Re: [Qemu-devel] [PATCH v2 11/14] spapr_pci: enable basic hotplug operations

[Alexey Kardashevskiy]

On 12/06/2013 09:33 AM, Michael Roth wrote:
> From: Mike Day <address@hidden>
> 
> This enables hotplug for PHB bridges. Upon hotplug we generate the
> OF-nodes required by PAPR specification and IEEE 1275-1994
> "PCI Bus Binding to Open Firmware" for the device.
> 
> We associate the corresponding FDT for these nodes with the DrcEntry
> corresponding to the slot, which will be fetched via
> ibm,configure-connector RTAS calls by the guest as described by PAPR
> specification. The FDT is cleaned up in the case of unplug.
> 
> Amongst the required OF-node properties for each device are the "reg"
> and "assigned-addresses" properties which describe the BAR-assignments
> for IO/MEM/ROM regions. To handle these assignments we scan the address
> space associated with each region for a contiguous range of appropriate
> size based on PCI specification and encode these in accordance with
> Open Firmware PCI Bus Binding spec.
> 
> These assignments will be used by the guest when the rpaphp hotplug
> module is used, but may be re-assigned by guests for cases where we
> rely on bus rescan.
> 
> Signed-off-by: Mike Day <address@hidden>
> Signed-off-by: Michael Roth <address@hidden>
> ---
>  hw/ppc/spapr_pci.c     |  375 
> ++++++++++++++++++++++++++++++++++++++++++++++--
>  include/hw/ppc/spapr.h |    1 +
>  2 files changed, 368 insertions(+), 8 deletions(-)
> 
> diff --git a/hw/ppc/spapr_pci.c b/hw/ppc/spapr_pci.c
> index 6e7ee31..9b4f829 100644
> --- a/hw/ppc/spapr_pci.c
> +++ b/hw/ppc/spapr_pci.c
> @@ -56,6 +56,17 @@
>  #define RTAS_TYPE_MSI           1
>  #define RTAS_TYPE_MSIX          2
>  
> +#define FDT_MAX_SIZE            0x10000
> +#define _FDT(exp) \
> +    do { \
> +        int ret = (exp);                                           \
> +        if (ret < 0) {                                             \
> +            return ret;                                            \
> +        }                                                          \
> +    } while (0)
> +
> +static void spapr_drc_state_reset(DrcEntry *drc_entry);
> +
>  static sPAPRPHBState *find_phb(sPAPREnvironment *spapr, uint64_t buid)
>  {
>      sPAPRPHBState *sphb;
> @@ -448,6 +459,22 @@ static void rtas_set_indicator(PowerPCCPU *cpu, 
> sPAPREnvironment *spapr,
>          /* encode the new value into the correct bit field */
>          shift = INDICATOR_ISOLATION_SHIFT;
>          mask = INDICATOR_ISOLATION_MASK;
> +        if (drc_entry) {
> +            /* transition from unisolated to isolated for a hotplug slot
> +             * entails completion of guest-side device unplug/cleanup, so
> +             * we can now safely remove the device if qemu is waiting for
> +             * it to be released
> +             */
> +            if (DECODE_DRC_STATE(*pind, mask, shift) != indicator_state) {
> +                if (indicator_state == 0 && drc_entry->awaiting_release) {
> +                    /* device_del has been called and host is waiting
> +                     * for guest to release/isolate device, go ahead
> +                     * and remove it now
> +                     */
> +                    spapr_drc_state_reset(drc_entry);
> +                }
> +            }
> +        }
>          break;
>      case 9002: /* DR */
>          shift = INDICATOR_DR_SHIFT;
> @@ -776,6 +803,345 @@ static AddressSpace *spapr_pci_dma_iommu(PCIBus *bus, 
> void *opaque, int devfn)
>      return &phb->iommu_as;
>  }
>  
> +/* for 'reg'/'assigned-addresses' OF properties */
> +#define RESOURCE_CELLS_SIZE 2
> +#define RESOURCE_CELLS_ADDRESS 3
> +#define RESOURCE_CELLS_TOTAL \
> +    (RESOURCE_CELLS_SIZE + RESOURCE_CELLS_ADDRESS)
> +
> +static void fill_resource_props(PCIDevice *d, int bus_num,
> +                                uint32_t *reg, int *reg_size,
> +                                uint32_t *assigned, int *assigned_size)
> +{
> +    uint32_t *reg_row, *assigned_row;
> +    uint32_t dev_id = ((bus_num << 8) |
> +                        (PCI_SLOT(d->devfn) << 3) | PCI_FUNC(d->devfn));
> +    int i, idx = 0;
> +
> +    reg[0] = cpu_to_be32(dev_id << 8);
> +
> +    for (i = 0; i < PCI_NUM_REGIONS; i++) {
> +        if (!d->io_regions[i].size) {
> +            continue;
> +        }
> +        reg_row = &reg[(idx + 1) * RESOURCE_CELLS_TOTAL];
> +        assigned_row = &assigned[idx * RESOURCE_CELLS_TOTAL];
> +        reg_row[0] = cpu_to_be32((dev_id << 8) | (pci_bar(d, i) & 0xff));
> +        if (d->io_regions[i].type & PCI_BASE_ADDRESS_SPACE_IO) {
> +            reg_row[0] |= cpu_to_be32(0x01000000);
> +        } else {
> +            reg_row[0] |= cpu_to_be32(0x02000000);
> +        }
> +        assigned_row[0] = cpu_to_be32(reg_row[0] | 0x80000000);


0x80000000 == relocatable? 0x01000000/0x02000000 - space codes? There are
macros (b_n, b_ss) in this file, can you please use them?


> +        assigned_row[3] = reg_row[3] = cpu_to_be32(d->io_regions[i].size >> 
> 32);
> +        assigned_row[4] = reg_row[4] = cpu_to_be32(d->io_regions[i].size);
> +        assigned_row[1] = cpu_to_be32(d->io_regions[i].addr >> 32);
> +        assigned_row[2] = cpu_to_be32(d->io_regions[i].addr);
> +        idx++;
> +    }
> +
> +    *reg_size = (idx + 1) * RESOURCE_CELLS_TOTAL * sizeof(uint32_t);
> +    *assigned_size = idx * RESOURCE_CELLS_TOTAL * sizeof(uint32_t);
> +}
> +
> +static hwaddr spapr_find_bar_addr(sPAPRPHBState *phb, PCIIORegion *r)


This does not use @pbh at all and therefore can go to hw/pci/pci.c may be
(which can be tricky though)?


> +{
> +    MemoryRegionSection mrs = { 0 };
> +    hwaddr search_addr;
> +    hwaddr size = r->size;
> +    hwaddr addr_mask = ~(size - 1);
> +    hwaddr increment = size;
> +    hwaddr limit;
> +
> +    if (r->type == PCI_BASE_ADDRESS_SPACE_MEMORY) {
> +        /* beginning portion of mmio address space for bus does not get
> +         * mapped into system memory, so calculate addr starting at the
> +         * corresponding offset into mmio as.
> +         */
> +        search_addr = (SPAPR_PCI_MEM_WIN_BUS_OFFSET + increment) & addr_mask;
> +    } else {
> +        search_addr = increment;
> +    }
> +    limit = memory_region_size(r->address_space);
> +
> +    do {
> +        mrs = memory_region_find_subregion(r->address_space, search_addr, 
> size);
> +        if (mrs.mr) {
> +            hwaddr mr_last_addr;
> +            mr_last_addr = mrs.mr->addr + memory_region_size(mrs.mr) - 1;
> +            search_addr = (mr_last_addr + 1) & addr_mask;
> +            if (search_addr <= mr_last_addr) {
> +                search_addr += increment;
> +            }
> +            /* this memory region overlaps, unref and continue searching */
> +            memory_region_unref(mrs.mr);
> +        }
> +    } while (int128_nz(mrs.size) && search_addr + size <= limit);
> +
> +    if (search_addr + size >= limit) {
> +        return PCI_BAR_UNMAPPED;
> +    }
> +
> +    return search_addr;
> +}
> +
> +static int spapr_map_bars(sPAPRPHBState *phb, PCIDevice *dev)

This does not use @phb, well, it uses to call spapr_find_bar_addr() but
that function does not use it either.

Yet another candidate to get moved to hw/pci/pci.c? If you do so, you'll
get even more reviews :)


> +{
> +    PCIIORegion *r;
> +    int i, ret = -1;
> +
> +    for (i = 0; i < PCI_NUM_REGIONS; i++) {
> +        uint32_t bar_address = pci_bar(dev, i);
> +        uint32_t bar_value;
> +        uint16_t cmd_value = pci_default_read_config(dev, PCI_COMMAND, 2);
> +        hwaddr addr;
> +
> +        r = &dev->io_regions[i];
> +
> +        /* this region isn't registered */
> +        if (!r->size) {
> +            continue;
> +        }
> +
> +        /* find a hw addr we can map */
> +        addr = spapr_find_bar_addr(phb, r);
> +        if (addr == PCI_BAR_UNMAPPED) {
> +            /* we can't find a free range within address space for this BAR 
> */
> +            fprintf(stderr,
> +                    "Unable to map BAR %d, no free range available\n", i);
> +            return -1;
> +        }
> +        /* we can probably map this region into memory if there is not
> +         * a race condition with some other allocator. write the address
> +         * to the device BAR which will force a call to pci_update_mappings
> +         */
> +        if (r->type & PCI_BASE_ADDRESS_SPACE_IO) {
> +            pci_default_write_config(dev, PCI_COMMAND,
> +                                     cmd_value | PCI_COMMAND_IO, 2);
> +        } else {
> +            pci_default_write_config(dev, PCI_COMMAND,
> +                                     cmd_value | PCI_COMMAND_MEMORY, 2);
> +        }
> +
> +        bar_value = addr;
> +
> +        if (i == PCI_ROM_SLOT) {
> +            bar_value |= PCI_ROM_ADDRESS_ENABLE;
> +        }
> +        /* write the new bar value */
> +        pci_default_write_config(dev, bar_address, bar_value, 4);
> +
> +        /* if this is a 64-bit BAR, we need to also write the
> +         * upper 32 bit value.
> +         */
> +        if (r->type & PCI_BASE_ADDRESS_MEM_TYPE_64) {
> +            bar_value = (addr >> 32) & 0xffffffffUL;
> +            pci_default_write_config(dev, bar_address + 4, bar_value, 4);
> +        }
> +        ret = 0;
> +    }
> +    return ret;
> +}
> +
> +static int spapr_populate_pci_child_dt(PCIDevice *dev, void *fdt, int offset,
> +                                       int phb_index)
> +{
> +    int slot = PCI_SLOT(dev->devfn);
> +    char slotname[16];
> +    bool is_bridge = 1;
> +    DrcEntry *drc_entry, *drc_entry_slot;
> +    uint32_t reg[RESOURCE_CELLS_TOTAL * 8] = { 0 };
> +    uint32_t assigned[RESOURCE_CELLS_TOTAL * 8] = { 0 };
> +    int reg_size, assigned_size;
> +
> +    drc_entry = spapr_phb_to_drc_entry(phb_index + SPAPR_PCI_BASE_BUID);
> +    g_assert(drc_entry);
> +    drc_entry_slot = &drc_entry->child_entries[slot];
> +
> +    if (pci_default_read_config(dev, PCI_HEADER_TYPE, 1) ==


s/1/PCI_HEADER_TYPE_BRIDGE/


> +        PCI_HEADER_TYPE_NORMAL) {
> +        is_bridge = 0;
> +    }
> +
> +    _FDT(fdt_setprop_cell(fdt, offset, "vendor-id",
> +                          pci_default_read_config(dev, PCI_VENDOR_ID, 2)));
> +    _FDT(fdt_setprop_cell(fdt, offset, "device-id",
> +                          pci_default_read_config(dev, PCI_DEVICE_ID, 2)));
> +    _FDT(fdt_setprop_cell(fdt, offset, "revision-id",
> +                          pci_default_read_config(dev, PCI_REVISION_ID, 1)));
> +    _FDT(fdt_setprop_cell(fdt, offset, "class-code",
> +                          pci_default_read_config(dev, PCI_CLASS_DEVICE, 2) 
> << 8));
> +
> +    _FDT(fdt_setprop_cell(fdt, offset, "interrupts",
> +                          pci_default_read_config(dev, PCI_INTERRUPT_PIN, 
> 1)));
> +
> +    /* if this device is NOT a bridge */
> +    if (!is_bridge) {


s/!is_bridge/pci_default_read_config(dev, PCI_HEADER_TYPE, 1) ==
PCI_HEADER_TYPE_NORMAL/

and get rid of is_bridge?



> +        _FDT(fdt_setprop_cell(fdt, offset, "min-grant",
> +            pci_default_read_config(dev, PCI_MIN_GNT, 1)));
> +        _FDT(fdt_setprop_cell(fdt, offset, "max-latency",
> +            pci_default_read_config(dev, PCI_MAX_LAT, 1)));
> +        _FDT(fdt_setprop_cell(fdt, offset, "subsystem-id",
> +            pci_default_read_config(dev, PCI_SUBSYSTEM_ID, 2)));
> +        _FDT(fdt_setprop_cell(fdt, offset, "subsystem-vendor-id",
> +            pci_default_read_config(dev, PCI_SUBSYSTEM_VENDOR_ID, 2)));
> +    }
> +
> +    _FDT(fdt_setprop_cell(fdt, offset, "cache-line-size",
> +        pci_default_read_config(dev, PCI_CACHE_LINE_SIZE, 1)));
> +
> +    /* the following fdt cells are masked off the pci status register */
> +    int pci_status = pci_default_read_config(dev, PCI_STATUS, 2);
> +    _FDT(fdt_setprop_cell(fdt, offset, "devsel-speed",
> +                          PCI_STATUS_DEVSEL_MASK & pci_status));
> +    _FDT(fdt_setprop_cell(fdt, offset, "fast-back-to-back",
> +                          PCI_STATUS_FAST_BACK & pci_status));
> +    _FDT(fdt_setprop_cell(fdt, offset, "66mhz-capable",
> +                          PCI_STATUS_66MHZ & pci_status));
> +    _FDT(fdt_setprop_cell(fdt, offset, "udf-supported",
> +                          PCI_STATUS_UDF & pci_status));
> +
> +    _FDT(fdt_setprop_string(fdt, offset, "name", "pci"));
> +    sprintf(slotname, "Slot %d", slot + phb_index * 32);
> +    _FDT(fdt_setprop(fdt, offset, "ibm,loc-code", slotname, 
> strlen(slotname)));
> +    _FDT(fdt_setprop_cell(fdt, offset, "ibm,my-drc-index",
> +                          drc_entry_slot->drc_index));
> +
> +    _FDT(fdt_setprop_cell(fdt, offset, "#address-cells",
> +                          RESOURCE_CELLS_ADDRESS));
> +    _FDT(fdt_setprop_cell(fdt, offset, "#size-cells",
> +                          RESOURCE_CELLS_SIZE));
> +    _FDT(fdt_setprop_cell(fdt, offset, "ibm,req#msi-x",
> +                          RESOURCE_CELLS_SIZE));
> +    fill_resource_props(dev, phb_index, reg, &reg_size,
> +                        assigned, &assigned_size);
> +    _FDT(fdt_setprop(fdt, offset, "reg", reg, reg_size));
> +    _FDT(fdt_setprop(fdt, offset, "assigned-addresses",
> +                     assigned, assigned_size));
> +
> +    return 0;
> +}
> +
> +static int spapr_device_hotplug_add(DeviceState *qdev, PCIDevice *dev)
> +{
> +    sPAPRPHBState *phb = SPAPR_PCI_HOST_BRIDGE(qdev);
> +    DrcEntry *drc_entry, *drc_entry_slot;
> +    ConfigureConnectorState *ccs;
> +    int slot = PCI_SLOT(dev->devfn);
> +    int offset, ret;
> +    void *fdt_orig, *fdt;
> +    char nodename[512];
> +    uint32_t encoded = ENCODE_DRC_STATE(INDICATOR_ENTITY_SENSE_PRESENT,
> +                                        INDICATOR_ENTITY_SENSE_MASK,
> +                                        INDICATOR_ENTITY_SENSE_SHIFT);
> +
> +    drc_entry = spapr_phb_to_drc_entry(phb->buid);
> +    g_assert(drc_entry);
> +    drc_entry_slot = &drc_entry->child_entries[slot];
> +
> +    drc_entry->state &= ~(uint32_t)INDICATOR_ENTITY_SENSE_MASK;
> +    drc_entry->state |= encoded; /* DR entity present */
> +    drc_entry_slot->state &= ~(uint32_t)INDICATOR_ENTITY_SENSE_MASK;
> +    drc_entry_slot->state |= encoded; /* and the slot */


"and the slot" what?
s/uint32_t encoded/const uint32_t present/ and remove the comments?


> +    /* need to allocate memory region for device BARs */
> +    spapr_map_bars(phb, dev);
> +
> +    /* add OF node for pci device and required OF DT properties */
> +    fdt_orig = g_malloc0(FDT_MAX_SIZE);
> +    offset = fdt_create(fdt_orig, FDT_MAX_SIZE);
> +    fdt_begin_node(fdt_orig, "");
> +    fdt_end_node(fdt_orig);
> +    fdt_finish(fdt_orig);
> +
> +    fdt = g_malloc0(FDT_MAX_SIZE);
> +    fdt_open_into(fdt_orig, fdt, FDT_MAX_SIZE);
> +    sprintf(nodename, "address@hidden", slot);
> +    offset = fdt_add_subnode(fdt, 0, nodename);
> +    ret = spapr_populate_pci_child_dt(dev, fdt, offset, phb->index);
> +    g_assert(!ret);
> +    g_free(fdt_orig);
> +
> +    /* hold on to node, configure_connector will pass it to the guest later 
> */
> +    ccs = &drc_entry_slot->cc_state;
> +    ccs->fdt = fdt;
> +    ccs->offset_start = offset;
> +    ccs->state = CC_STATE_PENDING;
> +    ccs->dev = dev;
> +
> +    return 0;
> +}
> +
> +/* check whether guest has released/isolated device */
> +static bool spapr_drc_state_is_releasable(DrcEntry *drc_entry)
> +{
> +    return !DECODE_DRC_STATE(drc_entry->state,
> +                             INDICATOR_ISOLATION_MASK,
> +                             INDICATOR_ISOLATION_SHIFT);
> +}

It looks like this is the only separated function which calls
DECODE_DRC_STATE, and it is used just once, and  others call
DECODE_DRC_STATE()/ENCODE_DRC_STATE() directly. I'd remove this function
and call DECODE_DRC_STATE() directly, below in the code.


> +
> +/* finalize device unplug/deletion */
> +static void spapr_drc_state_reset(DrcEntry *drc_entry)
> +{
> +    ConfigureConnectorState *ccs = &drc_entry->cc_state;
> +    uint32_t sense_empty = ENCODE_DRC_STATE(INDICATOR_ENTITY_SENSE_EMPTY,
> +                                            INDICATOR_ENTITY_SENSE_MASK,
> +                                            INDICATOR_ENTITY_SENSE_SHIFT);
> +
> +    g_free(ccs->fdt);
> +    ccs->fdt = NULL;
> +    object_unparent(OBJECT(ccs->dev));
> +    ccs->dev = NULL;
> +    ccs->state = CC_STATE_IDLE;
> +    drc_entry->state &= ~INDICATOR_ENTITY_SENSE_MASK;
> +    drc_entry->state |= sense_empty;
> +    drc_entry->awaiting_release = false;
> +}
> +
> +static void spapr_device_hotplug_remove(DeviceState *qdev, PCIDevice *dev)
> +{
> +    sPAPRPHBState *phb = SPAPR_PCI_HOST_BRIDGE(qdev);
> +    DrcEntry *drc_entry, *drc_entry_slot;
> +    ConfigureConnectorState *ccs;
> +    int slot = PCI_SLOT(dev->devfn);
> +
> +    drc_entry = spapr_phb_to_drc_entry(phb->buid);
> +    g_assert(drc_entry);
> +    drc_entry_slot = &drc_entry->child_entries[slot];
> +    ccs = &drc_entry_slot->cc_state;
> +    /* shouldn't be removing devices we haven't created an fdt for */
> +    g_assert(ccs->state != CC_STATE_IDLE);


Instead of g_assert(), would not it be better to return -1 here and then
return this return code from spapr_device_hotplug() and let common PCI code
handle this?

Or we are absolutely sure that spapr_device_hotplug() cannot possibly fail
so we are ready to kill the guest if it does? I do not know, just asking :)


> +    /* if the device has already been released/isolated by guest, go ahead
> +     * and remove it now. Otherwise, flag it as pending guest release so it
> +     * can be removed later
> +     */
> +    if (spapr_drc_state_is_releasable(drc_entry_slot)) {
> +        spapr_drc_state_reset(drc_entry_slot);
> +    } else {
> +        if (drc_entry_slot->awaiting_release) {
> +            fprintf(stderr, "waiting for guest to release the device");
> +        } else {
> +            drc_entry_slot->awaiting_release = true;
> +        }
> +    }
> +}
> +
> +static int spapr_device_hotplug(DeviceState *qdev, PCIDevice *dev,
> +                                PCIHotplugState state)
> +{

sPAPRPHBState *phb = SPAPR_PCI_HOST_BRIDGE(qdev);

> +    if (state == PCI_COLDPLUG_ENABLED) {
> +        return 0;
> +    }
> +
> +    if (state == PCI_HOTPLUG_ENABLED) {
> +        spapr_device_hotplug_add(qdev, dev);
> +    } else {
> +        spapr_device_hotplug_remove(qdev, dev);
> +    }

and here s/qdev/phb/? spapr_device_hotplug_(add|remove),
spapr_pci_hotplug_(add|remove)_event (from further patch(es)) do not use
qdev as a DeviceState anyway, they cast it to sPAPRPHBState and use that.



> +
> +    return 0;
> +}
> +
>  static int spapr_phb_init(SysBusDevice *s)
>  {
>      DeviceState *dev = DEVICE(s);
> @@ -889,6 +1255,7 @@ static int spapr_phb_init(SysBusDevice *s)
>                             &sphb->memspace, &sphb->iospace,
>                             PCI_DEVFN(0, 0), PCI_NUM_PINS, TYPE_PCI_BUS);
>      phb->bus = bus;
> +    pci_bus_hotplug(phb->bus, spapr_device_hotplug, DEVICE(sphb));
>  
>      sphb->dma_window_start = 0;
>      sphb->dma_window_size = 0x40000000;
> @@ -1181,14 +1548,6 @@ int spapr_populate_pci_dt(sPAPRPHBState *phb,
>          return bus_off;
>      }
>  
> -#define _FDT(exp) \
> -    do { \
> -        int ret = (exp);                                           \
> -        if (ret < 0) {                                             \
> -            return ret;                                            \
> -        }                                                          \
> -    } while (0)
> -
>      /* Write PHB properties */
>      _FDT(fdt_setprop_string(fdt, bus_off, "device_type", "pci"));
>      _FDT(fdt_setprop_string(fdt, bus_off, "compatible", "IBM,Logical_PHB"));
> diff --git a/include/hw/ppc/spapr.h b/include/hw/ppc/spapr.h
> index 7c8a521..1c9b725 100644
> --- a/include/hw/ppc/spapr.h
> +++ b/include/hw/ppc/spapr.h
> @@ -328,6 +328,7 @@ struct DrcEntry {
>      void *fdt;
>      int fdt_offset;
>      uint32_t state;
> +    bool awaiting_release;
>      ConfigureConnectorState cc_state;
>      DrcEntry *child_entries;
>  };
> 


-- 
Alexey