (i = 0; i <= s->max_peer; i++) {
>> +        for (j = 0; j < s->peers[i].nb_eventfds; j++) {
>> +            kvm_set_ioeventfd_mmio_long(s->peers[i].eventfds[j],
>> +                    s->mmio_addr + Doorbell, (i << 16) | j, 1);
>> +        }
>> +    }
>> +
>> +    /* setup irqfd for this VM's eventfds */
>> +    for (i = 0; i < s->vectors; i++) {
>> +        kvm_set_irqfd(s->dev.msix_irq_entries[i].gsi,
>> +                        s->peers[s->vm_id].eventfds[i], 1);
>
> kvm_set_irqfd() does not exist in HEAD.
>
>> +    }
>> +}
>> +
>> +
>> +/* this function increase the dynamic storage need to store data about other
>> + * guests */
>> +static void increase_dynamic_storage(IVShmemState *s, int new_min_size) {
>> +
>> +    int j, old_nb_alloc;
>> +
>> +    old_nb_alloc = s->nb_peers;
>> +
>> +    while (new_min_size >= s->nb_peers)
>> +        s->nb_peers = s->nb_peers * 2;
>> +
>> +    IVSHMEM_DPRINTF("bumping storage to %d guests\n", s->nb_peers);
>> +    s->peers = qemu_realloc(s->peers, s->nb_peers * sizeof(Peer));
>> +
>> +    if (s->peers == NULL) {
>> +        fprintf(stderr, "Allocation error - exiting\n");
>> +        exit(1);
>> +    }
>
> qemu_realloc will never return zero.
>
>> +
>> +    /* zero out new pointers */
>> +    for (j = old_nb_alloc; j < s->nb_peers; j++) {
>> +        s->peers[j].eventfds = NULL;
>> +        s->peers[j].nb_eventfds = 0;
>> +    }
>> +}
>> +
>> +static void ivshmem_read(void *opaque, const uint8_t * buf, int flags)
>> +{
>> +    IVShmemState *s = opaque;
>> +    int incoming_fd, tmp_fd;
>> +    int guest_curr_max;
>> +    long incoming_posn;
>> +
>> +    memcpy(&incoming_posn, buf, sizeof(long));
>> +    /* pick off s->server_chr->msgfd and store it, posn should accompany 
>> msg */
>> +    tmp_fd = qemu_chr_get_msgfd(s->server_chr);
>> +    IVSHMEM_DPRINTF("posn is %ld, fd is %d\n", incoming_posn, tmp_fd);
>> +
>> +    /* make sure we have enough space for this guest */
>> +    if (incoming_posn >= s->nb_peers) {
>> +        increase_dynamic_storage(s, incoming_posn);
>> +    }
>> +
>> +    if (tmp_fd == -1) {
>> +        /* if posn is positive and unseen before then this is our posn*/
>> +        if ((incoming_posn >= 0) && (s->peers[incoming_posn].eventfds == 
>> NULL)) {
>> +            /* receive our posn */
>> +            s->vm_id = incoming_posn;
>> +            return;
>> +        } else {
>> +            /* otherwise an fd == -1 means an existing guest has gone away 
>> */
>> +            IVSHMEM_DPRINTF("posn %ld has gone away\n", incoming_posn);
>> +            close_guest_eventfds(s, incoming_posn);
>> +            return;
>> +        }
>> +    }
>> +
>> +    /* because of the implementation of get_msgfd, we need a dup */
>> +    incoming_fd = dup(tmp_fd);
>> +
>> +    if (incoming_fd == -1) {
>> +        fprintf(stderr, "could not allocate file descriptor %s\n",
>> +                                                            
>> strerror(errno));
>> +        return;
>> +    }
>> +
>> +    /* if the position is -1, then it's shared memory region fd */
>> +    if (incoming_posn == -1) {
>> +
>> +        void * map_ptr;
>> +
>> +        s->max_peer = 0;
>> +
>> +        if (check_shm_size(s, incoming_fd) == -1) {
>> +            exit(-1);
>> +        }
>> +
>> +        /* mmap the region and map into the BAR2 */
>> +        map_ptr = mmap(0, s->ivshmem_size, PROT_READ|PROT_WRITE, MAP_SHARED,
>> +                                                                
>> incoming_fd, 0);
>> +        s->ivshmem_offset = qemu_ram_map(s->ivshmem_size, map_ptr);
>> +
>> +        IVSHMEM_DPRINTF("guest pci addr = %u, guest h/w addr = %u, size = 
>> %u\n",
>> +                        (uint32_t)s->shm_pci_addr, 
>> (uint32_t)s->ivshmem_offset,
>> +                        (uint32_t)s->ivshmem_size);
>> +
>> +        if (s->shm_pci_addr > 0) {
>> +            /* map memory into BAR2 */
>> +            cpu_register_physical_memory(s->shm_pci_addr, s->ivshmem_size,
>> +                                                            
>> s->ivshmem_offset);
>> +            if (s->role && strncmp(s->role, "peer", 4) == 0) {
>> +                IVSHMEM_DPRINTF("marking pages no migrate\n");
>> +                cpu_mark_pages_no_migrate(s->ivshmem_offset, 
>> s->ivshmem_size);
>> +            }
>> +
>> +        }
>> +
>> +        /* only store the fd if it is successfully mapped */
>> +        s->shm_fd = incoming_fd;
>> +
>> +        return;
>> +    }
>> +
>> +    /* each guest has an array of eventfds, and we keep track of how many
>> +     * guests for each VM */
>> +    guest_curr_max = s->peers[incoming_posn].nb_eventfds;
>> +    if (guest_curr_max == 0) {
>> +        /* one eventfd per MSI vector */
>> +        s->peers[incoming_posn].eventfds = (int *) qemu_malloc(s->vectors *
>> +                                                                
>> sizeof(int));
>
> Useless cast in C.
>
>> +    }
>> +
>> +    /* this is an eventfd for a particular guest VM */
>> +    IVSHMEM_DPRINTF("eventfds[%ld][%d] = %d\n", incoming_posn, 
>> guest_curr_max,
>> +                                                                
>> incoming_fd);
>> +    s->peers[incoming_posn].eventfds[guest_curr_max] = incoming_fd;
>> +
>> +    /* increment count for particular guest */
>> +    s->peers[incoming_posn].nb_eventfds++;
>> +
>> +    /* keep track of the maximum VM ID */
>> +    if (incoming_posn > s->max_peer) {
>> +        s->max_peer = incoming_posn;
>> +    }
>> +
>> +    if (incoming_posn == s->vm_id) {
>> +        int vector = guest_curr_max;
>
> Why add a new variable?

for clarity, so when looking at the code it's clear that the current
maxium is a MSI vector.  Perhaps I'll rename guest_curr_max to achieve
this.

>
>> +        if (!ivshmem_has_feature(s, IVSHMEM_IRQFD)) {
>> +            /* initialize char device for callback
>> +             * if this is one of my eventfds */
>> +            s->eventfd_chr[vector] = create_eventfd_chr_device(s,
>> +                       s->peers[s->vm_id].eventfds[vector], vector);
>> +        }
>> +    }
>> +
>> +    return;
>> +}
>> +
>> +static void ivshmem_reset(DeviceState *d)
>> +{
>> +    return;
>
> Nothing to do?

Oversight, will fix.

>
>> +}
>> +
>> +static void ivshmem_mmio_map(PCIDevice *pci_dev, int region_num,
>> +                       pcibus_t addr, pcibus_t size, int type)
>> +{
>> +    IVShmemState *s = DO_UPCAST(IVShmemState, dev, pci_dev);
>> +
>> +    s->mmio_addr = addr;
>> +    cpu_register_physical_memory(addr + 0, 0x400, s->ivshmem_mmio_io_addr);
>
> The 0x400 should be #defined earlier. Why 0x400 since you are only
> interested in the 0x100 first bytes?

We bumped to 1MB for a previous version that had multiple doorbells.
ioeventfd masks eliminated made multiple doorbells unnecessary.  I'll
put it back to 0x100.

>
>> +
>> +    /* ioeventfd and irqfd are enabled together,
>> +     * so the flag IRQFD refers to both */
>> +    if (ivshmem_has_feature(s, IVSHMEM_IRQFD)) {
>> +        setup_ioeventfds(s);
>> +    }
>> +}
>> +
>> +static uint64_t ivshmem_get_size(IVShmemState * s) {
>> +
>> +    uint64_t value;
>> +    char *ptr;
>> +
>> +    value = strtoul(s->sizearg, &ptr, 10);
>
> I'd use strtoull() but the whole function should be suppressed, see below.
>
>> +    switch (*ptr) {
>> +        case 0: case 'M': case 'm':
>> +            value <<= 20;
>> +            break;
>> +        case 'G': case 'g':
>> +            value <<= 30;
>> +            break;
>> +        default:
>> +            fprintf(stderr, "qemu: invalid ram size: %s\n", s->sizearg);
>> +            exit(1);
>> +    }
>> +
>> +    /* BARs must be a power of 2 */
>> +    if (!is_power_of_two(value)) {
>> +        fprintf(stderr, "ivshmem: size must be power of 2\n");
>> +        exit(1);
>> +    }
>
> Isn't the BAR check in pci.c enough?

This would produce a clearer error that it is the ivshmem device that
is the problem.

>
>> +
>> +    return value;
>> +
>> +}
>> +
>> +static void ivshmem_setup_msi(IVShmemState * s) {
>> +
>> +    int i;
>> +
>> +    /* allocate the MSI-X vectors */
>> +
>> +    if (!msix_init(&s->dev, s->vectors, 1, 0)) {
>> +        pci_register_bar(&s->dev, 1,
>> +                         msix_bar_size(&s->dev),
>> +                         PCI_BASE_ADDRESS_SPACE_MEMORY,
>> +                         msix_mmio_map);
>> +        IVSHMEM_DPRINTF("msix initialized (%d vectors)\n", s->vectors);
>> +    } else {
>> +        IVSHMEM_DPRINTF("msix initialization failed\n");
>
> Is this fatal considering the msix_vector_use() below?

Perhaps not, we could fall back to regular interrupts.  Could move the
msix_vector_use() into the "then" clause.

>
>> +    }
>> +
>> +    /* 'activate' the vectors */
>> +    for (i = 0; i < s->vectors; i++) {
>> +        msix_vector_use(&s->dev, i);
>> +    }
>> +
>> +    /* if IRQFDs are not supported, we'll have to trigger the interrupts
>> +     * via Qemu char devices */
>> +    if (!ivshmem_has_feature(s, IVSHMEM_IRQFD)) {
>> +        /* for handling interrupts when IRQFD is not available */
>> +        s->eventfd_table = qemu_mallocz(s->vectors * sizeof(EventfdEntry));
>> +    }
>> +}
>> +
>> +static void ivshmem_save(QEMUFile* f, void *opaque)
>> +{
>> +    IVShmemState *proxy = opaque;
>> +
>> +    IVSHMEM_DPRINTF("ivshmem_save\n");
>> +    pci_device_save(&proxy->dev, f);
>> +
>> +    if (ivshmem_has_feature(proxy, IVSHMEM_MSI)) {
>> +        msix_save(&proxy->dev, f);
>> +    } else {
>> +        qemu_put_be32(f, proxy->intrstatus);
>> +        qemu_put_be32(f, proxy->intrmask);
>> +    }
>> +
>> +}
>
> There may be VMState magic to handle conditional structures (or just
> make the structures unconditional), so VMState should be used instead.

MSI-X requires the use of the old-style savevm/loadvm functions.
Should VMState and the load/save be used together?

>
>> +
>> +static int ivshmem_load(QEMUFile* f, void *opaque, int version_id)
>> +{
>> +    IVSHMEM_DPRINTF("ivshmem_load\n");
>> +
>> +    IVShmemState *proxy = opaque;
>> +    int ret, i;
>> +
>
> Missing version check.
>
>> +    ret = pci_device_load(&proxy->dev, f);
>> +    if (ret) {
>> +        return ret;
>> +    }
>> +
>> +    if (ivshmem_has_feature(proxy, IVSHMEM_MSI)) {
>> +        msix_load(&proxy->dev, f);
>> +        for (i = 0; i < proxy->vectors; i++) {
>> +            msix_vector_use(&proxy->dev, i);
>> +        }
>> +    } else {
>> +        proxy->intrstatus = qemu_get_be32(f);
>> +        proxy->intrmask = qemu_get_be32(f);
>> +    }
>> +
>> +    return 0;
>> +}
>> +
>> +static int pci_ivshmem_init(PCIDevice *dev)
>> +{
>> +    IVShmemState *s = DO_UPCAST(IVShmemState, dev, dev);
>> +    uint8_t *pci_conf;
>> +
>> +    if (s->sizearg == NULL)
>> +        s->ivshmem_size = 4 << 20; /* 4 MB default */
>> +    else {
>> +        s->ivshmem_size = ivshmem_get_size(s);
>> +    }
>> +
>> +    register_savevm("ivshmem", 0, 0, ivshmem_save, ivshmem_load, dev);
>> +
>> +    /* IRQFD requires MSI */
>> +    if (ivshmem_has_feature(s, IVSHMEM_IRQFD) &&
>> +        !ivshmem_has_feature(s, IVSHMEM_MSI)) {
>> +        fprintf(stderr, "ivshmem: ioeventfd/irqfd requires MSI\n");
>> +        exit(1);
>> +    }
>> +
>> +    /* check that role is reasonable */
>> +    if (s->role && !((strncmp(s->role, "peer", 5) == 0) ||
>> +                        (strncmp(s->role, "master", 7) == 0))) {
>> +        fprintf(stderr, "ivshmem: 'role' must be 'peer' or 'master'\n");
>> +        exit(1);
>> +    }
>
> I'd add a scalar flag in IVShmemState for role so that further strcmps
> are avoided.
>
>> +
>> +    pci_conf = s->dev.config;
>> +    pci_conf[0x00] = 0xf4; /* Qumranet vendor ID 0x5002 */
>> +    pci_conf[0x01] = 0x1a;
>> +    pci_conf[0x02] = 0x10;
>> +    pci_conf[0x03] = 0x11;
>
> Please add the DID to hw/pci_ids.h and use pci_config_set_xyz() here.
>
>> +    pci_conf[0x04] = PCI_COMMAND_IO | PCI_COMMAND_MEMORY;
>> +    pci_conf[0x0a] = 0x00; /* RAM controller */
>> +    pci_conf[0x0b] = 0x05;
>> +    pci_conf[0x0e] = 0x00; /* header_type */
>> +
>> +    pci_conf[PCI_INTERRUPT_PIN] = 1;
>> +
>> +    s->shm_pci_addr = 0;
>> +    s->ivshmem_offset = 0;
>> +    s->shm_fd = 0;
>> +
>> +    s->ivshmem_mmio_io_addr = cpu_register_io_memory(ivshmem_mmio_read,
>> +                                    ivshmem_mmio_write, s);
>> +    /* region for registers*/
>> +    pci_register_bar(&s->dev, 0, 0x400,
>> +                           PCI_BASE_ADDRESS_SPACE_MEMORY, ivshmem_mmio_map);
>> +
>> +    if ((s->server_chr != NULL) &&
>> +                        (strncmp(s->server_chr->filename, "unix:", 5) == 
>> 0)) {
>> +        /* if we get a UNIX socket as the parameter we will talk
>> +         * to the ivshmem server to receive the memory region */
>> +
>> +        IVSHMEM_DPRINTF("using shared memory server (socket = %s)\n",
>> +                                                    
>> s->server_chr->filename);
>> +
>> +        if (ivshmem_has_feature(s, IVSHMEM_MSI)) {
>> +            ivshmem_setup_msi(s);
>> +        }
>> +
>> +        /* we allocate enough space for 16 guests and grow as needed */
>> +        s->nb_peers = 16;
>> +        s->vm_id = -1;
>> +
>> +        /* allocate/initialize space for interrupt handling */
>> +        s->peers = qemu_mallocz(s->nb_peers * sizeof(Peer));
>> +
>> +        pci_register_bar(&s->dev, 2, s->ivshmem_size,
>> +                                    PCI_BASE_ADDRESS_SPACE_MEMORY, 
>> ivshmem_map);
>> +
>> +        s->eventfd_chr = (CharDriverState **) qemu_mallocz(s->vectors *
>> +                                                sizeof(CharDriverState *));
>
> Useless cast in C.
>
>> +
>> +        qemu_chr_add_handlers(s->server_chr, ivshmem_can_receive, 
>> ivshmem_read,
>> +                     ivshmem_event, s);
>> +    } else {
>> +        /* just map the file immediately, we're not using a server */
>> +        int fd;
>> +
>> +        if (s->shmobj == NULL) {
>> +            fprintf(stderr, "Must specify 'chardev' or 'shm' to ivshmem\n");
>> +        }
>
> I'd rather have separate 'chardev' and 'shm_file' parameters. Then
> 'info qtree' could return more useful information about the chardev.

can you elaborate?  the command-line currently must be one of the following

server case:
-ivshmem chardev=x,...
-chardev id=x,...

or

non-server case:
-ivshmem shm=<name>,...

>
>> +
>> +        IVSHMEM_DPRINTF("using shm_open (shm object = %s)\n", s->shmobj);
>> +
>> +        /* try opening with O_EXCL and if it succeeds zero the memory
>> +         * by truncating to 0 */
>> +        if ((fd = shm_open(s->shmobj, O_CREAT|O_RDWR|O_EXCL,
>> +                        S_IRWXU|S_IRWXG|S_IRWXO)) > 0) {
>> +           /* truncate file to length PCI device's memory */
>> +            if (ftruncate(fd, s->ivshmem_size) != 0) {
>> +                fprintf(stderr, "kvm_ivshmem: could not truncate shared 
>> file\n");
>
> Why 'kvm_ivshmem'?

old name, will remove.

>
>> +            }
>> +
>> +        } else if ((fd = shm_open(s->shmobj, O_CREAT|O_RDWR,
>> +                        S_IRWXU|S_IRWXG|S_IRWXO)) < 0) {
>> +            fprintf(stderr, "kvm_ivshmem: could not open shared file\n");
>> +            exit(-1);
>> +
>> +        }
>> +
>> +        if (check_shm_size(s, fd) == -1) {
>> +            exit(-1);
>> +        }
>> +
>> +        create_shared_memory_BAR(s, fd);
>> +
>> +    }
>> +
>> +    return 0;
>> +}
>> +
>> +static int pci_ivshmem_uninit(PCIDevice *dev)
>> +{
>> +    IVShmemState *s = DO_UPCAST(IVShmemState, dev, dev);
>> +
>> +    cpu_unregister_io_memory(s->ivshmem_mmio_io_addr);
>> +
>> +    return 0;
>> +}
>> +
>> +static PCIDeviceInfo ivshmem_info = {
>> +    .qdev.name  = "ivshmem",
>> +    .qdev.size  = sizeof(IVShmemState),
>> +    .qdev.reset = ivshmem_reset,
>> +    .init       = pci_ivshmem_init,
>> +    .exit       = pci_ivshmem_uninit,
>> +    .qdev.props = (Property[]) {
>> +        DEFINE_PROP_CHR("chardev", IVShmemState, server_chr),
>> +        DEFINE_PROP_STRING("size", IVShmemState, sizearg),
>
> This should be scalar type, not string.

but it needs to handle the 'm' and 'g' suffixes for memory sizes.

>
>> +        DEFINE_PROP_UINT32("vectors", IVShmemState, vectors, 1),
>> +        DEFINE_PROP_BIT("irqfd", IVShmemState, features, IVSHMEM_IRQFD, 
>> false),
>> +        DEFINE_PROP_BIT("msi", IVShmemState, features, IVSHMEM_MSI, true),
>> +        DEFINE_PROP_STRING("shm", IVShmemState, shmobj),
>> +        DEFINE_PROP_STRING("role", IVShmemState, role),
>> +        DEFINE_PROP_END_OF_LIST(),
>> +    }
>> +};
>> +
>> +static void ivshmem_register_devices(void)
>> +{
>> +    pci_qdev_register(&ivshmem_info);
>> +}
>> +
>> +device_init(ivshmem_register_devices)
>> diff --git a/qemu-char.c b/qemu-char.c
>> index ac65a1c..b2e50d0 100644
>> --- a/qemu-char.c
>> +++ b/qemu-char.c
>> @@ -2093,6 +2093,12 @@ static void tcp_chr_read(void *opaque)
>>     }
>>  }
>>
>> +CharDriverState *qemu_chr_open_eventfd(int eventfd){
>> +
>> +    return qemu_chr_open_fd(eventfd, eventfd);
>> +
>> +}
>> +
>>  static void tcp_chr_connect(void *opaque)
>>  {
>>     CharDriverState *chr = opaque;
>> diff --git a/qemu-char.h b/qemu-char.h
>> index e3a0783..6ea01ba 100644
>> --- a/qemu-char.h
>> +++ b/qemu-char.h
>> @@ -94,6 +94,9 @@ void qemu_chr_info_print(Monitor *mon, const QObject 
>> *ret_data);
>>  void qemu_chr_info(Monitor *mon, QObject **ret_data);
>>  CharDriverState *qemu_chr_find(const char *name);
>>
>> +/* add an eventfd to the qemu devices that are polled */
>> +CharDriverState *qemu_chr_open_eventfd(int eventfd);
>
> Maybe this should be removed and just open coded with qemu_chr_open_fd.

I'm indifferent, it just looked a little funny passing the parameter twice.

>
>> +
>>  extern int term_escape_char;
>>
>>  /* async I/O support */
>> diff --git a/qemu-doc.texi b/qemu-doc.texi
>> index 6647b7b..24f8748 100644
>> --- a/qemu-doc.texi
>> +++ b/qemu-doc.texi
>> @@ -706,6 +706,49 @@ Using the @option{-net socket} option, it is possible 
>> to make VLANs
>>  that span several QEMU instances. See @ref{sec_invocation} to have a
>>  basic example.
>>
>> address@hidden Other Devices
>> +
>> address@hidden Inter-VM Shared Memory device
>> +
>> +With KVM enabled on a Linux host, a shared memory device is available.  
>> Guests
>> +map a POSIX shared memory region into the guest as a PCI device that enables
>> +zero-copy communication to the application level of the guests.  The basic
>> +syntax is:
>> +
>> address@hidden
>> +qemu -device ivshmem,size=<size in format accepted by -m>[,shm=<shm name>]
>> address@hidden example
>> +
>> +If desired, interrupts can be sent between guest VMs accessing the same 
>> shared
>> +memory region.  Interrupt support requires using a shared memory server and
>> +using a chardev socket to connect to it.  The code for the shared memory 
>> server
>> +is qemu.git/contrib/ivshmem-server.  An example syntax when using the shared
>> +memory server is:
>> +
>> address@hidden
>> +qemu -device ivshmem,size=<size in format accepted by -m>[,chardev=<id>]
>> +                        [,msi=on][,irqfd=on][,vectors=n][,role=peer|master]
>> +qemu -chardev socket,path=<path>,id=<id>
>> address@hidden example
>> +
>> +When using the server, the guest will be assigned a VM ID (>=0) that allows 
>> guests
>> +using the same server to communicate via interrupts.  Guests can read their
>> +VM ID from a device register (see example code).  Since receiving the shared
>> +memory region from the server is asynchronous, there is a (small) chance the
>> +guest may boot before the shared memory is attached.  To allow an 
>> application
>> +to ensure shared memory is attached, the VM ID register will return -1 (an
>> +invalid VM ID) until the memory is attached.  Once the shared memory is
>> +attached, the VM ID will return the guest's valid VM ID.  With these 
>> semantics,
>> +the guest application can check to ensure the shared memory is attached to 
>> the
>> +guest before proceeding.
>> +
>> +The @option{role} argument can be set to either master or peer and will 
>> affect
>> +how the shared memory is migrated.  With @option{role=master}, the guest 
>> will
>> +copy the shared memory on migration to the destination host.  With
>> address@hidden, the shared memory will not be copied on migration.  Only
>> +one guest should be specified as
>> +the master.
>> +
>> address@hidden direct_linux_boot
>> address@hidden Direct Linux Boot
>>
>> --
>> 1.6.3.2.198.g6096d
>>
>>
>>
>
>