Skip to contentSkip to conversation

boot order specification

PATCH Superseded Gleb Natapov 4 participants 55 messages Nov 14, 2010 – Nov 18, 2010 KVM

v4 of 8 Superseded Latest is v8 →

lore
Share & tools
lore
#0 lore raw

v4 cover boot order specification

This is current state of the patch series for people to comment on. I am using open firmware naming scheme to specify device path names. In this submission I addressed all comment from previous one and added option rom support and rebased to qemu upstream.

Kevin can you double check that the names are usable by Seabios? Reading PC boot specification it looks like Seabios will not be able to take full advantage of this though. Only one BCV can be bootable, so only disk with lowest boot index will be bootable by Seabios. Is this correct?

Names look like this on pci machine:
/pci@i0cf8/ide@1,1/drive@1/disk@0
/pci@i0cf8/isa@1/fdc@03f1/floppy@1
/pci@i0cf8/isa@1/fdc@03f1/floppy@0
/pci@i0cf8/ide@1,1/drive@1/disk@1
/pci@i0cf8/ide@1,1/drive@0/disk@0
/pci@i0cf8/scsi@3/disk@0
/pci@i0cf8/ethernet@4/ethernet-phy@0
/pci@i0cf8/ethernet@5/ethernet-phy@0
/pci@i0cf8/ide@1,1/drive@0/disk@1
/pci@i0cf8/isa@1/ide@01e8/drive@0/disk@0
/pci@i0cf8/usb@1,2/network@0/ethernet@0
/pci@i0cf8/usb@1,2/hub@1/network@0/ethernet@0
/rom@genroms/linuxboot.bin

and on isa machine:
/isa/ide@0170/drive@0/disk@0
/isa/fdc@03f1/floppy@1
/isa/fdc@03f1/floppy@0
/isa/ide@0170/drive@0/disk@1

Instead of using get_dev_path() callback I introduces another one get_fw_dev_path. Unfortunately the way get_dev_path() callback is used in migration code makes it hard to reuse it for other purposes. First of all it is not called recursively so caller expects it to provide unique name by itself. Device path though is inherently recursive. Each individual element may not be unique, but the whole path will be. On the other hand to call get_dev_path() recursively in migration code we should implement it for all possible buses first. Other problem is compatibility. If we change get_dev_path() output format now we will not be able to migrate from old qemu to new one without some additional compatibility layer.

Gleb Natapov (15):
Introduce fw_name field to DeviceInfo structure. Introduce new BusInfo callback get_fw_dev_path.
Keep track of ISA ports ISA device is using in qdev. Add get_fw_dev_path callback to ISA bus in qdev. Store IDE bus id in IDEBus structure for easy access. Add get_fw_dev_path callback to IDE bus.
Add get_dev_path callback for system bus.
Add get_fw_dev_path callback for pci bus.
Record which USBDevice USBPort belongs too.
Add get_dev_path callback for usb bus.
Add bootindex parameter to net/block/fd device
Change fw_cfg_add_file() to get full file path as a parameter. Add bootindex for option roms.
Add notifier that will be called when machine is fully created. Pass boot device list to firmware.

 block_int.h       |    4 +-
 hw/cs4231a.c      |    1 +
 hw/e1000.c        |    4 ++
 hw/eepro100.c     |    3 +
 hw/fdc.c          |   12 +++++
 hw/fw_cfg.c       |   30 ++++++++-----
 hw/fw_cfg.h       |    8 ++-
 hw/gus.c          |    4 ++
 hw/ide/cmd646.c   |    4 +-
 hw/ide/internal.h |    3 +-
 hw/ide/isa.c      |    5 ++-
 hw/ide/piix.c     |    4 +-
 hw/ide/qdev.c     |   22 +++++++++-
 hw/ide/via.c      |    4 +-
 hw/isa-bus.c      |   42 +++++++++++++++++++
 hw/isa.h          |    4 ++
 hw/lance.c        |    1 +
 hw/loader.c       |   32 +++++++++++---
 hw/loader.h       |    8 ++--
 hw/m48t59.c       |    1 +
 hw/mc146818rtc.c  |    1 +
 hw/multiboot.c    |    3 +-
 hw/ne2000-isa.c   |    3 +
 hw/ne2000.c       |    5 ++-
 hw/nseries.c      |    2 +-
 hw/parallel.c     |    5 ++
 hw/pc.c           |    7 ++-
 hw/pci.c          |  110 +++++++++++++++++++++++++++++++++++++++-----------
 hw/pci_host.c     |    2 +
 hw/pckbd.c        |    3 +
 hw/pcnet.c        |    6 ++-
 hw/piix_pci.c     |    1 +
 hw/qdev.c         |   32 ++++++++++++++
 hw/qdev.h         |    9 ++++
 hw/rtl8139.c      |    4 ++
 hw/sb16.c         |    4 ++
 hw/serial.c       |    1 +
 hw/sysbus.c       |   30 ++++++++++++++
 hw/sysbus.h       |    4 ++
 hw/usb-bus.c      |   45 ++++++++++++++++++++-
 hw/usb-hub.c      |    3 +-
 hw/usb-musb.c     |    2 +-
 hw/usb-net.c      |    3 +
 hw/usb-ohci.c     |    2 +-
 hw/usb-uhci.c     |    2 +-
 hw/usb.h          |    3 +-
 hw/virtio-blk.c   |    2 +
 hw/virtio-net.c   |    2 +
 hw/virtio-pci.c   |    1 +
 net.h             |    4 +-
 qemu-config.c     |   17 ++++++++
 sysemu.h          |   11 +++++-
 vl.c              |  117 ++++++++++++++++++++++++++++++++++++++++++++++++++++-
 53 files changed, 565 insertions(+), 77 deletions(-)
#1 lore raw

PATCH v4 1/15 Introduce fw_name field to DeviceInfo structure.

Add "fw_name" to DeviceInfo to use in device path building. In contrast to "name" "fw_name" should refer to functionality device provides instead of particular device model like "name" does.

---
 hw/fdc.c        |    1 +
 hw/ide/isa.c    |    1 +
 hw/ide/qdev.c   |    1 +
 hw/isa-bus.c    |    1 +
 hw/lance.c      |    1 +
 hw/piix_pci.c   |    1 +
 hw/qdev.h       |    6 ++++++
 hw/usb-hub.c    |    1 +
 hw/usb-net.c    |    1 +
 hw/virtio-pci.c |    1 +
 10 files changed, 15 insertions(+), 0 deletions(-)
hw/fdc.c +1 -0
index c159dcb..a467c4b 100644
@@ -2040,6 +2040,7 @@ static const VMStateDescription vmstate_isa_fdc ={
 static ISADeviceInfo isa_fdc_info = {
     .init = isabus_fdc_init1,
     .qdev.name  = "isa-fdc",
+    .qdev.fw_name  = "fdc",
     .qdev.size  = sizeof(FDCtrlISABus),
     .qdev.no_user = 1,
     .qdev.vmsd  = &vmstate_isa_fdc,
hw/ide/isa.c +1 -0
index 6b57e0d..9856435 100644
@@ -98,6 +98,7 @@ ISADevice *isa_ide_init(int iobase, int iobase2, int isairq,
 
 static ISADeviceInfo isa_ide_info = {
     .qdev.name  = "isa-ide",
+    .qdev.fw_name  = "ide",
     .qdev.size  = sizeof(ISAIDEState),
     .init       = isa_ide_initfn,
     .qdev.reset = isa_ide_reset,
hw/ide/qdev.c +1 -0
index 0808760..6d27b60 100644
@@ -134,6 +134,7 @@ static int ide_drive_initfn(IDEDevice *dev)
 
 static IDEDeviceInfo ide_drive_info = {
     .qdev.name  = "ide-drive",
+    .qdev.fw_name  = "drive",
     .qdev.size  = sizeof(IDEDrive),
     .init       = ide_drive_initfn,
     .qdev.props = (Property[]) {
hw/isa-bus.c +1 -0
index 4e306de..26036e0 100644
@@ -153,6 +153,7 @@ static int isabus_bridge_init(SysBusDevice *dev)
 static SysBusDeviceInfo isabus_bridge_info = {
     .init = isabus_bridge_init,
     .qdev.name  = "isabus-bridge",
+    .qdev.fw_name  = "isa",
     .qdev.size  = sizeof(SysBusDevice),
     .qdev.no_user = 1,
 };
hw/lance.c +1 -0
index dc12144..1a3bb1a 100644
@@ -141,6 +141,7 @@ static void lance_reset(DeviceState *dev)
 static SysBusDeviceInfo lance_info = {
     .init       = lance_init,
     .qdev.name  = "lance",
+    .qdev.fw_name  = "ethernet",
     .qdev.size  = sizeof(SysBusPCNetState),
     .qdev.reset = lance_reset,
     .qdev.vmsd  = &vmstate_lance,
hw/piix_pci.c +1 -0
index b5589b9..38f9d9e 100644
@@ -365,6 +365,7 @@ static PCIDeviceInfo i440fx_info[] = {
 static SysBusDeviceInfo i440fx_pcihost_info = {
     .init         = i440fx_pcihost_initfn,
     .qdev.name    = "i440FX-pcihost",
+    .qdev.fw_name = "pci",
     .qdev.size    = sizeof(I440FXState),
     .qdev.no_user = 1,
 };
hw/qdev.h +6 -0
index 579328a..9f90efe 100644
@@ -139,6 +139,7 @@ typedef void (*qdev_resetfn)(DeviceState *dev);
 
 struct DeviceInfo {
     const char *name;
+    const char *fw_name;
     const char *alias;
     const char *desc;
     size_t size;
@@ -288,6 +289,11 @@ void qdev_prop_set_defaults(DeviceState *dev, Property *props);
 void qdev_prop_register_global_list(GlobalProperty *props);
 void qdev_prop_set_globals(DeviceState *dev);
 
+static inline const char *qdev_fw_name(DeviceState *dev)
+{
+    return dev->info->fw_name ? : dev->info->alias ? : dev->info->name;
+}
+
 /* This is a nasty hack to allow passing a NULL bus to qdev_create.  */
 extern struct BusInfo system_bus_info;
 
hw/usb-hub.c +1 -0
index 2a1edfc..8e3a96b 100644
@@ -545,6 +545,7 @@ static int usb_hub_initfn(USBDevice *dev)
 static struct USBDeviceInfo hub_info = {
     .product_desc   = "QEMU USB Hub",
     .qdev.name      = "usb-hub",
+    .qdev.fw_name    = "hub",
     .qdev.size      = sizeof(USBHubState),
     .init           = usb_hub_initfn,
     .handle_packet  = usb_hub_handle_packet,
hw/usb-net.c +1 -0
index 70f9263..2287ee1 100644
@@ -1496,6 +1496,7 @@ static USBDevice *usb_net_init(const char *cmdline)
 static struct USBDeviceInfo net_info = {
     .product_desc   = "QEMU USB Network Interface",
     .qdev.name      = "usb-net",
+    .qdev.fw_name    = "network",
     .qdev.size      = sizeof(USBNetState),
     .init           = usb_net_initfn,
     .handle_packet  = usb_generic_handle_packet,
hw/virtio-pci.c +1 -0
index 729917d..be2c92f 100644
@@ -697,6 +697,7 @@ static int virtio_9p_init_pci(PCIDevice *pci_dev)
 static PCIDeviceInfo virtio_info[] = {
     {
         .qdev.name = "virtio-blk-pci",
+        .qdev.alias = "virtio-blk",
         .qdev.size = sizeof(VirtIOPCIProxy),
         .init      = virtio_blk_init_pci,
         .exit      = virtio_blk_exit_pci,
Signature · 1 line
1.7.1
#2 lore raw

PATCH v4 2/15 Introduce new BusInfo callback get_fw_dev_path.

In reply to Gleb Natapov

New get_fw_dev_path callback will be used for build device path usable by firmware in contrast to qdev qemu internal device path.

---
 hw/qdev.h |    2 ++
 1 files changed, 2 insertions(+), 0 deletions(-)
hw/qdev.h +2 -0
index 9f90efe..dc669b3 100644
@@ -49,12 +49,14 @@ struct DeviceState {
 
 typedef void (*bus_dev_printfn)(Monitor *mon, DeviceState *dev, int indent);
 typedef char *(*bus_get_dev_path)(DeviceState *dev);
+typedef char *(*bus_get_fw_dev_path)(DeviceState *dev);
 
 struct BusInfo {
     const char *name;
     size_t size;
     bus_dev_printfn print_dev;
     bus_get_dev_path get_dev_path;
+    bus_get_fw_dev_path get_fw_dev_path;
     Property *props;
 };
 
Signature · 1 line
1.7.1
#3 lore raw

PATCH v4 3/15 Keep track of ISA ports ISA device is using in qdev.

In reply to Gleb Natapov

Store all io ports used by device in ISADevice structure.

---
 hw/cs4231a.c     |    1 +
 hw/fdc.c         |    3 +++
 hw/gus.c         |    4 ++++
 hw/ide/isa.c     |    2 ++
 hw/isa-bus.c     |   25 +++++++++++++++++++++++++
 hw/isa.h         |    4 ++++
 hw/m48t59.c      |    1 +
 hw/mc146818rtc.c |    1 +
 hw/ne2000-isa.c  |    3 +++
 hw/parallel.c    |    5 +++++
 hw/pckbd.c       |    3 +++
 hw/sb16.c        |    4 ++++
 hw/serial.c      |    1 +
 13 files changed, 57 insertions(+), 0 deletions(-)
hw/cs4231a.c +1 -0
index 4d5ce5c..598f032 100644
@@ -645,6 +645,7 @@ static int cs4231a_initfn (ISADevice *dev)
     isa_init_irq (dev, &s->pic, s->irq);
 
     for (i = 0; i < 4; i++) {
+        isa_init_ioport(dev, i);
         register_ioport_write (s->port + i, 1, 1, cs_write, s);
         register_ioport_read (s->port + i, 1, 1, cs_read, s);
     }
hw/fdc.c +3 -0
index a467c4b..5ab754b 100644
@@ -1983,6 +1983,9 @@ static int isabus_fdc_init1(ISADevice *dev)
                           &fdctrl_write_port, fdctrl);
     register_ioport_write(iobase + 0x07, 1, 1,
                           &fdctrl_write_port, fdctrl);
+    isa_init_ioport_range(dev, iobase + 1, 5);
+    isa_init_ioport(dev, iobase + 7);
+
     isa_init_irq(&isa->busdev, &fdctrl->irq, isairq);
     fdctrl->dma_chann = dma_chann;
 
hw/gus.c +4 -0
index e9016d8..ff9e7c7 100644
@@ -264,20 +264,24 @@ static int gus_initfn (ISADevice *dev)
 
     register_ioport_write (s->port, 1, 1, gus_writeb, s);
     register_ioport_write (s->port, 1, 2, gus_writew, s);
+    isa_init_ioport_range(dev, s->port, 2);
 
     register_ioport_read ((s->port + 0x100) & 0xf00, 1, 1, gus_readb, s);
     register_ioport_read ((s->port + 0x100) & 0xf00, 1, 2, gus_readw, s);
+    isa_init_ioport_range(dev, (s->port + 0x100) & 0xf00, 2);
 
     register_ioport_write (s->port + 6, 10, 1, gus_writeb, s);
     register_ioport_write (s->port + 6, 10, 2, gus_writew, s);
     register_ioport_read (s->port + 6, 10, 1, gus_readb, s);
     register_ioport_read (s->port + 6, 10, 2, gus_readw, s);
+    isa_init_ioport_range(dev, s->port + 6, 10);
 
 
     register_ioport_write (s->port + 0x100, 8, 1, gus_writeb, s);
     register_ioport_write (s->port + 0x100, 8, 2, gus_writew, s);
     register_ioport_read (s->port + 0x100, 8, 1, gus_readb, s);
     register_ioport_read (s->port + 0x100, 8, 2, gus_readw, s);
+    isa_init_ioport_range(dev, s->port + 0x100, 8);
 
     DMA_register_channel (s->emu.gusdma, GUS_read_DMA, s);
     s->emu.himemaddr = s->himem;
hw/ide/isa.c +2 -0
index 9856435..4206afd 100644
@@ -70,6 +70,8 @@ static int isa_ide_initfn(ISADevice *dev)
     ide_bus_new(&s->bus, &s->dev.qdev);
     ide_init_ioport(&s->bus, s->iobase, s->iobase2);
     isa_init_irq(dev, &s->irq, s->isairq);
+    isa_init_ioport_range(dev, s->iobase, 8);
+    isa_init_ioport(dev, s->iobase2);
     ide_init2(&s->bus, s->irq);
     vmstate_register(&dev->qdev, 0, &vmstate_ide_isa, s);
     return 0;
hw/isa-bus.c +25 -0
index 26036e0..c0ac7e9 100644
@@ -92,6 +92,31 @@ void isa_init_irq(ISADevice *dev, qemu_irq *p, int isairq)
     dev->nirqs++;
 }
 
+static void isa_init_ioport_one(ISADevice *dev, uint16_t ioport)
+{
+    assert(dev->nioports < ARRAY_SIZE(dev->ioports));
+    dev->ioports[dev->nioports++] = ioport;
+}
+
+static int isa_cmp_ports(const void *p1, const void *p2)
+{
+    return *(uint16_t*)p1 - *(uint16_t*)p2;
+}
+
+void isa_init_ioport_range(ISADevice *dev, uint16_t start, uint16_t length)
+{
+    int i;
+    for (i = start; i < start + length; i++) {
+        isa_init_ioport_one(dev, i);
+    }
+    qsort(dev->ioports, dev->nioports, sizeof(dev->ioports[0]), isa_cmp_ports);
+}
+
+void isa_init_ioport(ISADevice *dev, uint16_t ioport)
+{
+    isa_init_ioport_range(dev, ioport, 1);
+}
+
 static int isa_qdev_init(DeviceState *qdev, DeviceInfo *base)
 {
     ISADevice *dev = DO_UPCAST(ISADevice, qdev, qdev);
hw/isa.h +4 -0
index aaf0272..4794b76 100644
@@ -14,6 +14,8 @@ struct ISADevice {
     DeviceState qdev;
     uint32_t isairq[2];
     int nirqs;
+    uint16_t ioports[32];
+    int nioports;
 };
 
 typedef int (*isa_qdev_initfn)(ISADevice *dev);
@@ -26,6 +28,8 @@ ISABus *isa_bus_new(DeviceState *dev);
 void isa_bus_irqs(qemu_irq *irqs);
 qemu_irq isa_reserve_irq(int isairq);
 void isa_init_irq(ISADevice *dev, qemu_irq *p, int isairq);
+void isa_init_ioport(ISADevice *dev, uint16_t ioport);
+void isa_init_ioport_range(ISADevice *dev, uint16_t start, uint16_t length);
 void isa_qdev_register(ISADeviceInfo *info);
 ISADevice *isa_create(const char *name);
 ISADevice *isa_create_simple(const char *name);
hw/m48t59.c +1 -0
index c7492a6..75a94e1 100644
@@ -680,6 +680,7 @@ M48t59State *m48t59_init_isa(uint32_t io_base, uint16_t size, int type)
     if (io_base != 0) {
         register_ioport_read(io_base, 0x04, 1, NVRAM_readb, s);
         register_ioport_write(io_base, 0x04, 1, NVRAM_writeb, s);
+        isa_init_ioport_range(dev, io_base, 4);
     }
 
     return s;
hw/mc146818rtc.c +1 -0
index 2b91fa8..6466aff 100644
@@ -605,6 +605,7 @@ static int rtc_initfn(ISADevice *dev)
 
     register_ioport_write(base, 2, 1, cmos_ioport_write, s);
     register_ioport_read(base, 2, 1, cmos_ioport_read, s);
+    isa_init_ioport_range(dev, base, 2);
 
     qdev_set_legacy_instance_id(&dev->qdev, base, 2);
     qemu_register_reset(rtc_reset, s);
hw/ne2000-isa.c +3 -0
index 03a5a1f..3ff0d89 100644
@@ -68,14 +68,17 @@ static int isa_ne2000_initfn(ISADevice *dev)
 
     register_ioport_write(isa->iobase, 16, 1, ne2000_ioport_write, s);
     register_ioport_read(isa->iobase, 16, 1, ne2000_ioport_read, s);
+    isa_init_ioport_range(dev, isa->iobase, 16);
 
     register_ioport_write(isa->iobase + 0x10, 1, 1, ne2000_asic_ioport_write, s);
     register_ioport_read(isa->iobase + 0x10, 1, 1, ne2000_asic_ioport_read, s);
     register_ioport_write(isa->iobase + 0x10, 2, 2, ne2000_asic_ioport_write, s);
     register_ioport_read(isa->iobase + 0x10, 2, 2, ne2000_asic_ioport_read, s);
+    isa_init_ioport_range(dev, isa->iobase + 0x10, 2);
 
     register_ioport_write(isa->iobase + 0x1f, 1, 1, ne2000_reset_ioport_write, s);
     register_ioport_read(isa->iobase + 0x1f, 1, 1, ne2000_reset_ioport_read, s);
+    isa_init_ioport(dev, isa->iobase + 0x1f);
 
     isa_init_irq(dev, &s->irq, isa->isairq);
 
hw/parallel.c +5 -0
index 6b11672..6270b53 100644
@@ -481,16 +481,21 @@ static int parallel_isa_initfn(ISADevice *dev)
     if (s->hw_driver) {
         register_ioport_write(base, 8, 1, parallel_ioport_write_hw, s);
         register_ioport_read(base, 8, 1, parallel_ioport_read_hw, s);
+        isa_init_ioport_range(dev, base, 8);
+
         register_ioport_write(base+4, 1, 2, parallel_ioport_eppdata_write_hw2, s);
         register_ioport_read(base+4, 1, 2, parallel_ioport_eppdata_read_hw2, s);
         register_ioport_write(base+4, 1, 4, parallel_ioport_eppdata_write_hw4, s);
         register_ioport_read(base+4, 1, 4, parallel_ioport_eppdata_read_hw4, s);
+        isa_init_ioport(dev, base+4);
         register_ioport_write(base+0x400, 8, 1, parallel_ioport_ecp_write, s);
         register_ioport_read(base+0x400, 8, 1, parallel_ioport_ecp_read, s);
+        isa_init_ioport_range(dev, base+0x400, 8);
     }
     else {
         register_ioport_write(base, 8, 1, parallel_ioport_write_sw, s);
         register_ioport_read(base, 8, 1, parallel_ioport_read_sw, s);
+        isa_init_ioport_range(dev, base, 8);
     }
     return 0;
 }
hw/pckbd.c +3 -0
index 6e4e406..e736505 100644
@@ -484,10 +484,13 @@ static int i8042_initfn(ISADevice *dev)
 
     register_ioport_read(0x60, 1, 1, kbd_read_data, s);
     register_ioport_write(0x60, 1, 1, kbd_write_data, s);
+    isa_init_ioport(dev, 0x60);
     register_ioport_read(0x64, 1, 1, kbd_read_status, s);
     register_ioport_write(0x64, 1, 1, kbd_write_command, s);
+    isa_init_ioport(dev, 0x64);
     register_ioport_read(0x92, 1, 1, ioport92_read, s);
     register_ioport_write(0x92, 1, 1, ioport92_write, s);
+    isa_init_ioport(dev, 0x92);
 
     s->kbd = ps2_kbd_init(kbd_update_kbd_irq, s);
     s->mouse = ps2_mouse_init(kbd_update_aux_irq, s);
hw/sb16.c +4 -0
index 78590a7..c9d37ad 100644
@@ -1368,16 +1368,20 @@ static int sb16_initfn (ISADevice *dev)
 
     for (i = 0; i < ARRAY_SIZE (dsp_write_ports); i++) {
         register_ioport_write (s->port + dsp_write_ports[i], 1, 1, dsp_write, s);
+        isa_init_ioport(dev, s->port + dsp_write_ports[i]);
     }
 
     for (i = 0; i < ARRAY_SIZE (dsp_read_ports); i++) {
         register_ioport_read (s->port + dsp_read_ports[i], 1, 1, dsp_read, s);
+        isa_init_ioport(dev, s->port + dsp_read_ports[i]);
     }
 
     register_ioport_write (s->port + 0x4, 1, 1, mixer_write_indexb, s);
     register_ioport_write (s->port + 0x4, 1, 2, mixer_write_indexw, s);
+    isa_init_ioport(dev, s->port + 0x4);
     register_ioport_read (s->port + 0x5, 1, 1, mixer_read, s);
     register_ioport_write (s->port + 0x5, 1, 1, mixer_write_datab, s);
+    isa_init_ioport(dev, s->port + 0x5);
 
     DMA_register_channel (s->hdma, SB_read_DMA, s);
     DMA_register_channel (s->dma, SB_read_DMA, s);
hw/serial.c +1 -0
index 9ebc452..20902ae 100644
@@ -778,6 +778,7 @@ static int serial_isa_initfn(ISADevice *dev)
 
     register_ioport_write(isa->iobase, 8, 1, serial_ioport_write, s);
     register_ioport_read(isa->iobase, 8, 1, serial_ioport_read, s);
+    isa_init_ioport_range(dev, isa->iobase, 8);
     return 0;
 }
 
Signature · 1 line
1.7.1
#4 lore raw

PATCH v4 4/15 Add get_fw_dev_path callback to ISA bus in qdev.

In reply to Gleb Natapov

Use device ioports to create unique device path.

---
 hw/isa-bus.c |   16 ++++++++++++++++
 1 files changed, 16 insertions(+), 0 deletions(-)
hw/isa-bus.c +16 -0
index c0ac7e9..c423c1b 100644
@@ -31,11 +31,13 @@ static ISABus *isabus;
 target_phys_addr_t isa_mem_base = 0;
 
 static void isabus_dev_print(Monitor *mon, DeviceState *dev, int indent);
+static char *isabus_get_fw_dev_path(DeviceState *dev);
 
 static struct BusInfo isa_bus_info = {
     .name      = "ISA",
     .size      = sizeof(ISABus),
     .print_dev = isabus_dev_print,
+    .get_fw_dev_path = isabus_get_fw_dev_path,
 };
 
 ISABus *isa_bus_new(DeviceState *dev)
@@ -188,4 +190,18 @@ static void isabus_register_devices(void)
     sysbus_register_withprop(&isabus_bridge_info);
 }
 
+static char *isabus_get_fw_dev_path(DeviceState *dev)
+{
+    ISADevice *d = (ISADevice*)dev;
+    char path[40];
+    int off;
+
+    off = snprintf(path, sizeof(path), "%s", qdev_fw_name(dev));
+    if (d->nioports) {
+        snprintf(path + off, sizeof(path) - off, "@%04x", d->ioports[0]);
+    }
+
+    return strdup(path);
+}
+
 device_init(isabus_register_devices)
Signature · 1 line
1.7.1
#5 lore raw

PATCH v4 5/15 Store IDE bus id in IDEBus structure for easy access.

In reply to Gleb Natapov
---
 hw/ide/cmd646.c   |    4 ++--
 hw/ide/internal.h |    3 ++-
 hw/ide/isa.c      |    2 +-
 hw/ide/piix.c     |    4 ++--
 hw/ide/qdev.c     |    3 ++-
 hw/ide/via.c      |    4 ++--
 6 files changed, 11 insertions(+), 9 deletions(-)
hw/ide/cmd646.c +2 -2
index ff80dd5..b2cbdbc 100644
@@ -257,8 +257,8 @@ static int pci_cmd646_ide_initfn(PCIDevice *dev)
     pci_conf[PCI_INTERRUPT_PIN] = 0x01; // interrupt on pin 1
 
     irq = qemu_allocate_irqs(cmd646_set_irq, d, 2);
-    ide_bus_new(&d->bus[0], &d->dev.qdev);
-    ide_bus_new(&d->bus[1], &d->dev.qdev);
+    ide_bus_new(&d->bus[0], &d->dev.qdev, 0);
+    ide_bus_new(&d->bus[1], &d->dev.qdev, 1);
     ide_init2(&d->bus[0], irq[0]);
     ide_init2(&d->bus[1], irq[1]);
 
hw/ide/internal.h +2 -1
index d652e06..c0a1abc 100644
@@ -448,6 +448,7 @@ struct IDEBus {
     IDEDevice *slave;
     BMDMAState *bmdma;
     IDEState ifs[2];
+    int bus_id;
     uint8_t unit;
     uint8_t cmd;
     qemu_irq irq;
@@ -565,7 +566,7 @@ void ide_init2_with_non_qdev_drives(IDEBus *bus, DriveInfo *hd0,
 void ide_init_ioport(IDEBus *bus, int iobase, int iobase2);
 
 /* hw/ide/qdev.c */
-void ide_bus_new(IDEBus *idebus, DeviceState *dev);
+void ide_bus_new(IDEBus *idebus, DeviceState *dev, int bus_id);
 IDEDevice *ide_create_drive(IDEBus *bus, int unit, DriveInfo *drive);
 
 #endif /* HW_IDE_INTERNAL_H */
hw/ide/isa.c +1 -1
index 4206afd..8c59c5a 100644
@@ -67,7 +67,7 @@ static int isa_ide_initfn(ISADevice *dev)
 {
     ISAIDEState *s = DO_UPCAST(ISAIDEState, dev, dev);
 
-    ide_bus_new(&s->bus, &s->dev.qdev);
+    ide_bus_new(&s->bus, &s->dev.qdev, 0);
     ide_init_ioport(&s->bus, s->iobase, s->iobase2);
     isa_init_irq(dev, &s->irq, s->isairq);
     isa_init_ioport_range(dev, s->iobase, 8);
hw/ide/piix.c +2 -2
index 07483e8..d0b04a3 100644
@@ -129,8 +129,8 @@ static int pci_piix_ide_initfn(PCIIDEState *d)
 
     vmstate_register(&d->dev.qdev, 0, &vmstate_ide_pci, d);
 
-    ide_bus_new(&d->bus[0], &d->dev.qdev);
-    ide_bus_new(&d->bus[1], &d->dev.qdev);
+    ide_bus_new(&d->bus[0], &d->dev.qdev, 0);
+    ide_bus_new(&d->bus[1], &d->dev.qdev, 1);
     ide_init_ioport(&d->bus[0], 0x1f0, 0x3f6);
     ide_init_ioport(&d->bus[1], 0x170, 0x376);
 
hw/ide/qdev.c +2 -1
index 6d27b60..88ff657 100644
@@ -29,9 +29,10 @@ static struct BusInfo ide_bus_info = {
     .size  = sizeof(IDEBus),
 };
 
-void ide_bus_new(IDEBus *idebus, DeviceState *dev)
+void ide_bus_new(IDEBus *idebus, DeviceState *dev, int bus_id)
 {
     qbus_create_inplace(&idebus->qbus, &ide_bus_info, dev, NULL);
+    idebus->bus_id = bus_id;
 }
 
 static int ide_qdev_init(DeviceState *qdev, DeviceInfo *base)
hw/ide/via.c +2 -2
index b2c7cad..cc48b2b 100644
@@ -158,8 +158,8 @@ static int vt82c686b_ide_initfn(PCIDevice *dev)
 
     vmstate_register(&dev->qdev, 0, &vmstate_ide_pci, d);
 
-    ide_bus_new(&d->bus[0], &d->dev.qdev);
-    ide_bus_new(&d->bus[1], &d->dev.qdev);
+    ide_bus_new(&d->bus[0], &d->dev.qdev, 0);
+    ide_bus_new(&d->bus[1], &d->dev.qdev, 1);
     ide_init2(&d->bus[0], isa_reserve_irq(14));
     ide_init2(&d->bus[1], isa_reserve_irq(15));
     ide_init_ioport(&d->bus[0], 0x1f0, 0x3f6);
Signature · 1 line
1.7.1
#6 lore raw

PATCH v4 6/15 Add get_fw_dev_path callback to IDE bus.

In reply to Gleb Natapov
---
 hw/ide/qdev.c |   13 +++++++++++++
 1 files changed, 13 insertions(+), 0 deletions(-)
hw/ide/qdev.c +13 -0
index 88ff657..01a181b 100644
@@ -24,9 +24,12 @@
 
 /* --------------------------------- */
 
+static char *idebus_get_fw_dev_path(DeviceState *dev);
+
 static struct BusInfo ide_bus_info = {
     .name  = "IDE",
     .size  = sizeof(IDEBus),
+    .get_fw_dev_path = idebus_get_fw_dev_path,
 };
 
 void ide_bus_new(IDEBus *idebus, DeviceState *dev, int bus_id)
@@ -35,6 +38,16 @@ void ide_bus_new(IDEBus *idebus, DeviceState *dev, int bus_id)
     idebus->bus_id = bus_id;
 }
 
+static char *idebus_get_fw_dev_path(DeviceState *dev)
+{
+    char path[30];
+
+    snprintf(path, sizeof(path), "%s@%d", qdev_fw_name(dev),
+             ((IDEBus*)dev->parent_bus)->bus_id);
+
+    return strdup(path);
+}
+
 static int ide_qdev_init(DeviceState *qdev, DeviceInfo *base)
 {
     IDEDevice *dev = DO_UPCAST(IDEDevice, qdev, qdev);
Signature · 1 line
1.7.1
#7 lore raw

PATCH v4 7/15 Add get_dev_path callback for system bus.

In reply to Gleb Natapov

Prints out mmio or pio used to access child device.

---
 hw/pci_host.c |    2 ++
 hw/sysbus.c   |   30 ++++++++++++++++++++++++++++++
 hw/sysbus.h   |    4 ++++
 3 files changed, 36 insertions(+), 0 deletions(-)
hw/pci_host.c +2 -0
index bc5b771..28d45bf 100644
@@ -197,6 +197,7 @@ void pci_host_conf_register_ioport(pio_addr_t ioport, PCIHostState *s)
 {
     pci_host_init(s);
     register_ioport_simple(&s->conf_noswap_handler, ioport, 4, 4);
+    sysbus_init_ioports(&s->busdev, ioport, 4);
 }
 
 int pci_host_data_register_mmio(PCIHostState *s, int swap)
@@ -215,4 +216,5 @@ void pci_host_data_register_ioport(pio_addr_t ioport, PCIHostState *s)
     register_ioport_simple(&s->data_noswap_handler, ioport, 4, 1);
     register_ioport_simple(&s->data_noswap_handler, ioport, 4, 2);
     register_ioport_simple(&s->data_noswap_handler, ioport, 4, 4);
+    sysbus_init_ioports(&s->busdev, ioport, 4);
 }
hw/sysbus.c +30 -0
index d817721..1583bd8 100644
@@ -22,11 +22,13 @@
 #include "monitor.h"
 
 static void sysbus_dev_print(Monitor *mon, DeviceState *dev, int indent);
+static char *sysbus_get_fw_dev_path(DeviceState *dev);
 
 struct BusInfo system_bus_info = {
     .name       = "System",
     .size       = sizeof(BusState),
     .print_dev  = sysbus_dev_print,
+    .get_fw_dev_path = sysbus_get_fw_dev_path,
 };
 
 void sysbus_connect_irq(SysBusDevice *dev, int n, qemu_irq irq)
@@ -106,6 +108,16 @@ void sysbus_init_mmio_cb(SysBusDevice *dev, target_phys_addr_t size,
     dev->mmio[n].cb = cb;
 }
 
+void sysbus_init_ioports(SysBusDevice *dev, pio_addr_t ioport, pio_addr_t size)
+{
+    pio_addr_t i;
+
+    for (i = 0; i < size; i++) {
+        assert(dev->num_pio < QDEV_MAX_PIO);
+        dev->pio[dev->num_pio++] = ioport++;
+    }
+}
+
 static int sysbus_device_init(DeviceState *dev, DeviceInfo *base)
 {
     SysBusDeviceInfo *info = container_of(base, SysBusDeviceInfo, qdev);
@@ -171,3 +183,21 @@ static void sysbus_dev_print(Monitor *mon, DeviceState *dev, int indent)
                        indent, "", s->mmio[i].addr, s->mmio[i].size);
     }
 }
+
+static char *sysbus_get_fw_dev_path(DeviceState *dev)
+{
+    SysBusDevice *s = sysbus_from_qdev(dev);
+    char path[40];
+    int off;
+
+    off = snprintf(path, sizeof(path), "%s", qdev_fw_name(dev));
+
+    if (s->num_mmio) {
+        snprintf(path + off, sizeof(path) - off, "@"TARGET_FMT_plx,
+                 s->mmio[0].addr);
+    } else if (s->num_pio) {
+        snprintf(path + off, sizeof(path) - off, "@i%04x", s->pio[0]);
+    }
+
+    return strdup(path);
+}
hw/sysbus.h +4 -0
index 5980901..e9eb618 100644
@@ -6,6 +6,7 @@
 #include "qdev.h"
 
 #define QDEV_MAX_MMIO 32
+#define QDEV_MAX_PIO 32
 #define QDEV_MAX_IRQ 256
 
 typedef struct SysBusDevice SysBusDevice;
@@ -23,6 +24,8 @@ struct SysBusDevice {
         mmio_mapfunc cb;
         ram_addr_t iofunc;
     } mmio[QDEV_MAX_MMIO];
+    int num_pio;
+    pio_addr_t pio[QDEV_MAX_PIO];
 };
 
 typedef int (*sysbus_initfn)(SysBusDevice *dev);
@@ -45,6 +48,7 @@ void sysbus_init_mmio_cb(SysBusDevice *dev, target_phys_addr_t size,
                             mmio_mapfunc cb);
 void sysbus_init_irq(SysBusDevice *dev, qemu_irq *p);
 void sysbus_pass_irq(SysBusDevice *dev, SysBusDevice *target);
+void sysbus_init_ioports(SysBusDevice *dev, pio_addr_t ioport, pio_addr_t size);
 
 
 void sysbus_connect_irq(SysBusDevice *dev, int n, qemu_irq irq);
Signature · 1 line
1.7.1
#8 lore raw

PATCH v4 8/15 Add get_fw_dev_path callback for pci bus.

In reply to Gleb Natapov
---
 hw/pci.c |  108 ++++++++++++++++++++++++++++++++++++++++++++++++-------------
 1 files changed, 85 insertions(+), 23 deletions(-)
hw/pci.c +85 -23
index 962886e..114b435 100644
@@ -43,12 +43,14 @@
 
 static void pcibus_dev_print(Monitor *mon, DeviceState *dev, int indent);
 static char *pcibus_get_dev_path(DeviceState *dev);
+static char *pcibus_get_fw_dev_path(DeviceState *dev);
 
 struct BusInfo pci_bus_info = {
     .name       = "PCI",
     .size       = sizeof(PCIBus),
     .print_dev  = pcibus_dev_print,
     .get_dev_path = pcibus_get_dev_path,
+    .get_fw_dev_path = pcibus_get_fw_dev_path,
     .props      = (Property[]) {
         DEFINE_PROP_PCI_DEVFN("addr", PCIDevice, devfn, -1),
         DEFINE_PROP_STRING("romfile", PCIDevice, romfile),
@@ -1061,45 +1063,63 @@ void pci_msi_notify(PCIDevice *dev, unsigned int vector)
 typedef struct {
     uint16_t class;
     const char *desc;
+    const char *fw_name;
+    uint16_t fw_ign_bits;
 } pci_class_desc;
 
 static const pci_class_desc pci_class_descriptions[] =
 {
-    { 0x0100, "SCSI controller"},
-    { 0x0101, "IDE controller"},
-    { 0x0102, "Floppy controller"},
-    { 0x0103, "IPI controller"},
-    { 0x0104, "RAID controller"},
+    { 0x0001, "VGA controller", "display"},
+    { 0x0100, "SCSI controller", "scsi"},
+    { 0x0101, "IDE controller", "ide"},
+    { 0x0102, "Floppy controller", "fdc"},
+    { 0x0103, "IPI controller", "ipi"},
+    { 0x0104, "RAID controller", "raid"},
     { 0x0106, "SATA controller"},
     { 0x0107, "SAS controller"},
     { 0x0180, "Storage controller"},
-    { 0x0200, "Ethernet controller"},
-    { 0x0201, "Token Ring controller"},
-    { 0x0202, "FDDI controller"},
-    { 0x0203, "ATM controller"},
+    { 0x0200, "Ethernet controller", "ethernet"},
+    { 0x0201, "Token Ring controller", "token-ring"},
+    { 0x0202, "FDDI controller", "fddi"},
+    { 0x0203, "ATM controller", "atm"},
     { 0x0280, "Network controller"},
-    { 0x0300, "VGA controller"},
+    { 0x0300, "VGA controller", "display", 0x00ff},
     { 0x0301, "XGA controller"},
     { 0x0302, "3D controller"},
     { 0x0380, "Display controller"},
-    { 0x0400, "Video controller"},
-    { 0x0401, "Audio controller"},
+    { 0x0400, "Video controller", "video"},
+    { 0x0401, "Audio controller", "sound"},
     { 0x0402, "Phone"},
     { 0x0480, "Multimedia controller"},
-    { 0x0500, "RAM controller"},
-    { 0x0501, "Flash controller"},
+    { 0x0500, "RAM controller", "memory"},
+    { 0x0501, "Flash controller", "flash"},
     { 0x0580, "Memory controller"},
-    { 0x0600, "Host bridge"},
-    { 0x0601, "ISA bridge"},
-    { 0x0602, "EISA bridge"},
-    { 0x0603, "MC bridge"},
-    { 0x0604, "PCI bridge"},
-    { 0x0605, "PCMCIA bridge"},
-    { 0x0606, "NUBUS bridge"},
-    { 0x0607, "CARDBUS bridge"},
+    { 0x0600, "Host bridge", "host"},
+    { 0x0601, "ISA bridge", "isa"},
+    { 0x0602, "EISA bridge", "eisa"},
+    { 0x0603, "MC bridge", "mca"},
+    { 0x0604, "PCI bridge", "pci"},
+    { 0x0605, "PCMCIA bridge", "pcmcia"},
+    { 0x0606, "NUBUS bridge", "nubus"},
+    { 0x0607, "CARDBUS bridge", "cardbus"},
     { 0x0608, "RACEWAY bridge"},
     { 0x0680, "Bridge"},
-    { 0x0c03, "USB controller"},
+    { 0x0700, "Serial port", "serial"},
+    { 0x0701, "Parallel port", "parallel"},
+    { 0x0800, "Interrupt controller", "interrupt-controller"},
+    { 0x0801, "DMA controller", "dma-controller"},
+    { 0x0802, "Timer", "timer"},
+    { 0x0803, "RTC", "rtc"},
+    { 0x0900, "Keyboard", "keyboard"},
+    { 0x0901, "Pen", "pen"},
+    { 0x0902, "Mouse", "mouse"},
+    { 0x0A00, "Dock station", "dock", 0x00ff},
+    { 0x0B00, "i386 cpu", "cpu", 0x00ff},
+    { 0x0c00, "Fireware contorller", "fireware"},
+    { 0x0c01, "Access bus controller", "access-bus"},
+    { 0x0c02, "SSA controller", "ssa"},
+    { 0x0c03, "USB controller", "usb"},
+    { 0x0c04, "Fibre channel controller", "fibre-channel"},
     { 0, NULL}
 };
 
@@ -1825,6 +1845,48 @@ static void pcibus_dev_print(Monitor *mon, DeviceState *dev, int indent)
     }
 }
 
+static char *pci_dev_fw_name(DeviceState *dev, char *buf, int len)
+{
+    PCIDevice *d = (PCIDevice *)dev;
+    const char *name = NULL;
+    const pci_class_desc *desc =  pci_class_descriptions;
+    int class = pci_get_word(d->config + PCI_CLASS_DEVICE);
+
+    while (desc->desc &&
+          (class & ~desc->fw_ign_bits) !=
+          (desc->class & ~desc->fw_ign_bits)) {
+        desc++;
+    }
+
+    if (desc->desc) {
+        name = desc->fw_name;
+    }
+
+    if (name) {
+        pstrcpy(buf, len, name);
+    } else {
+        snprintf(buf, len, "pci%04x,%04x",
+                 pci_get_word(d->config + PCI_VENDOR_ID),
+                 pci_get_word(d->config + PCI_DEVICE_ID));
+    }
+
+    return buf;
+}
+
+static char *pcibus_get_fw_dev_path(DeviceState *dev)
+{
+    PCIDevice *d = (PCIDevice *)dev;
+    char path[50], name[33];
+    int off;
+
+    off = snprintf(path, sizeof(path), "%s@%x",
+                   pci_dev_fw_name(dev, name, sizeof name),
+                   PCI_SLOT(d->devfn));
+    if (PCI_FUNC(d->devfn))
+        snprintf(path + off, sizeof(path) + off, ",%x", PCI_FUNC(d->devfn));
+    return strdup(path);
+}
+
 static char *pcibus_get_dev_path(DeviceState *dev)
 {
     PCIDevice *d = (PCIDevice *)dev;
Signature · 1 line
1.7.1
#9 lore raw

Re: [PATCHv4 08/15] Add get_fw_dev_path callback for pci bus.

Quoted 145 lines · Gleb Natapov
---
 hw/pci.c |  108 ++++++++++++++++++++++++++++++++++++++++++++++++-------------
 1 files changed, 85 insertions(+), 23 deletions(-)
hw/pci.c +72 -23
index 962886e..114b435 100644
@@ -43,12 +43,14 @@
 
 static void pcibus_dev_print(Monitor *mon, DeviceState *dev, int indent);
 static char *pcibus_get_dev_path(DeviceState *dev);
+static char *pcibus_get_fw_dev_path(DeviceState *dev);
 
 struct BusInfo pci_bus_info = {
     .name       = "PCI",
     .size       = sizeof(PCIBus),
     .print_dev  = pcibus_dev_print,
     .get_dev_path = pcibus_get_dev_path,
+    .get_fw_dev_path = pcibus_get_fw_dev_path,
     .props      = (Property[]) {
         DEFINE_PROP_PCI_DEVFN("addr", PCIDevice, devfn, -1),
         DEFINE_PROP_STRING("romfile", PCIDevice, romfile),
@@ -1061,45 +1063,63 @@ void pci_msi_notify(PCIDevice *dev, unsigned int vector)
 typedef struct {
     uint16_t class;
     const char *desc;
+    const char *fw_name;
+    uint16_t fw_ign_bits;
 } pci_class_desc;
 
 static const pci_class_desc pci_class_descriptions[] =
 {
-    { 0x0100, "SCSI controller"},
-    { 0x0101, "IDE controller"},
-    { 0x0102, "Floppy controller"},
-    { 0x0103, "IPI controller"},
-    { 0x0104, "RAID controller"},
+    { 0x0001, "VGA controller", "display"},
+    { 0x0100, "SCSI controller", "scsi"},
+    { 0x0101, "IDE controller", "ide"},
+    { 0x0102, "Floppy controller", "fdc"},
+    { 0x0103, "IPI controller", "ipi"},
+    { 0x0104, "RAID controller", "raid"},
     { 0x0106, "SATA controller"},
     { 0x0107, "SAS controller"},
     { 0x0180, "Storage controller"},
-    { 0x0200, "Ethernet controller"},
-    { 0x0201, "Token Ring controller"},
-    { 0x0202, "FDDI controller"},
-    { 0x0203, "ATM controller"},
+    { 0x0200, "Ethernet controller", "ethernet"},
+    { 0x0201, "Token Ring controller", "token-ring"},
+    { 0x0202, "FDDI controller", "fddi"},
+    { 0x0203, "ATM controller", "atm"},
     { 0x0280, "Network controller"},
-    { 0x0300, "VGA controller"},
+    { 0x0300, "VGA controller", "display", 0x00ff},
     { 0x0301, "XGA controller"},
     { 0x0302, "3D controller"},
     { 0x0380, "Display controller"},
-    { 0x0400, "Video controller"},
-    { 0x0401, "Audio controller"},
+    { 0x0400, "Video controller", "video"},
+    { 0x0401, "Audio controller", "sound"},
     { 0x0402, "Phone"},
     { 0x0480, "Multimedia controller"},
-    { 0x0500, "RAM controller"},
-    { 0x0501, "Flash controller"},
+    { 0x0500, "RAM controller", "memory"},
+    { 0x0501, "Flash controller", "flash"},
     { 0x0580, "Memory controller"},
-    { 0x0600, "Host bridge"},
-    { 0x0601, "ISA bridge"},
-    { 0x0602, "EISA bridge"},
-    { 0x0603, "MC bridge"},
-    { 0x0604, "PCI bridge"},
-    { 0x0605, "PCMCIA bridge"},
-    { 0x0606, "NUBUS bridge"},
-    { 0x0607, "CARDBUS bridge"},
+    { 0x0600, "Host bridge", "host"},
+    { 0x0601, "ISA bridge", "isa"},
+    { 0x0602, "EISA bridge", "eisa"},
+    { 0x0603, "MC bridge", "mca"},
+    { 0x0604, "PCI bridge", "pci"},
+    { 0x0605, "PCMCIA bridge", "pcmcia"},
+    { 0x0606, "NUBUS bridge", "nubus"},
+    { 0x0607, "CARDBUS bridge", "cardbus"},
     { 0x0608, "RACEWAY bridge"},
     { 0x0680, "Bridge"},
-    { 0x0c03, "USB controller"},
+    { 0x0700, "Serial port", "serial"},
+    { 0x0701, "Parallel port", "parallel"},
+    { 0x0800, "Interrupt controller", "interrupt-controller"},
+    { 0x0801, "DMA controller", "dma-controller"},
+    { 0x0802, "Timer", "timer"},
+    { 0x0803, "RTC", "rtc"},
+    { 0x0900, "Keyboard", "keyboard"},
+    { 0x0901, "Pen", "pen"},
+    { 0x0902, "Mouse", "mouse"},
+    { 0x0A00, "Dock station", "dock", 0x00ff},
+    { 0x0B00, "i386 cpu", "cpu", 0x00ff},
+    { 0x0c00, "Fireware contorller", "fireware"},
+    { 0x0c01, "Access bus controller", "access-bus"},
+    { 0x0c02, "SSA controller", "ssa"},
+    { 0x0c03, "USB controller", "usb"},
+    { 0x0c04, "Fibre channel controller", "fibre-channel"},
     { 0, NULL}
 };
 
@@ -1825,6 +1845,48 @@ static void pcibus_dev_print(Monitor *mon, DeviceState *dev, int indent)
     }
 }
 
+static char *pci_dev_fw_name(DeviceState *dev, char *buf, int len)
+{
+    PCIDevice *d = (PCIDevice *)dev;
+    const char *name = NULL;
+    const pci_class_desc *desc =  pci_class_descriptions;
+    int class = pci_get_word(d->config + PCI_CLASS_DEVICE);
+
+    while (desc->desc &&
+          (class & ~desc->fw_ign_bits) !=
+          (desc->class & ~desc->fw_ign_bits)) {
+        desc++;
+    }
+
+    if (desc->desc) {
+        name = desc->fw_name;
+    }
+
+    if (name) {
+        pstrcpy(buf, len, name);
+    } else {
+        snprintf(buf, len, "pci%04x,%04x",
+                 pci_get_word(d->config + PCI_VENDOR_ID),
+                 pci_get_word(d->config + PCI_DEVICE_ID));
+    }
+
+    return buf;
+}
+
+static char *pcibus_get_fw_dev_path(DeviceState *dev)

hw/pci.c static void pcibus_dev_print(Monitor *mon, DeviceState *dev, int indent) · line 1886

Quoted 10 lines
hw/pci.c +10 -0
+{
+    PCIDevice *d = (PCIDevice *)dev;
+    char path[50], name[33];
+    int off;
+
+    off = snprintf(path, sizeof(path), "%s@%x",
+                   pci_dev_fw_name(dev, name, sizeof name),
+                   PCI_SLOT(d->devfn));
+    if (PCI_FUNC(d->devfn))
+        snprintf(path + off, sizeof(path) + off, ",%x", PCI_FUNC(d->devfn));

Can we *always* specify a slot number in the name?
If yes I think we should, because I think we saw that the short form is
ambiguous: if there's a device at a slot != 0, openfirmware will to match that against the name.
Right?

Quoted 8 lines
hw/pci.c +3 -0
+    return strdup(path);
+}
+
 static char *pcibus_get_dev_path(DeviceState *dev)
 {
     PCIDevice *d = (PCIDevice *)dev;
Signature · 1 line
1.7.1
#10 lore raw

Re: [PATCHv4 08/15] Add get_fw_dev_path callback for pci bus.

Quoted 17 lines · Michael S. Tsirkin

hw/pci.c static void pcibus_dev_print(Monitor *mon, DeviceState *dev, int indent) · line 1886

Quoted 11 lines · Gleb Natapov
hw/pci.c +11 -0
+static char *pcibus_get_fw_dev_path(DeviceState *dev)
+{
+    PCIDevice *d = (PCIDevice *)dev;
+    char path[50], name[33];
+    int off;
+
+    off = snprintf(path, sizeof(path), "%s@%x",
+                   pci_dev_fw_name(dev, name, sizeof name),
+                   PCI_SLOT(d->devfn));
+    if (PCI_FUNC(d->devfn))
+        snprintf(path + off, sizeof(path) + off, ",%x", PCI_FUNC(d->devfn));

Can we *always* specify a slot number in the name?
If yes I think we should, because I think we saw that the short form is
ambiguous: if there's a device at a slot != 0, openfirmware will to match that against the name.

You accidentally something here.

Quoted 2 lines

Right?

I do not see what is ambiguous here. Can you explain?

--
Gleb.

#11 lore raw

PATCH v4 9/15 Record which USBDevice USBPort belongs too.

In reply to Gleb Natapov

Ports on root hub will have NULL here. This is needed to reconstruct path from device to its root hub to build device path.

---
 hw/usb-bus.c  |    3 ++-
 hw/usb-hub.c  |    2 +-
 hw/usb-musb.c |    2 +-
 hw/usb-ohci.c |    2 +-
 hw/usb-uhci.c |    2 +-
 hw/usb.h      |    3 ++-
 6 files changed, 8 insertions(+), 6 deletions(-)
hw/usb-bus.c +2 -1
index b692503..256b881 100644
@@ -110,11 +110,12 @@ USBDevice *usb_create_simple(USBBus *bus, const char *name)
 }
 
 void usb_register_port(USBBus *bus, USBPort *port, void *opaque, int index,
-                       usb_attachfn attach)
+                       USBDevice *pdev, usb_attachfn attach)
 {
     port->opaque = opaque;
     port->index = index;
     port->attach = attach;
+    port->pdev = pdev;
     QTAILQ_INSERT_TAIL(&bus->free, port, next);
     bus->nfree++;
 }
hw/usb-hub.c +1 -1
index 8e3a96b..8a3f829 100644
@@ -535,7 +535,7 @@ static int usb_hub_initfn(USBDevice *dev)
     for (i = 0; i < s->nb_ports; i++) {
         port = &s->ports[i];
         usb_register_port(usb_bus_from_device(dev),
-                          &port->port, s, i, usb_hub_attach);
+                          &port->port, s, i, &s->dev, usb_hub_attach);
         port->wPortStatus = PORT_STAT_POWER;
         port->wPortChange = 0;
     }
hw/usb-musb.c +1 -1
index 7f15842..9efe7a6 100644
@@ -343,7 +343,7 @@ struct MUSBState {
     }
 
     usb_bus_new(&s->bus, NULL /* FIXME */);
-    usb_register_port(&s->bus, &s->port, s, 0, musb_attach);
+    usb_register_port(&s->bus, &s->port, s, 0, NULL, musb_attach);
 
     return s;
 }
hw/usb-ohci.c +1 -1
index c60fd8d..59604cf 100644
@@ -1705,7 +1705,7 @@ static void usb_ohci_init(OHCIState *ohci, DeviceState *dev,
     usb_bus_new(&ohci->bus, dev);
     ohci->num_ports = num_ports;
     for (i = 0; i < num_ports; i++) {
-        usb_register_port(&ohci->bus, &ohci->rhport[i].port, ohci, i, ohci_attach);
+        usb_register_port(&ohci->bus, &ohci->rhport[i].port, ohci, i, NULL, ohci_attach);
     }
 
     ohci->async_td = 0;
hw/usb-uhci.c +1 -1
index 1d83400..b9b822f 100644
@@ -1115,7 +1115,7 @@ static int usb_uhci_common_initfn(UHCIState *s)
 
     usb_bus_new(&s->bus, &s->dev.qdev);
     for(i = 0; i < NB_PORTS; i++) {
-        usb_register_port(&s->bus, &s->ports[i].port, s, i, uhci_attach);
+        usb_register_port(&s->bus, &s->ports[i].port, s, i, NULL, uhci_attach);
     }
     s->frame_timer = qemu_new_timer(vm_clock, uhci_frame_timer, s);
     s->expire_time = qemu_get_clock(vm_clock) +
hw/usb.h +2 -1
index 00d2802..0b32d77 100644
@@ -203,6 +203,7 @@ struct USBPort {
     USBDevice *dev;
     usb_attachfn attach;
     void *opaque;
+    USBDevice *pdev;
     int index; /* internal port index, may be used with the opaque */
     QTAILQ_ENTRY(USBPort) next;
 };
@@ -312,7 +313,7 @@ USBDevice *usb_create(USBBus *bus, const char *name);
 USBDevice *usb_create_simple(USBBus *bus, const char *name);
 USBDevice *usbdevice_create(const char *cmdline);
 void usb_register_port(USBBus *bus, USBPort *port, void *opaque, int index,
-                       usb_attachfn attach);
+                       USBDevice *pdev, usb_attachfn attach);
 void usb_unregister_port(USBBus *bus, USBPort *port);
 int usb_device_attach(USBDevice *dev);
 int usb_device_detach(USBDevice *dev);
Signature · 1 line
1.7.1
#12 lore raw

PATCH v4 10/15 Add get_dev_path callback for usb bus.

In reply to Gleb Natapov
---
 hw/usb-bus.c |   42 ++++++++++++++++++++++++++++++++++++++++++
 1 files changed, 42 insertions(+), 0 deletions(-)
hw/usb-bus.c +42 -0
index 256b881..8b4583c 100644
@@ -5,11 +5,13 @@
 #include "monitor.h"
 
 static void usb_bus_dev_print(Monitor *mon, DeviceState *qdev, int indent);
+static char *usbbus_get_fw_dev_path(DeviceState *dev);
 
 static struct BusInfo usb_bus_info = {
     .name      = "USB",
     .size      = sizeof(USBBus),
     .print_dev = usb_bus_dev_print,
+    .get_fw_dev_path = usbbus_get_fw_dev_path,
 };
 static int next_usb_bus = 0;
 static QTAILQ_HEAD(, USBBus) busses = QTAILQ_HEAD_INITIALIZER(busses);
@@ -307,3 +309,43 @@ USBDevice *usbdevice_create(const char *cmdline)
     }
     return usb->usbdevice_init(params);
 }
+
+static int usbbus_get_fw_dev_path_helper(USBDevice *d, USBBus *bus, char *p,
+                                         int len)
+{
+    int l = 0;
+    USBPort *port;
+
+    QTAILQ_FOREACH(port, &bus->used, next) {
+        if (port->dev == d) {
+            if (port->pdev) {
+                l = usbbus_get_fw_dev_path_helper(port->pdev, bus, p, len);
+            }
+            l += snprintf(p + l, len - l, "%s@%x/", qdev_fw_name(&d->qdev),
+                          port->index);
+            break;
+        }
+    }
+
+    return l;
+}
+
+static char *usbbus_get_fw_dev_path(DeviceState *dev)
+{
+    USBDevice *d = (USBDevice*)dev;
+    USBBus *bus = usb_bus_from_device(d);
+    char path[100];
+    int l;
+
+    assert(d->attached != 0);
+
+    l = usbbus_get_fw_dev_path_helper(d, bus, path, sizeof(path));
+
+    if (l == 0) {
+        abort();
+    }
+
+    path[l-1] = '\0';
+
+    return strdup(path);
+}
Signature · 1 line
1.7.1
#13 lore raw

PATCH v4 11/15 Add bootindex parameter to net/block/fd device

In reply to Gleb Natapov

If bootindex is specified on command line a string that describes device in firmware readable way is added into sorted list. Later this list will be passed into firmware to control boot order.

---
 block_int.h     |    4 +++-
 hw/e1000.c      |    4 ++++
 hw/eepro100.c   |    3 +++
 hw/fdc.c        |    8 ++++++++
 hw/ide/qdev.c   |    5 +++++
 hw/ne2000.c     |    3 +++
 hw/pcnet.c      |    4 ++++
 hw/qdev.c       |   32 ++++++++++++++++++++++++++++++++
 hw/qdev.h       |    1 +
 hw/rtl8139.c    |    4 ++++
 hw/usb-net.c    |    2 ++
 hw/virtio-blk.c |    2 ++
 hw/virtio-net.c |    2 ++
 net.h           |    4 +++-
 sysemu.h        |    2 ++
 vl.c            |   40 ++++++++++++++++++++++++++++++++++++++++
 16 files changed, 118 insertions(+), 2 deletions(-)
block_int.h +3 -1
index 3c3adb5..0a0e47d 100644
@@ -227,6 +227,7 @@ typedef struct BlockConf {
     uint16_t logical_block_size;
     uint16_t min_io_size;
     uint32_t opt_io_size;
+    int32_t bootindex;
 } BlockConf;
 
 static inline unsigned int get_physical_block_exp(BlockConf *conf)
@@ -249,6 +250,7 @@ static inline unsigned int get_physical_block_exp(BlockConf *conf)
     DEFINE_PROP_UINT16("physical_block_size", _state,                   \
                        _conf.physical_block_size, 512),                 \
     DEFINE_PROP_UINT16("min_io_size", _state, _conf.min_io_size, 0),  \
-    DEFINE_PROP_UINT32("opt_io_size", _state, _conf.opt_io_size, 0)
+    DEFINE_PROP_UINT32("opt_io_size", _state, _conf.opt_io_size, 0),    \
+    DEFINE_PROP_INT32("bootindex", _state, _conf.bootindex, -1)         \
 
 #endif /* BLOCK_INT_H */
hw/e1000.c +4 -0
index 532efdc..053f33e 100644
@@ -30,6 +30,7 @@
 #include "net.h"
 #include "net/checksum.h"
 #include "loader.h"
+#include "sysemu.h"
 
 #include "e1000_hw.h"
 
@@ -1148,6 +1149,9 @@ static int pci_e1000_init(PCIDevice *pci_dev)
                           d->dev.qdev.info->name, d->dev.qdev.id, d);
 
     qemu_format_nic_info_str(&d->nic->nc, macaddr);
+
+    add_boot_device_path(d->conf.bootindex, &pci_dev->qdev, "ethernet-phy@0");
+
     return 0;
 }
 
hw/eepro100.c +3 -0
index 41d792a..80adac6 100644
@@ -46,6 +46,7 @@
 #include "pci.h"
 #include "net.h"
 #include "eeprom93xx.h"
+#include "sysemu.h"
 
 #define KiB 1024
 
@@ -1907,6 +1908,8 @@ static int e100_nic_init(PCIDevice *pci_dev)
     s->vmstate->name = s->nic->nc.model;
     vmstate_register(&pci_dev->qdev, -1, s->vmstate, s);
 
+    add_boot_device_path(s->conf.bootindex, &pci_dev->qdev, "ethernet-phy@0");
+
     return 0;
 }
 
hw/fdc.c +8 -0
index 5ab754b..7b1349f 100644
@@ -35,6 +35,7 @@
 #include "sysbus.h"
 #include "qdev-addr.h"
 #include "blockdev.h"
+#include "sysemu.h"
 
 /********************************************************/
 /* debug Floppy devices */
@@ -523,6 +524,8 @@ typedef struct FDCtrlSysBus {
 typedef struct FDCtrlISABus {
     ISADevice busdev;
     struct FDCtrl state;
+    int32_t bootindexA;
+    int32_t bootindexB;
 } FDCtrlISABus;
 
 static uint32_t fdctrl_read (void *opaque, uint32_t reg)
@@ -1992,6 +1995,9 @@ static int isabus_fdc_init1(ISADevice *dev)
     qdev_set_legacy_instance_id(&dev->qdev, iobase, 2);
     ret = fdctrl_init_common(fdctrl);
 
+    add_boot_device_path(isa->bootindexA, &dev->qdev, "floppy@0");
+    add_boot_device_path(isa->bootindexB, &dev->qdev, "floppy@1");
+
     return ret;
 }
 
@@ -2051,6 +2057,8 @@ static ISADeviceInfo isa_fdc_info = {
     .qdev.props = (Property[]) {
         DEFINE_PROP_DRIVE("driveA", FDCtrlISABus, state.drives[0].bs),
         DEFINE_PROP_DRIVE("driveB", FDCtrlISABus, state.drives[1].bs),
+        DEFINE_PROP_INT32("bootindexA", FDCtrlISABus, bootindexA, -1),
+        DEFINE_PROP_INT32("bootindexB", FDCtrlISABus, bootindexB, -1),
         DEFINE_PROP_END_OF_LIST(),
     },
 };
hw/ide/qdev.c +5 -0
index 01a181b..69a00e2 100644
@@ -21,6 +21,7 @@
 #include "qemu-error.h"
 #include <hw/ide/internal.h>
 #include "blockdev.h"
+#include "sysemu.h"
 
 /* --------------------------------- */
 
@@ -143,6 +144,10 @@ static int ide_drive_initfn(IDEDevice *dev)
     if (!dev->serial) {
         dev->serial = qemu_strdup(s->drive_serial_str);
     }
+
+    add_boot_device_path(dev->conf.bootindex, &dev->qdev,
+                         dev->unit ? "disk@1" : "disk@0");
+
     return 0;
 }
 
hw/ne2000.c +3 -0
index 126e7cf..f4bbac2 100644
@@ -26,6 +26,7 @@
 #include "net.h"
 #include "ne2000.h"
 #include "loader.h"
+#include "sysemu.h"
 
 /* debug NE2000 card */
 //#define DEBUG_NE2000
@@ -746,6 +747,8 @@ static int pci_ne2000_init(PCIDevice *pci_dev)
         }
     }
 
+    add_boot_device_path(s->c.bootindex, &pci_dev->qdev, "ethernet-phy@0");
+
     return 0;
 }
 
hw/pcnet.c +4 -0
index b52935a..a988f4f 100644
@@ -40,6 +40,7 @@
 #include "loader.h"
 #include "qemu-timer.h"
 #include "qemu_socket.h"
+#include "sysemu.h"
 
 #include "pcnet.h"
 
@@ -1898,6 +1899,9 @@ int pcnet_common_init(DeviceState *dev, PCNetState *s, NetClientInfo *info)
     qemu_macaddr_default_if_unset(&s->conf.macaddr);
     s->nic = qemu_new_nic(info, &s->conf, dev->info->name, dev->id, s);
     qemu_format_nic_info_str(&s->nic->nc, s->conf.macaddr.a);
+
+    add_boot_device_path(s->conf.bootindex, dev, "ethernet-phy@0");
+
     return 0;
 }
 
hw/qdev.c +32 -0
index 35858cb..cc9bf43 100644
@@ -820,3 +820,35 @@ int do_device_del(Monitor *mon, const QDict *qdict, QObject **ret_data)
     }
     return qdev_unplug(dev);
 }
+
+static int qdev_get_fw_dev_path_helper(DeviceState *dev, char *p, int size)
+{
+    int l = 0;
+
+    if (dev && dev->parent_bus) {
+        char *d;
+        l = qdev_get_fw_dev_path_helper(dev->parent_bus->parent, p, size);
+        if (dev->parent_bus->info->get_fw_dev_path) {
+            d = dev->parent_bus->info->get_fw_dev_path(dev);
+            l += snprintf(p + l, size - l, "%s", d);
+            qemu_free(d);
+        } else {
+            l += snprintf(p + l, size - l, "%s", dev->info->name);
+        }
+    }
+    l += snprintf(p + l , size - l, "/");
+
+    return l;
+}
+
+char* qdev_get_fw_dev_path(DeviceState *dev)
+{
+    char path[128];
+    int l;
+
+    l = qdev_get_fw_dev_path_helper(dev, path, 128);
+
+    path[l-1] = '\0';
+
+    return strdup(path);
+}
hw/qdev.h +1 -0
index dc669b3..13f3597 100644
@@ -296,6 +296,7 @@ static inline const char *qdev_fw_name(DeviceState *dev)
     return dev->info->fw_name ? : dev->info->alias ? : dev->info->name;
 }
 
+char *qdev_get_fw_dev_path(DeviceState *dev);
 /* This is a nasty hack to allow passing a NULL bus to qdev_create.  */
 extern struct BusInfo system_bus_info;
 
hw/rtl8139.c +4 -0
index d92981d..e44ffb8 100644
@@ -52,6 +52,7 @@
 #include "qemu-timer.h"
 #include "net.h"
 #include "loader.h"
+#include "sysemu.h"
 
 /* debug RTL8139 card */
 //#define DEBUG_RTL8139 1
@@ -3387,6 +3388,9 @@ static int pci_rtl8139_init(PCIDevice *dev)
     s->TimerExpire = 0;
     s->timer = qemu_new_timer(vm_clock, rtl8139_timer, s);
     rtl8139_set_next_tctr_time(s, qemu_get_clock(vm_clock));
+
+    add_boot_device_path(s->conf.bootindex, &dev->qdev, "ethernet-phy@0");
+
     return 0;
 }
 
hw/usb-net.c +2 -0
index 2287ee1..dac20a7 100644
@@ -27,6 +27,7 @@
 #include "usb.h"
 #include "net.h"
 #include "qemu-queue.h"
+#include "sysemu.h"
 
 /*#define TRAFFIC_DEBUG*/
 /* Thanks to NetChip Technologies for donating this product ID.
@@ -1463,6 +1464,7 @@ static int usb_net_initfn(USBDevice *dev)
              s->conf.macaddr.a[4],
              s->conf.macaddr.a[5]);
 
+    add_boot_device_path(s->conf.bootindex, &dev->qdev, "ethernet@0");
     return 0;
 }
 
hw/virtio-blk.c +2 -0
index 49528a9..ba11e29 100644
@@ -548,6 +548,8 @@ VirtIODevice *virtio_blk_init(DeviceState *dev, BlockConf *conf)
     bdrv_set_removable(s->bs, 0);
     s->bs->buffer_alignment = conf->logical_block_size;
 
+    add_boot_device_path(conf->bootindex, dev, "disk@0");
+
     return &s->vdev;
 }
 
hw/virtio-net.c +2 -0
index 7e1688c..a2ba40b 100644
@@ -1018,6 +1018,8 @@ VirtIODevice *virtio_net_init(DeviceState *dev, NICConf *conf,
                     virtio_net_save, virtio_net_load, n);
     n->vmstate = qemu_add_vm_change_state_handler(virtio_net_vmstate_change, n);
 
+    add_boot_device_path(conf->bootindex, dev, "ethernet-phy@0");
+
     return &n->vdev;
 }
 
net.h +3 -1
index 44c31a9..6ceca50 100644
@@ -17,12 +17,14 @@ typedef struct NICConf {
     MACAddr macaddr;
     VLANState *vlan;
     VLANClientState *peer;
+    int32_t bootindex;
 } NICConf;
 
 #define DEFINE_NIC_PROPERTIES(_state, _conf)                            \
     DEFINE_PROP_MACADDR("mac",   _state, _conf.macaddr),                \
     DEFINE_PROP_VLAN("vlan",     _state, _conf.vlan),                   \
-    DEFINE_PROP_NETDEV("netdev", _state, _conf.peer)
+    DEFINE_PROP_NETDEV("netdev", _state, _conf.peer),                   \
+    DEFINE_PROP_INT32("bootindex", _state, _conf.bootindex, -1)
 
 /* VLANs support */
 
sysemu.h +2 -0
index b81a70e..fe9a582 100644
@@ -188,4 +188,6 @@ void rtc_change_mon_event(struct tm *tm);
 
 void register_devices(void);
 
+void add_boot_device_path(int32_t bootindex, DeviceState *dev,
+                          const char *suffix);
 #endif
vl.c +40 -0
index c58583d..4c159f0 100644
@@ -228,6 +228,17 @@ unsigned int nb_prom_envs = 0;
 const char *prom_envs[MAX_PROM_ENVS];
 int boot_menu;
 
+typedef struct FWBootEntry FWBootEntry;
+
+struct FWBootEntry {
+    QTAILQ_ENTRY(FWBootEntry) link;
+    int32_t bootindex;
+    DeviceState *dev;
+    char *suffix;
+};
+
+QTAILQ_HEAD(, FWBootEntry) fw_boot_order = QTAILQ_HEAD_INITIALIZER(fw_boot_order);
+
 int nb_numa_nodes;
 uint64_t node_mem[MAX_NODES];
 uint64_t node_cpumask[MAX_NODES];
@@ -692,6 +703,35 @@ static void restore_boot_devices(void *opaque)
     qemu_free(standard_boot_devices);
 }
 
+void add_boot_device_path(int32_t bootindex, DeviceState *dev,
+                          const char *suffix)
+{
+    FWBootEntry *node, *i;
+
+    if (bootindex < 0) {
+        return;
+    }
+
+    assert(dev != NULL || suffix != NULL);
+
+    node = qemu_mallocz(sizeof(FWBootEntry));
+    node->bootindex = bootindex;
+    node->suffix = strdup(suffix);
+    node->dev = dev;
+
+    QTAILQ_FOREACH(i, &fw_boot_order, link) {
+        if (i->bootindex == bootindex) {
+            fprintf(stderr, "Two devices with same boot index %d\n", bootindex);
+            exit(1);
+        } else if (i->bootindex < bootindex) {
+            continue;
+        }
+        QTAILQ_INSERT_BEFORE(i, node, link);
+        return;
+    }
+    QTAILQ_INSERT_TAIL(&fw_boot_order, node, link);
+}
+
 static void numa_add(const char *optarg)
 {
     char option[128];
Signature · 1 line
1.7.1
#14 lore raw

PATCH v4 12/15 Change fw_cfg_add_file() to get full file path as a parameter.

In reply to Gleb Natapov

Change fw_cfg_add_file() to get full file path as a parameter instead of building one internally. Two reasons for that. First caller may need to know how file is named. Second this moves policy of file naming out from fw_cfg. Platform may want to use more then two levels of directories for instance.

---
 hw/fw_cfg.c |   16 ++++------------
 hw/fw_cfg.h |    4 ++--
 hw/loader.c |   16 ++++++++++++++--
 3 files changed, 20 insertions(+), 16 deletions(-)
hw/fw_cfg.c +4 -12
index 72866ae..7b9434f 100644
@@ -277,10 +277,9 @@ int fw_cfg_add_callback(FWCfgState *s, uint16_t key, FWCfgCallback callback,
     return 1;
 }
 
-int fw_cfg_add_file(FWCfgState *s,  const char *dir, const char *filename,
-                    uint8_t *data, uint32_t len)
+int fw_cfg_add_file(FWCfgState *s,  const char *filename, uint8_t *data,
+                    uint32_t len)
 {
-    const char *basename;
     int i, index;
 
     if (!s->files) {
@@ -297,15 +296,8 @@ int fw_cfg_add_file(FWCfgState *s,  const char *dir, const char *filename,
 
     fw_cfg_add_bytes(s, FW_CFG_FILE_FIRST + index, data, len);
 
-    basename = strrchr(filename, '/');
-    if (basename) {
-        basename++;
-    } else {
-        basename = filename;
-    }
-
-    snprintf(s->files->f[index].name, sizeof(s->files->f[index].name),
-             "%s/%s", dir, basename);
+    pstrcpy(s->files->f[index].name, sizeof(s->files->f[index].name),
+            filename);
     for (i = 0; i < index; i++) {
         if (strcmp(s->files->f[index].name, s->files->f[i].name) == 0) {
             FW_CFG_DPRINTF("%s: skip duplicate: %s\n", __FUNCTION__,
hw/fw_cfg.h +2 -2
index 4d13a4f..856bf91 100644
@@ -60,8 +60,8 @@ int fw_cfg_add_i32(FWCfgState *s, uint16_t key, uint32_t value);
 int fw_cfg_add_i64(FWCfgState *s, uint16_t key, uint64_t value);
 int fw_cfg_add_callback(FWCfgState *s, uint16_t key, FWCfgCallback callback,
                         void *callback_opaque, uint8_t *data, size_t len);
-int fw_cfg_add_file(FWCfgState *s, const char *dir, const char *filename,
-                    uint8_t *data, uint32_t len);
+int fw_cfg_add_file(FWCfgState *s, const char *filename, uint8_t *data,
+                    uint32_t len);
 FWCfgState *fw_cfg_init(uint32_t ctl_port, uint32_t data_port,
                         target_phys_addr_t crl_addr, target_phys_addr_t data_addr);
 
hw/loader.c +14 -2
index 49ac1fa..1e98326 100644
@@ -592,8 +592,20 @@ int rom_add_file(const char *file, const char *fw_dir,
     }
     close(fd);
     rom_insert(rom);
-    if (rom->fw_file && fw_cfg)
-        fw_cfg_add_file(fw_cfg, rom->fw_dir, rom->fw_file, rom->data, rom->romsize);
+    if (rom->fw_file && fw_cfg) {
+        const char *basename;
+        char fw_file_name[56];
+
+        basename = strrchr(rom->fw_file, '/');
+        if (basename) {
+            basename++;
+        } else {
+            basename = rom->fw_file;
+        }
+        snprintf(fw_file_name, sizeof(fw_file_name), "%s/%s", rom->fw_dir,
+                 basename);
+        fw_cfg_add_file(fw_cfg, fw_file_name, rom->data, rom->romsize);
+    }
     return 0;
 
 err:
Signature · 1 line
1.7.1
#15 lore raw

PATCH v4 13/15 Add bootindex for option roms.

In reply to Gleb Natapov

Extend -option-rom command to have additional parameter ,bootindex=.

---
 hw/loader.c    |   16 +++++++++++-----
 hw/loader.h    |    8 ++++----
 hw/multiboot.c |    3 ++-
 hw/ne2000.c    |    2 +-
 hw/nseries.c   |    2 +-
 hw/pc.c        |    7 ++++---
 hw/pci.c       |    2 +-
 hw/pcnet.c     |    2 +-
 qemu-config.c  |   17 +++++++++++++++++
 sysemu.h       |    6 +++++-
 vl.c           |   11 +++++++++--
 11 files changed, 56 insertions(+), 20 deletions(-)
hw/loader.c +11 -5
index 1e98326..eb198f6 100644
@@ -107,7 +107,7 @@ int load_image_targphys(const char *filename,
 
     size = get_image_size(filename);
     if (size > 0)
-        rom_add_file_fixed(filename, addr);
+        rom_add_file_fixed(filename, addr, -1);
     return size;
 }
 
@@ -557,10 +557,11 @@ static void rom_insert(Rom *rom)
 }
 
 int rom_add_file(const char *file, const char *fw_dir,
-                 target_phys_addr_t addr)
+                 target_phys_addr_t addr, int32_t bootindex)
 {
     Rom *rom;
     int rc, fd = -1;
+    char devpath[100];
 
     rom = qemu_mallocz(sizeof(*rom));
     rom->name = qemu_strdup(file);
@@ -605,7 +606,12 @@ int rom_add_file(const char *file, const char *fw_dir,
         snprintf(fw_file_name, sizeof(fw_file_name), "%s/%s", rom->fw_dir,
                  basename);
         fw_cfg_add_file(fw_cfg, fw_file_name, rom->data, rom->romsize);
+        snprintf(devpath, sizeof(devpath), "/rom@%s", fw_file_name);
+    } else {
+        snprintf(devpath, sizeof(devpath), "/rom@" TARGET_FMT_plx, addr);
     }
+
+    add_boot_device_path(bootindex, NULL, devpath);
     return 0;
 
 err:
@@ -635,12 +641,12 @@ int rom_add_blob(const char *name, const void *blob, size_t len,
 
 int rom_add_vga(const char *file)
 {
-    return rom_add_file(file, "vgaroms", 0);
+    return rom_add_file(file, "vgaroms", 0, -1);
 }
 
-int rom_add_option(const char *file)
+int rom_add_option(const char *file, int32_t bootindex)
 {
-    return rom_add_file(file, "genroms", 0);
+    return rom_add_file(file, "genroms", 0, bootindex);
 }
 
 static void rom_reset(void *unused)
hw/loader.h +4 -4
index 1f82fc5..fc6bdff 100644
@@ -22,7 +22,7 @@ void pstrcpy_targphys(const char *name,
 
 
 int rom_add_file(const char *file, const char *fw_dir,
-                 target_phys_addr_t addr);
+                 target_phys_addr_t addr, int32_t bootindex);
 int rom_add_blob(const char *name, const void *blob, size_t len,
                  target_phys_addr_t addr);
 int rom_load_all(void);
@@ -31,8 +31,8 @@ int rom_copy(uint8_t *dest, target_phys_addr_t addr, size_t size);
 void *rom_ptr(target_phys_addr_t addr);
 void do_info_roms(Monitor *mon);
 
-#define rom_add_file_fixed(_f, _a)              \
-    rom_add_file(_f, NULL, _a)
+#define rom_add_file_fixed(_f, _a, _i)          \
+    rom_add_file(_f, NULL, _a, _i)
 #define rom_add_blob_fixed(_f, _b, _l, _a)      \
     rom_add_blob(_f, _b, _l, _a)
 
@@ -43,6 +43,6 @@ void do_info_roms(Monitor *mon);
 #define PC_ROM_SIZE        (PC_ROM_MAX - PC_ROM_MIN_VGA)
 
 int rom_add_vga(const char *file);
-int rom_add_option(const char *file);
+int rom_add_option(const char *file, int32_t bootindex);
 
 #endif
hw/multiboot.c +2 -1
index f9097a2..b438019 100644
@@ -325,7 +325,8 @@ int load_multiboot(void *fw_cfg,
     fw_cfg_add_bytes(fw_cfg, FW_CFG_INITRD_DATA, mb_bootinfo_data,
                      sizeof(bootinfo));
 
-    option_rom[nb_option_roms] = "multiboot.bin";
+    option_rom[nb_option_roms].name = "multiboot.bin";
+    option_rom[nb_option_roms].bootindex = 0;
     nb_option_roms++;
 
     return 1; /* yes, we are multiboot */
hw/ne2000.c +1 -1
index f4bbac2..67e0cb0 100644
@@ -742,7 +742,7 @@ static int pci_ne2000_init(PCIDevice *pci_dev)
     if (!pci_dev->qdev.hotplugged) {
         static int loaded = 0;
         if (!loaded) {
-            rom_add_option("pxe-ne2k_pci.bin");
+            rom_add_option("pxe-ne2k_pci.bin", -1);
             loaded = 1;
         }
     }
hw/nseries.c +1 -1
index 04a028d..39780ef 100644
@@ -1326,7 +1326,7 @@ static void n8x0_init(ram_addr_t ram_size, const char *boot_device,
         qemu_register_reset(n8x0_boot_init, s);
     }
 
-    if (option_rom[0] && (boot_device[0] == 'n' || !kernel_filename)) {
+    if (option_rom[0].name && (boot_device[0] == 'n' || !kernel_filename)) {
         int rom_size;
         uint8_t nolo_tags[0x10000];
         /* No, wait, better start at the ROM.  */
hw/pc.c +4 -3
index 69b13bf..5111a76 100644
@@ -733,7 +733,8 @@ static void load_linux(void *fw_cfg,
     fw_cfg_add_i32(fw_cfg, FW_CFG_SETUP_SIZE, setup_size);
     fw_cfg_add_bytes(fw_cfg, FW_CFG_SETUP_DATA, setup, setup_size);
 
-    option_rom[nb_option_roms] = "linuxboot.bin";
+    option_rom[nb_option_roms].name = "linuxboot.bin";
+    option_rom[nb_option_roms].bootindex = 0;
     nb_option_roms++;
 }
 
@@ -937,7 +938,7 @@ void pc_memory_init(ram_addr_t ram_size,
         goto bios_error;
     }
     bios_offset = qemu_ram_alloc(NULL, "pc.bios", bios_size);
-    ret = rom_add_file_fixed(bios_name, (uint32_t)(-bios_size));
+    ret = rom_add_file_fixed(bios_name, (uint32_t)(-bios_size), -1);
     if (ret != 0) {
     bios_error:
         fprintf(stderr, "qemu: could not load PC BIOS '%s'\n", bios_name);
@@ -969,7 +970,7 @@ void pc_memory_init(ram_addr_t ram_size,
     }
 
     for (i = 0; i < nb_option_roms; i++) {
-        rom_add_option(option_rom[i]);
+        rom_add_option(option_rom[i].name, option_rom[i].bootindex);
     }
 }
 
hw/pci.c +1 -1
index 114b435..49e9298 100644
@@ -1701,7 +1701,7 @@ static int pci_add_option_rom(PCIDevice *pdev)
         if (class == 0x0300) {
             rom_add_vga(pdev->romfile);
         } else {
-            rom_add_option(pdev->romfile);
+            rom_add_option(pdev->romfile, -1);
         }
         return 0;
     }
hw/pcnet.c +1 -1
index a988f4f..38e6591 100644
@@ -2014,7 +2014,7 @@ static int pci_pcnet_init(PCIDevice *pci_dev)
     if (!pci_dev->qdev.hotplugged) {
         static int loaded = 0;
         if (!loaded) {
-            rom_add_option("pxe-pcnet.bin");
+            rom_add_option("pxe-pcnet.bin", -1);
             loaded = 1;
         }
     }
qemu-config.c +17 -0
index 52f18be..965fa46 100644
@@ -429,6 +429,22 @@ QemuOptsList qemu_spice_opts = {
     },
 };
 
+QemuOptsList qemu_option_rom_opts = {
+    .name = "option-rom",
+    .implied_opt_name = "romfile",
+    .head = QTAILQ_HEAD_INITIALIZER(qemu_option_rom_opts.head),
+    .desc = {
+        {
+            .name = "bootindex",
+            .type = QEMU_OPT_NUMBER,
+        }, {
+            .name = "romfile",
+            .type = QEMU_OPT_STRING,
+        },
+        { /* end if list */ }
+    },
+};
+
 static QemuOptsList *vm_config_groups[32] = {
     &qemu_drive_opts,
     &qemu_chardev_opts,
@@ -442,6 +458,7 @@ static QemuOptsList *vm_config_groups[32] = {
 #ifdef CONFIG_SIMPLE_TRACE
     &qemu_trace_opts,
 #endif
+    &qemu_option_rom_opts,
     NULL,
 };
 
sysemu.h +5 -1
index fe9a582..48f8eee 100644
@@ -139,7 +139,11 @@ extern uint64_t node_mem[MAX_NODES];
 extern uint64_t node_cpumask[MAX_NODES];
 
 #define MAX_OPTION_ROMS 16
-extern const char *option_rom[MAX_OPTION_ROMS];
+typedef struct QEMUOptionRom {
+    const char *name;
+    int32_t bootindex;
+} QEMUOptionRom;
+extern QEMUOptionRom option_rom[MAX_OPTION_ROMS];
 extern int nb_option_roms;
 
 #define MAX_PROM_ENVS 128
vl.c +9 -2
index 4c159f0..e8ada75 100644
@@ -217,7 +217,7 @@ int cursor_hide = 1;
 int graphic_rotate = 0;
 uint8_t irq0override = 1;
 const char *watchdog;
-const char *option_rom[MAX_OPTION_ROMS];
+QEMUOptionRom option_rom[MAX_OPTION_ROMS];
 int nb_option_roms;
 int semihosting_enabled = 0;
 int old_param = 0;
@@ -2516,7 +2516,14 @@ int main(int argc, char **argv, char **envp)
 		    fprintf(stderr, "Too many option ROMs\n");
 		    exit(1);
 		}
-		option_rom[nb_option_roms] = optarg;
+                opts = qemu_opts_parse(qemu_find_opts("option-rom"), optarg, 1);
+                option_rom[nb_option_roms].name = qemu_opt_get(opts, "romfile");
+                option_rom[nb_option_roms].bootindex =
+                    qemu_opt_get_number(opts, "bootindex", -1);
+                if (!option_rom[nb_option_roms].name) {
+                    fprintf(stderr, "Option ROM file is not specified\n");
+                    exit(1);
+                }
 		nb_option_roms++;
 		break;
             case QEMU_OPTION_semihosting:
Signature · 1 line
1.7.1
#16 lore raw

Re: [PATCHv4 13/15] Add bootindex for option roms.

Quoted 1 line · Gleb Natapov

Extend -option-rom command to have additional parameter ,bootindex=.

This patch is broken:
CC arm-softmmu/palm.o
/src/qemu/hw/palm.c: In function 'palmte_init':
/src/qemu/hw/palm.c:237: error: incompatible type for argument 1 of 'get_image_size'
/src/qemu/hw/palm.c:245: error: incompatible type for argument 1 of 'load_image_targphys'
cc1: warnings being treated as errors /src/qemu/hw/palm.c:250: error: format '%s' expects type 'char *', but argument 4 has type 'QEMUOptionRom'
CC arm-softmmu/nseries.o
/src/qemu/hw/nseries.c: In function 'n8x0_init':
/src/qemu/hw/nseries.c:1346: error: incompatible type for argument 1 of 'load_image_targphys'

#17 lore raw

Re: [PATCHv4 13/15] Add bootindex for option roms.

Quoted 17 lines · Blue Swirl
Quoted 1 line · Gleb Natapov

Extend -option-rom command to have additional parameter ,bootindex=.

This patch is broken:
CC arm-softmmu/palm.o
/src/qemu/hw/palm.c: In function 'palmte_init':
/src/qemu/hw/palm.c:237: error: incompatible type for argument 1 of 'get_image_size'
/src/qemu/hw/palm.c:245: error: incompatible type for argument 1 of 'load_image_targphys'
cc1: warnings being treated as errors /src/qemu/hw/palm.c:250: error: format '%s' expects type 'char *', but argument 4 has type 'QEMUOptionRom'
CC arm-softmmu/nseries.o
/src/qemu/hw/nseries.c: In function 'n8x0_init':
/src/qemu/hw/nseries.c:1346: error: incompatible type for argument 1 of 'load_image_targphys'

Fixed in new version.

--
Gleb.

#18 lore raw

PATCH v4 14/15 Add notifier that will be called when machine is fully created.

In reply to Gleb Natapov

Action that depends on fully initialized device model should register with this notifier chain.

---
 sysemu.h |    2 ++
 vl.c     |   15 +++++++++++++++
 2 files changed, 17 insertions(+), 0 deletions(-)
sysemu.h +2 -0
index 48f8eee..c42f33a 100644
@@ -60,6 +60,8 @@ void qemu_system_reset(void);
 void qemu_add_exit_notifier(Notifier *notify);
 void qemu_remove_exit_notifier(Notifier *notify);
 
+void qemu_add_machine_init_done_notifier(Notifier *notify);
+
 void do_savevm(Monitor *mon, const QDict *qdict);
 int load_vmstate(const char *name);
 void do_delvm(Monitor *mon, const QDict *qdict);
vl.c +15 -0
index e8ada75..918d988 100644
@@ -253,6 +253,9 @@ static void *boot_set_opaque;
 static NotifierList exit_notifiers =
     NOTIFIER_LIST_INITIALIZER(exit_notifiers);
 
+static NotifierList machine_init_done_notifiers =
+    NOTIFIER_LIST_INITIALIZER(machine_init_done_notifiers);
+
 int kvm_allowed = 0;
 uint32_t xen_domid;
 enum xen_mode xen_mode = XEN_EMULATE;
@@ -1778,6 +1781,16 @@ static void qemu_run_exit_notifiers(void)
     notifier_list_notify(&exit_notifiers);
 }
 
+void qemu_add_machine_init_done_notifier(Notifier *notify)
+{
+    notifier_list_add(&machine_init_done_notifiers, notify);
+}
+
+static void qemu_run_machine_init_done_notifiers(void)
+{
+    notifier_list_notify(&machine_init_done_notifiers);
+}
+
 static const QEMUOption *lookup_opt(int argc, char **argv,
                                     const char **poptarg, int *poptind)
 {
@@ -3023,6 +3036,8 @@ int main(int argc, char **argv, char **envp)
         exit(1);
     }
 
+    qemu_run_machine_init_done_notifiers();
+
     qemu_system_reset();
     if (loadvm) {
         if (load_vmstate(loadvm) < 0) {
Signature · 1 line
1.7.1
#19 lore raw

PATCH v4 15/15 Pass boot device list to firmware.

In reply to Gleb Natapov
---
 hw/fw_cfg.c |   14 ++++++++++++++
 hw/fw_cfg.h |    4 +++-
 sysemu.h    |    1 +
 vl.c        |   51 +++++++++++++++++++++++++++++++++++++++++++++++++++
 4 files changed, 69 insertions(+), 1 deletions(-)
hw/fw_cfg.c +14 -0
index 7b9434f..f6a67db 100644
@@ -53,6 +53,7 @@ struct FWCfgState {
     FWCfgFiles *files;
     uint16_t cur_entry;
     uint32_t cur_offset;
+    Notifier machine_ready;
 };
 
 static void fw_cfg_write(FWCfgState *s, uint8_t value)
@@ -315,6 +316,15 @@ int fw_cfg_add_file(FWCfgState *s,  const char *filename, uint8_t *data,
     return 1;
 }
 
+static void fw_cfg_machine_ready(struct Notifier* n)
+{
+    uint32_t len;
+    char *bootindex = get_boot_devices_list(&len);
+
+    fw_cfg_add_bytes(container_of(n, FWCfgState, machine_ready),
+                     FW_CFG_BOOTINDEX, (uint8_t*)bootindex, len);
+}
+
 FWCfgState *fw_cfg_init(uint32_t ctl_port, uint32_t data_port,
                         target_phys_addr_t ctl_addr, target_phys_addr_t data_addr)
 {
@@ -343,6 +353,10 @@ FWCfgState *fw_cfg_init(uint32_t ctl_port, uint32_t data_port,
     fw_cfg_add_i16(s, FW_CFG_MAX_CPUS, (uint16_t)max_cpus);
     fw_cfg_add_i16(s, FW_CFG_BOOT_MENU, (uint16_t)boot_menu);
 
+
+    s->machine_ready.notify = fw_cfg_machine_ready;
+    qemu_add_machine_init_done_notifier(&s->machine_ready);
+
     return s;
 }
 
hw/fw_cfg.h +3 -1
index 856bf91..4d61410 100644
@@ -30,7 +30,9 @@
 
 #define FW_CFG_FILE_FIRST       0x20
 #define FW_CFG_FILE_SLOTS       0x10
-#define FW_CFG_MAX_ENTRY        (FW_CFG_FILE_FIRST+FW_CFG_FILE_SLOTS)
+#define FW_CFG_FILE_LAST_SLOT   (FW_CFG_FILE_FIRST+FW_CFG_FILE_SLOTS)
+#define FW_CFG_BOOTINDEX        (FW_CFG_FILE_LAST_SLOT + 1)
+#define FW_CFG_MAX_ENTRY        FW_CFG_BOOTINDEX 
 
 #define FW_CFG_WRITE_CHANNEL    0x4000
 #define FW_CFG_ARCH_LOCAL       0x8000
sysemu.h +1 -0
index c42f33a..38a20a3 100644
@@ -196,4 +196,5 @@ void register_devices(void);
 
 void add_boot_device_path(int32_t bootindex, DeviceState *dev,
                           const char *suffix);
+char *get_boot_devices_list(uint32_t *size);
 #endif
vl.c +51 -0

57 lines not on this page · raw message

Signature · 1 line
1.7.1
#20 lore raw

Re: [PATCHv4 15/15] Pass boot device list to firmware.

Quoted 78 lines · Gleb Natapov
---
 hw/fw_cfg.c |   14 ++++++++++++++
 hw/fw_cfg.h |    4 +++-
 sysemu.h    |    1 +
 vl.c        |   51 +++++++++++++++++++++++++++++++++++++++++++++++++++
 4 files changed, 69 insertions(+), 1 deletions(-)
hw/fw_cfg.c +14 -0
index 7b9434f..f6a67db 100644
@@ -53,6 +53,7 @@ struct FWCfgState {
     FWCfgFiles *files;
     uint16_t cur_entry;
     uint32_t cur_offset;
+    Notifier machine_ready;
 };
 
 static void fw_cfg_write(FWCfgState *s, uint8_t value)
@@ -315,6 +316,15 @@ int fw_cfg_add_file(FWCfgState *s,  const char *filename, uint8_t *data,
     return 1;
 }
 
+static void fw_cfg_machine_ready(struct Notifier* n)
+{
+    uint32_t len;
+    char *bootindex = get_boot_devices_list(&len);
+
+    fw_cfg_add_bytes(container_of(n, FWCfgState, machine_ready),
+                     FW_CFG_BOOTINDEX, (uint8_t*)bootindex, len);
+}
+
 FWCfgState *fw_cfg_init(uint32_t ctl_port, uint32_t data_port,
                         target_phys_addr_t ctl_addr, target_phys_addr_t data_addr)
 {
@@ -343,6 +353,10 @@ FWCfgState *fw_cfg_init(uint32_t ctl_port, uint32_t data_port,
     fw_cfg_add_i16(s, FW_CFG_MAX_CPUS, (uint16_t)max_cpus);
     fw_cfg_add_i16(s, FW_CFG_BOOT_MENU, (uint16_t)boot_menu);
 
+
+    s->machine_ready.notify = fw_cfg_machine_ready;
+    qemu_add_machine_init_done_notifier(&s->machine_ready);
+
     return s;
 }
 
hw/fw_cfg.h +3 -1
index 856bf91..4d61410 100644
@@ -30,7 +30,9 @@
 
 #define FW_CFG_FILE_FIRST       0x20
 #define FW_CFG_FILE_SLOTS       0x10
-#define FW_CFG_MAX_ENTRY        (FW_CFG_FILE_FIRST+FW_CFG_FILE_SLOTS)
+#define FW_CFG_FILE_LAST_SLOT   (FW_CFG_FILE_FIRST+FW_CFG_FILE_SLOTS)
+#define FW_CFG_BOOTINDEX        (FW_CFG_FILE_LAST_SLOT + 1)
+#define FW_CFG_MAX_ENTRY        FW_CFG_BOOTINDEX 
 
 #define FW_CFG_WRITE_CHANNEL    0x4000
 #define FW_CFG_ARCH_LOCAL       0x8000
sysemu.h +1 -0
index c42f33a..38a20a3 100644
@@ -196,4 +196,5 @@ void register_devices(void);
 
 void add_boot_device_path(int32_t bootindex, DeviceState *dev,
                           const char *suffix);
+char *get_boot_devices_list(uint32_t *size);
 #endif
vl.c +0 -0
index 918d988..cca1e76 100644

vl.c void add_boot_device_path(int32_t bootindex, DeviceState *dev, · line 742

Quoted 8 lines
vl.c +5 -0
@@ -735,6 +735,57 @@ void add_boot_device_path(int32_t bootindex, DeviceState *dev,
     QTAILQ_INSERT_TAIL(&fw_boot_order, node, link);
 }
 
+/*
+ * This function returns device list as an array in a below format:
+ * +-----+-----+---------------+-----+---------------+--
+ * |  n  |  l1 |   devpath1    |  l2 |  devpath2     | ...
+ * +-----+-----+---------------+-----+---------------+--

No one will ever want > 256 devices? Let's make it 4 byte or something.

vl.c void add_boot_device_path(int32_t bootindex, DeviceState *dev, · line 745

Quoted 3 lines
vl.c +3 -0
+ * where:
+ *   n - a number of devise pathes (one byte)
+ *   l - length of following device path string (one byte)

Might not fit: with pci we can have 256 nested buses. How about simply null-terminating each path?

vl.c void add_boot_device_path(int32_t bootindex, DeviceState *dev, · line 748

Quoted 3 lines
vl.c +3 -0
+ *   devpath - non-null terminated string of length l representing
+ *             one device path
+ */

Document return value + parameters as well?

Quoted 9 lines
vl.c +9 -0
+char *get_boot_devices_list(uint32_t *size)
+{
+    FWBootEntry *i;
+    uint32_t total = 1, c = 0;
+    char *list = qemu_malloc(1);
+
+    QTAILQ_FOREACH(i, &fw_boot_order, link) {
+        char *devpath = NULL, *bootpath;
+        int len;

vl.c void add_boot_device_path(int32_t bootindex, DeviceState *dev, · line 767

Quoted 10 lines
vl.c +10 -0
+
+        if (i->dev) {
+            devpath = qdev_get_fw_dev_path(i->dev);
+            assert(devpath);
+        }
+
+        if (i->suffix && devpath) {
+            bootpath = qemu_malloc(strlen(devpath) + strlen(i->suffix) + 2);
+            sprintf(bootpath, "%s/%s", devpath, i->suffix);
+            qemu_free(devpath);

devpath is allocated with strdup, not qemu_malloc, so I guess it should be freed with free? Alternatively, let's add qemu_strdup
Might be a good idea: fix error handling here and elsewhere.

vl.c void add_boot_device_path(int32_t bootindex, DeviceState *dev, · line 772

Quoted 5 lines
vl.c +5 -0
+        } else if (devpath) {
+            bootpath = devpath;
+        } else {
+            bootpath = strdup(i->suffix);
+        }

assert(bootpath).

vl.c void add_boot_device_path(int32_t bootindex, DeviceState *dev, · line 780

Quoted 8 lines
vl.c +8 -0
+
+        len = strlen(bootpath);
+        list = qemu_realloc(list, total + len + 1);
+        list[total++] = len;
+        memcpy(&list[total], bootpath, len);
+        total += len;
+        c++;
+        qemu_free(bootpath);

Man, is this tricky.

Quoted 13 lines
vl.c +8 -0
+    }
+
+    list[0] = c;
+    *size = total;
+
+    return list;
+}
+
 static void numa_add(const char *optarg)
 {
     char option[128];
Signature · 1 line
1.7.1
#21 lore raw

Re: [PATCHv4 15/15] Pass boot device list to firmware.

Quoted 82 lines · Michael S. Tsirkin
Quoted 81 lines · Gleb Natapov
---
 hw/fw_cfg.c |   14 ++++++++++++++
 hw/fw_cfg.h |    4 +++-
 sysemu.h    |    1 +
 vl.c        |   51 +++++++++++++++++++++++++++++++++++++++++++++++++++
 4 files changed, 69 insertions(+), 1 deletions(-)
hw/fw_cfg.c +14 -0
index 7b9434f..f6a67db 100644
@@ -53,6 +53,7 @@ struct FWCfgState {
     FWCfgFiles *files;
     uint16_t cur_entry;
     uint32_t cur_offset;
+    Notifier machine_ready;
 };
 
 static void fw_cfg_write(FWCfgState *s, uint8_t value)
@@ -315,6 +316,15 @@ int fw_cfg_add_file(FWCfgState *s,  const char *filename, uint8_t *data,
     return 1;
 }
 
+static void fw_cfg_machine_ready(struct Notifier* n)
+{
+    uint32_t len;
+    char *bootindex = get_boot_devices_list(&len);
+
+    fw_cfg_add_bytes(container_of(n, FWCfgState, machine_ready),
+                     FW_CFG_BOOTINDEX, (uint8_t*)bootindex, len);
+}
+
 FWCfgState *fw_cfg_init(uint32_t ctl_port, uint32_t data_port,
                         target_phys_addr_t ctl_addr, target_phys_addr_t data_addr)
 {
@@ -343,6 +353,10 @@ FWCfgState *fw_cfg_init(uint32_t ctl_port, uint32_t data_port,
     fw_cfg_add_i16(s, FW_CFG_MAX_CPUS, (uint16_t)max_cpus);
     fw_cfg_add_i16(s, FW_CFG_BOOT_MENU, (uint16_t)boot_menu);
 
+
+    s->machine_ready.notify = fw_cfg_machine_ready;
+    qemu_add_machine_init_done_notifier(&s->machine_ready);
+
     return s;
 }
 
hw/fw_cfg.h +3 -1
index 856bf91..4d61410 100644
@@ -30,7 +30,9 @@
 
 #define FW_CFG_FILE_FIRST       0x20
 #define FW_CFG_FILE_SLOTS       0x10
-#define FW_CFG_MAX_ENTRY        (FW_CFG_FILE_FIRST+FW_CFG_FILE_SLOTS)
+#define FW_CFG_FILE_LAST_SLOT   (FW_CFG_FILE_FIRST+FW_CFG_FILE_SLOTS)
+#define FW_CFG_BOOTINDEX        (FW_CFG_FILE_LAST_SLOT + 1)
+#define FW_CFG_MAX_ENTRY        FW_CFG_BOOTINDEX 
 
 #define FW_CFG_WRITE_CHANNEL    0x4000
 #define FW_CFG_ARCH_LOCAL       0x8000
sysemu.h +1 -0
index c42f33a..38a20a3 100644
@@ -196,4 +196,5 @@ void register_devices(void);
 
 void add_boot_device_path(int32_t bootindex, DeviceState *dev,
                           const char *suffix);
+char *get_boot_devices_list(uint32_t *size);
 #endif
vl.c +0 -0
index 918d988..cca1e76 100644
@@ -735,6 +735,57 @@ void add_boot_device_path(int32_t bootindex, DeviceState *dev,
     QTAILQ_INSERT_TAIL(&fw_boot_order, node, link);
 }
 
Quoted 8 lines

vl.c void add_boot_device_path(int32_t bootindex, DeviceState *dev, · line 742

Quoted 5 lines
vl.c +5 -0
+/*
+ * This function returns device list as an array in a below format:
+ * +-----+-----+---------------+-----+---------------+--
+ * |  n  |  l1 |   devpath1    |  l2 |  devpath2     | ...
+ * +-----+-----+---------------+-----+---------------+--

No one will ever want > 256 devices? Let's make it 4 byte or something.

More then 256 _boot_ devices. I think this is more then reasonable.

Quoted 5 lines

vl.c void add_boot_device_path(int32_t bootindex, DeviceState *dev, · line 745

Quoted 3 lines
vl.c +3 -0
+ * where:
+ *   n - a number of devise pathes (one byte)
+ *   l - length of following device path string (one byte)

Might not fit: with pci we can have 256 nested buses.

Theoretically. But will it practically happen?

Quoted 2 lines

How about simply null-terminating each path?

Will be harder for Seabios. I can use more then one byte for length, but I tried to avoid endianess handling.

Quoted 6 lines

vl.c void add_boot_device_path(int32_t bootindex, DeviceState *dev, · line 748

Quoted 3 lines
vl.c +3 -0
+ *   devpath - non-null terminated string of length l representing
+ *             one device path
+ */

Document return value + parameters as well?

Return value is documented above :)

Quoted 14 lines
Quoted 14 lines
vl.c +14 -0
+char *get_boot_devices_list(uint32_t *size)
+{
+    FWBootEntry *i;
+    uint32_t total = 1, c = 0;
+    char *list = qemu_malloc(1);
+
+    QTAILQ_FOREACH(i, &fw_boot_order, link) {
+        char *devpath = NULL, *bootpath;
+        int len;
+
+        if (i->dev) {
+            devpath = qdev_get_fw_dev_path(i->dev);
+            assert(devpath);
+        }
Quoted 8 lines

vl.c void add_boot_device_path(int32_t bootindex, DeviceState *dev, · line 767

Quoted 5 lines
vl.c +5 -0
+
+        if (i->suffix && devpath) {
+            bootpath = qemu_malloc(strlen(devpath) + strlen(i->suffix) + 2);
+            sprintf(bootpath, "%s/%s", devpath, i->suffix);
+            qemu_free(devpath);

devpath is allocated with strdup, not qemu_malloc, so I guess it should be freed with free?

There is code that do it like this all over qemu.

Quoted 11 lines

Alternatively, let's add qemu_strdup
Might be a good idea: fix error handling here and elsewhere.

vl.c void add_boot_device_path(int32_t bootindex, DeviceState *dev, · line 772

Quoted 5 lines
vl.c +5 -0
+        } else if (devpath) {
+            bootpath = devpath;
+        } else {
+            bootpath = strdup(i->suffix);
+        }

assert(bootpath).

OK

Quoted 11 lines

vl.c void add_boot_device_path(int32_t bootindex, DeviceState *dev, · line 780

Quoted 8 lines
vl.c +8 -0
+
+        len = strlen(bootpath);
+        list = qemu_realloc(list, total + len + 1);
+        list[total++] = len;
+        memcpy(&list[total], bootpath, len);
+        total += len;
+        c++;
+        qemu_free(bootpath);

Man, is this tricky.

Nah, not at all.

--
Gleb.

#22 lore raw

Re: [PATCHv4 15/15] Pass boot device list to firmware.

Quoted 91 lines · Gleb Natapov
Quoted 82 lines · Michael S. Tsirkin
Quoted 81 lines · Gleb Natapov
---
 hw/fw_cfg.c |   14 ++++++++++++++
 hw/fw_cfg.h |    4 +++-
 sysemu.h    |    1 +
 vl.c        |   51 +++++++++++++++++++++++++++++++++++++++++++++++++++
 4 files changed, 69 insertions(+), 1 deletions(-)
hw/fw_cfg.c +14 -0
index 7b9434f..f6a67db 100644
@@ -53,6 +53,7 @@ struct FWCfgState {
     FWCfgFiles *files;
     uint16_t cur_entry;
     uint32_t cur_offset;
+    Notifier machine_ready;
 };
 
 static void fw_cfg_write(FWCfgState *s, uint8_t value)
@@ -315,6 +316,15 @@ int fw_cfg_add_file(FWCfgState *s,  const char *filename, uint8_t *data,
     return 1;
 }
 
+static void fw_cfg_machine_ready(struct Notifier* n)
+{
+    uint32_t len;
+    char *bootindex = get_boot_devices_list(&len);
+
+    fw_cfg_add_bytes(container_of(n, FWCfgState, machine_ready),
+                     FW_CFG_BOOTINDEX, (uint8_t*)bootindex, len);
+}
+
 FWCfgState *fw_cfg_init(uint32_t ctl_port, uint32_t data_port,
                         target_phys_addr_t ctl_addr, target_phys_addr_t data_addr)
 {
@@ -343,6 +353,10 @@ FWCfgState *fw_cfg_init(uint32_t ctl_port, uint32_t data_port,
     fw_cfg_add_i16(s, FW_CFG_MAX_CPUS, (uint16_t)max_cpus);
     fw_cfg_add_i16(s, FW_CFG_BOOT_MENU, (uint16_t)boot_menu);
 
+
+    s->machine_ready.notify = fw_cfg_machine_ready;
+    qemu_add_machine_init_done_notifier(&s->machine_ready);
+
     return s;
 }
 
hw/fw_cfg.h +3 -1
index 856bf91..4d61410 100644
@@ -30,7 +30,9 @@
 
 #define FW_CFG_FILE_FIRST       0x20
 #define FW_CFG_FILE_SLOTS       0x10
-#define FW_CFG_MAX_ENTRY        (FW_CFG_FILE_FIRST+FW_CFG_FILE_SLOTS)
+#define FW_CFG_FILE_LAST_SLOT   (FW_CFG_FILE_FIRST+FW_CFG_FILE_SLOTS)
+#define FW_CFG_BOOTINDEX        (FW_CFG_FILE_LAST_SLOT + 1)
+#define FW_CFG_MAX_ENTRY        FW_CFG_BOOTINDEX 
 
 #define FW_CFG_WRITE_CHANNEL    0x4000
 #define FW_CFG_ARCH_LOCAL       0x8000
sysemu.h +1 -0
index c42f33a..38a20a3 100644
@@ -196,4 +196,5 @@ void register_devices(void);
 
 void add_boot_device_path(int32_t bootindex, DeviceState *dev,
                           const char *suffix);
+char *get_boot_devices_list(uint32_t *size);
 #endif
vl.c +0 -0
index 918d988..cca1e76 100644
@@ -735,6 +735,57 @@ void add_boot_device_path(int32_t bootindex, DeviceState *dev,
     QTAILQ_INSERT_TAIL(&fw_boot_order, node, link);
 }
 
Quoted 8 lines

vl.c void add_boot_device_path(int32_t bootindex, DeviceState *dev, · line 742

Quoted 5 lines
vl.c +5 -0
+/*
+ * This function returns device list as an array in a below format:
+ * +-----+-----+---------------+-----+---------------+--
+ * |  n  |  l1 |   devpath1    |  l2 |  devpath2     | ...
+ * +-----+-----+---------------+-----+---------------+--

No one will ever want > 256 devices? Let's make it 4 byte or something.

Quoted 8 lines

More then 256 _boot_ devices. I think this is more then reasonable.

Quoted 5 lines

vl.c void add_boot_device_path(int32_t bootindex, DeviceState *dev, · line 745

Quoted 3 lines
vl.c +3 -0
+ * where:
+ *   n - a number of devise pathes (one byte)
+ *   l - length of following device path string (one byte)

Might not fit: with pci we can have 256 nested buses.

Theoretically. But will it practically happen?

Why not? It's easy to specify this on qemu command line. You do nothing to detect this and gracefully fail either, do you?

Signature · 1 line
MST
#23 lore raw

Re: [PATCHv4 15/15] Pass boot device list to firmware.

In reply to Gleb Natapov
Quoted 9 lines · Gleb Natapov
Quoted 9 lines
Quoted 8 lines
vl.c +8 -0
+
+        len = strlen(bootpath);
+        list = qemu_realloc(list, total + len + 1);
+        list[total++] = len;
+        memcpy(&list[total], bootpath, len);
+        total += len;
+        c++;
+        qemu_free(bootpath);
Quoted 3 lines
Quoted 2 lines

Man, is this tricky.

Nah, not at all.

I think it will be easier if we don't try to do this in one pass.
1. pass: calculate total length and # of devices
2. allocate
2. pass fill in

Quoted 2 lines

--
Gleb.

#24 lore raw

Re: [PATCHv4 15/15] Pass boot device list to firmware.

In reply to Gleb Natapov
Quoted 22 lines · Gleb Natapov

---
 hw/fw_cfg.c |   14 ++++++++++++++
 hw/fw_cfg.h |    4 +++-
 sysemu.h    |    1 +
 vl.c        |   51 +++++++++++++++++++++++++++++++++++++++++++++++++++

 4 files changed, 69 insertions(+), 1 deletions(-)
hw/fw_cfg.c +1 -0
index 7b9434f..f6a67db 100644
@@ -53,6 +53,7 @@ struct FWCfgState {
     FWCfgFiles *files;
     uint16_t cur_entry;
     uint32_t cur_offset;
+    Notifier machine_ready;
  };
 
  static void fw_cfg_write(FWCfgState *s, uint8_t value)

hw/fw_cfg.c int fw_cfg_add_file(FWCfgState *s,  const char *filename, uint8_t *data, · line 325

Quoted 10 lines
hw/fw_cfg.c +7 -0
@@ -315,6 +316,15 @@ int fw_cfg_add_file(FWCfgState *s,  const char *filename, uint8_t *data,
     return 1;
  }
 
+static void fw_cfg_machine_ready(struct Notifier* n)
+{
+    uint32_t len;
+    char *bootindex = get_boot_devices_list(&len);
+
+    fw_cfg_add_bytes(container_of(n, FWCfgState, machine_ready),
+                     FW_CFG_BOOTINDEX, (uint8_t*)bootindex, len);

I started to implement this to OpenBIOS but I noticed a small issue. First the first byte must be read to determine length. Then the read routine will be called again to read the correct amount of bytes. This would work, but since there is no shortage of IDs, I'd prefer a system where one ID is used to query the length and another ID is used to read the data, without the length byte. This is similar how command line, initrd etc. are handled.

This would have the advantage that since fw_cfg uses little endian format, the length value would easily scale to for example 64 bits to support terabytes of boot device lists. ;-)

#25 lore raw

Re: [PATCHv4 15/15] Pass boot device list to firmware.

Quoted 42 lines · Blue Swirl
Quoted 22 lines · Gleb Natapov

---
 hw/fw_cfg.c |   14 ++++++++++++++
 hw/fw_cfg.h |    4 +++-
 sysemu.h    |    1 +
 vl.c        |   51 +++++++++++++++++++++++++++++++++++++++++++++++++++

 4 files changed, 69 insertions(+), 1 deletions(-)
hw/fw_cfg.c +1 -0
index 7b9434f..f6a67db 100644
@@ -53,6 +53,7 @@ struct FWCfgState {
     FWCfgFiles *files;
     uint16_t cur_entry;
     uint32_t cur_offset;
+    Notifier machine_ready;
  };
 
  static void fw_cfg_write(FWCfgState *s, uint8_t value)

hw/fw_cfg.c int fw_cfg_add_file(FWCfgState *s,  const char *filename, uint8_t *data, · line 325

Quoted 10 lines
hw/fw_cfg.c +7 -0
@@ -315,6 +316,15 @@ int fw_cfg_add_file(FWCfgState *s,  const char *filename, uint8_t *data,
     return 1;
  }
 
+static void fw_cfg_machine_ready(struct Notifier* n)
+{
+    uint32_t len;
+    char *bootindex = get_boot_devices_list(&len);
+
+    fw_cfg_add_bytes(container_of(n, FWCfgState, machine_ready),
+                     FW_CFG_BOOTINDEX, (uint8_t*)bootindex, len);

I started to implement this to OpenBIOS but I noticed a small issue. First the first byte must be read to determine length. Then the read routine will be called again to read the correct amount of bytes. This would work, but since there is no shortage of IDs, I'd prefer a system where one ID is used to query the length and another ID is used to read the data, without the length byte. This is similar how command line, initrd etc. are handled.

Quoted 3 lines

This would have the advantage that since fw_cfg uses little endian format, the length value would easily scale to for example 64 bits to support terabytes of boot device lists. ;-)

Yea. Let's just print # of devices as a property, in ASCII. No endian-ness, no nothing.
Also - can we just NULL-terminate each ID?

#26 lore raw

Re: [PATCHv4 15/15] Pass boot device list to firmware.

Quoted 43 lines · Michael S. Tsirkin
Quoted 42 lines · Blue Swirl
Quoted 22 lines · Gleb Natapov

---
 hw/fw_cfg.c |   14 ++++++++++++++
 hw/fw_cfg.h |    4 +++-
 sysemu.h    |    1 +
 vl.c        |   51 +++++++++++++++++++++++++++++++++++++++++++++++++++

 4 files changed, 69 insertions(+), 1 deletions(-)
hw/fw_cfg.c +1 -0
index 7b9434f..f6a67db 100644
@@ -53,6 +53,7 @@ struct FWCfgState {
     FWCfgFiles *files;
     uint16_t cur_entry;
     uint32_t cur_offset;
+    Notifier machine_ready;
  };
 
  static void fw_cfg_write(FWCfgState *s, uint8_t value)

hw/fw_cfg.c int fw_cfg_add_file(FWCfgState *s,  const char *filename, uint8_t *data, · line 325

Quoted 10 lines
hw/fw_cfg.c +7 -0
@@ -315,6 +316,15 @@ int fw_cfg_add_file(FWCfgState *s,  const char *filename, uint8_t *data,
     return 1;
  }
 
+static void fw_cfg_machine_ready(struct Notifier* n)
+{
+    uint32_t len;
+    char *bootindex = get_boot_devices_list(&len);
+
+    fw_cfg_add_bytes(container_of(n, FWCfgState, machine_ready),
+                     FW_CFG_BOOTINDEX, (uint8_t*)bootindex, len);

I started to implement this to OpenBIOS but I noticed a small issue. First the first byte must be read to determine length. Then the read routine will be called again to read the correct amount of bytes. This would work, but since there is no shortage of IDs, I'd prefer a system where one ID is used to query the length and another ID is used to read the data, without the length byte. This is similar how command line, initrd etc. are handled.

Quoted 7 lines
Quoted 3 lines

This would have the advantage that since fw_cfg uses little endian format, the length value would easily scale to for example 64 bits to support terabytes of boot device lists. ;-)

Yea. Let's just print # of devices as a property, in ASCII. No endian-ness, no nothing.
Also - can we just NULL-terminate each ID?

No, we should use LE numbers like other IDs. To be more specific, this is what I meant (instead of FW_CFG_BOOTINDEX):
FW_CFG_BOOTINDEX_LEN: get LE integer length of the boot device data.
FW_CFG_BOOTINDEX_DATA: get the boot device data as NUL terminated C strings, all strings back-to-back. The reader can determine number of strings.

#27 lore raw

Re: [PATCHv4 15/15] Pass boot device list to firmware.

Quoted 52 lines · Blue Swirl
Quoted 43 lines · Michael S. Tsirkin
Quoted 42 lines · Blue Swirl
Quoted 22 lines · Gleb Natapov

---
 hw/fw_cfg.c |   14 ++++++++++++++
 hw/fw_cfg.h |    4 +++-
 sysemu.h    |    1 +
 vl.c        |   51 +++++++++++++++++++++++++++++++++++++++++++++++++++

 4 files changed, 69 insertions(+), 1 deletions(-)
hw/fw_cfg.c +1 -0
index 7b9434f..f6a67db 100644
@@ -53,6 +53,7 @@ struct FWCfgState {
     FWCfgFiles *files;
     uint16_t cur_entry;
     uint32_t cur_offset;
+    Notifier machine_ready;
  };
 
  static void fw_cfg_write(FWCfgState *s, uint8_t value)

hw/fw_cfg.c int fw_cfg_add_file(FWCfgState *s,  const char *filename, uint8_t *data, · line 325

Quoted 10 lines
hw/fw_cfg.c +7 -0
@@ -315,6 +316,15 @@ int fw_cfg_add_file(FWCfgState *s,  const char *filename, uint8_t *data,
     return 1;
  }
 
+static void fw_cfg_machine_ready(struct Notifier* n)
+{
+    uint32_t len;
+    char *bootindex = get_boot_devices_list(&len);
+
+    fw_cfg_add_bytes(container_of(n, FWCfgState, machine_ready),
+                     FW_CFG_BOOTINDEX, (uint8_t*)bootindex, len);

I started to implement this to OpenBIOS but I noticed a small issue. First the first byte must be read to determine length. Then the read routine will be called again to read the correct amount of bytes. This would work, but since there is no shortage of IDs, I'd prefer a system where one ID is used to query the length and another ID is used to read the data, without the length byte. This is similar how command line, initrd etc. are handled.

Quoted 7 lines
Quoted 3 lines

This would have the advantage that since fw_cfg uses little endian format, the length value would easily scale to for example 64 bits to support terabytes of boot device lists. ;-)

Yea. Let's just print # of devices as a property, in ASCII. No endian-ness, no nothing.
Also - can we just NULL-terminate each ID?

Quoted 6 lines

No, we should use LE numbers like other IDs. To be more specific, this is what I meant (instead of FW_CFG_BOOTINDEX):
FW_CFG_BOOTINDEX_LEN: get LE integer length of the boot device data.
FW_CFG_BOOTINDEX_DATA: get the boot device data as NUL terminated C strings, all strings back-to-back. The reader can determine number of strings.

Sounds good.

#28 lore raw

Re: [PATCHv4 15/15] Pass boot device list to firmware.

In reply to Gleb Natapov
Quoted 53 lines · Gleb Natapov

---
 hw/fw_cfg.c |   14 ++++++++++++++
 hw/fw_cfg.h |    4 +++-
 sysemu.h    |    1 +
 vl.c        |   51 +++++++++++++++++++++++++++++++++++++++++++++++++++

 4 files changed, 69 insertions(+), 1 deletions(-)
hw/fw_cfg.c +14 -0
index 7b9434f..f6a67db 100644
@@ -53,6 +53,7 @@ struct FWCfgState {
     FWCfgFiles *files;
     uint16_t cur_entry;
     uint32_t cur_offset;
+    Notifier machine_ready;
  };
 
  static void fw_cfg_write(FWCfgState *s, uint8_t value)
@@ -315,6 +316,15 @@ int fw_cfg_add_file(FWCfgState *s,  const char *filename, uint8_t *data,
     return 1;
  }
 
+static void fw_cfg_machine_ready(struct Notifier* n)
+{
+    uint32_t len;
+    char *bootindex = get_boot_devices_list(&len);
+
+    fw_cfg_add_bytes(container_of(n, FWCfgState, machine_ready),
+                     FW_CFG_BOOTINDEX, (uint8_t*)bootindex, len);
+}
+
  FWCfgState *fw_cfg_init(uint32_t ctl_port, uint32_t data_port,
                         target_phys_addr_t ctl_addr, target_phys_addr_t data_addr)
  {
@@ -343,6 +353,10 @@ FWCfgState *fw_cfg_init(uint32_t ctl_port, uint32_t data_port,
     fw_cfg_add_i16(s, FW_CFG_MAX_CPUS, (uint16_t)max_cpus);
     fw_cfg_add_i16(s, FW_CFG_BOOT_MENU, (uint16_t)boot_menu);
 
+
+    s->machine_ready.notify = fw_cfg_machine_ready;
+    qemu_add_machine_init_done_notifier(&s->machine_ready);
+
     return s;
  }
 
hw/fw_cfg.h +0 -0
index 856bf91..4d61410 100644

hw/fw_cfg.h line 35

Quoted 7 lines
hw/fw_cfg.h +3 -1
@@ -30,7 +30,9 @@
 
  #define FW_CFG_FILE_FIRST       0x20
  #define FW_CFG_FILE_SLOTS       0x10
-#define FW_CFG_MAX_ENTRY        (FW_CFG_FILE_FIRST+FW_CFG_FILE_SLOTS)
+#define FW_CFG_FILE_LAST_SLOT   (FW_CFG_FILE_FIRST+FW_CFG_FILE_SLOTS)
+#define FW_CFG_BOOTINDEX        (FW_CFG_FILE_LAST_SLOT + 1)
+#define FW_CFG_MAX_ENTRY        FW_CFG_BOOTINDEX

This should be
#define FW_CFG_MAX_ENTRY (FW_CFG_BOOTINDEX + 1)
because the check is like this:

    if ((key & FW_CFG_ENTRY_MASK) >= FW_CFG_MAX_ENTRY) {
        s->cur_entry = FW_CFG_INVALID;

With that change, I got the bootindex passed to OpenBIOS:
OpenBIOS for Sparc64
Configuration device id QEMU version 1 machine id 0 kernel cmdline
CPUs: 1 x SUNW,UltraSPARC-IIi
UUID: 00000000-0000-0000-0000-000000000000 bootindex num_strings 1
bootindex /pbm@000001fe00000000/ide@5/drive@1/disk@0

The device path does not match exactly, but it's close:
/pci@1fe,0/pci-ata@5/ide1@600/disk@0

#29 lore raw

Re: [PATCHv4 15/15] Pass boot device list to firmware.

Quoted 65 lines · Blue Swirl
Quoted 53 lines · Gleb Natapov

---
 hw/fw_cfg.c |   14 ++++++++++++++
 hw/fw_cfg.h |    4 +++-
 sysemu.h    |    1 +
 vl.c        |   51 +++++++++++++++++++++++++++++++++++++++++++++++++++

 4 files changed, 69 insertions(+), 1 deletions(-)
hw/fw_cfg.c +14 -0
index 7b9434f..f6a67db 100644
@@ -53,6 +53,7 @@ struct FWCfgState {
     FWCfgFiles *files;
     uint16_t cur_entry;
     uint32_t cur_offset;
+    Notifier machine_ready;
  };
 
  static void fw_cfg_write(FWCfgState *s, uint8_t value)
@@ -315,6 +316,15 @@ int fw_cfg_add_file(FWCfgState *s,  const char *filename, uint8_t *data,
     return 1;
  }
 
+static void fw_cfg_machine_ready(struct Notifier* n)
+{
+    uint32_t len;
+    char *bootindex = get_boot_devices_list(&len);
+
+    fw_cfg_add_bytes(container_of(n, FWCfgState, machine_ready),
+                     FW_CFG_BOOTINDEX, (uint8_t*)bootindex, len);
+}
+
  FWCfgState *fw_cfg_init(uint32_t ctl_port, uint32_t data_port,
                         target_phys_addr_t ctl_addr, target_phys_addr_t data_addr)
  {
@@ -343,6 +353,10 @@ FWCfgState *fw_cfg_init(uint32_t ctl_port, uint32_t data_port,
     fw_cfg_add_i16(s, FW_CFG_MAX_CPUS, (uint16_t)max_cpus);
     fw_cfg_add_i16(s, FW_CFG_BOOT_MENU, (uint16_t)boot_menu);
 
+
+    s->machine_ready.notify = fw_cfg_machine_ready;
+    qemu_add_machine_init_done_notifier(&s->machine_ready);
+
     return s;
  }
 
hw/fw_cfg.h +0 -0
index 856bf91..4d61410 100644

hw/fw_cfg.h line 35

Quoted 7 lines
hw/fw_cfg.h +3 -1
@@ -30,7 +30,9 @@
 
  #define FW_CFG_FILE_FIRST       0x20
  #define FW_CFG_FILE_SLOTS       0x10
-#define FW_CFG_MAX_ENTRY        (FW_CFG_FILE_FIRST+FW_CFG_FILE_SLOTS)
+#define FW_CFG_FILE_LAST_SLOT   (FW_CFG_FILE_FIRST+FW_CFG_FILE_SLOTS)
+#define FW_CFG_BOOTINDEX        (FW_CFG_FILE_LAST_SLOT + 1)
+#define FW_CFG_MAX_ENTRY        FW_CFG_BOOTINDEX

This should be
#define FW_CFG_MAX_ENTRY (FW_CFG_BOOTINDEX + 1)
because the check is like this:

Quoted 3 lines
    if ((key & FW_CFG_ENTRY_MASK) >= FW_CFG_MAX_ENTRY) {
        s->cur_entry = FW_CFG_INVALID;

Yeah, will fix.

Quoted 11 lines

With that change, I got the bootindex passed to OpenBIOS:
OpenBIOS for Sparc64
Configuration device id QEMU version 1 machine id 0 kernel cmdline
CPUs: 1 x SUNW,UltraSPARC-IIi
UUID: 00000000-0000-0000-0000-000000000000 bootindex num_strings 1
bootindex /pbm@000001fe00000000/ide@5/drive@1/disk@0

The device path does not match exactly, but it's close:
/pci@1fe,0/pci-ata@5/ide1@600/disk@0

pbm->pci should be solvable by the patch at the end. Were in the spec it is allowed to abbreviate 1fe00000000 as 1fe,0? Spec allows to drop starting zeroes but TARGET_FMT_plx definition in targphys.h has 0 after %. I can define another one without leading zeroes. Can you suggest a name? TARGET_FMT_lx is poisoned. As of ATA there is no open firmware binding spec for ATA, so everyone does what he pleases. I based my implementation on what open firmware showing when running on qemu x86. "pci-ata" should be "ide" according to PCI binding spec :)

hw/apb_pci.c +1 -0
index c619112..643aa49 100644
@@ -453,6 +453,7 @@ static PCIDeviceInfo pbm_pci_host_info = {
 
 static SysBusDeviceInfo pbm_host_info = {
     .qdev.name = "pbm",
+    .qdev.fw_name = "pci",
     .qdev.size = sizeof(APBState),
     .qdev.reset = pci_pbm_reset,
     .init = pci_pbm_init_device,

--
Gleb.

#30 lore raw

Re: [PATCHv4 15/15] Pass boot device list to firmware.

2010/11/15 Gleb Natapov <[email protected]>:

Quoted 80 lines · Blue Swirl
Quoted 55 lines · Gleb Natapov
Quoted 54 lines

---
 hw/fw_cfg.c |   14 ++++++++++++++
 hw/fw_cfg.h |    4 +++-
 sysemu.h    |    1 +
 vl.c        |   51 +++++++++++++++++++++++++++++++++++++++++++++++++++

 4 files changed, 69 insertions(+), 1 deletions(-)
hw/fw_cfg.c +14 -0
index 7b9434f..f6a67db 100644
@@ -53,6 +53,7 @@ struct FWCfgState {
     FWCfgFiles *files;
     uint16_t cur_entry;
     uint32_t cur_offset;
+    Notifier machine_ready;
  };
 
  static void fw_cfg_write(FWCfgState *s, uint8_t value)
@@ -315,6 +316,15 @@ int fw_cfg_add_file(FWCfgState *s,  const char *filename, uint8_t *data,
     return 1;
  }
 
+static void fw_cfg_machine_ready(struct Notifier* n)
+{
+    uint32_t len;
+    char *bootindex = get_boot_devices_list(&len);
+
+    fw_cfg_add_bytes(container_of(n, FWCfgState, machine_ready),
+                     FW_CFG_BOOTINDEX, (uint8_t*)bootindex, len);
+}
+
  FWCfgState *fw_cfg_init(uint32_t ctl_port, uint32_t data_port,
                         target_phys_addr_t ctl_addr, target_phys_addr_t data_addr)
  {
@@ -343,6 +353,10 @@ FWCfgState *fw_cfg_init(uint32_t ctl_port, uint32_t data_port,
     fw_cfg_add_i16(s, FW_CFG_MAX_CPUS, (uint16_t)max_cpus);
     fw_cfg_add_i16(s, FW_CFG_BOOT_MENU, (uint16_t)boot_menu);
 
+
+    s->machine_ready.notify = fw_cfg_machine_ready;
+    qemu_add_machine_init_done_notifier(&s->machine_ready);
+
     return s;
  }
 
hw/fw_cfg.h +0 -0
index 856bf91..4d61410 100644
@@ -30,7 +30,9 @@
 
Quoted 13 lines

hw/fw_cfg.h line 35

Quoted 6 lines
hw/fw_cfg.h +3 -1
  #define FW_CFG_FILE_FIRST       0x20
  #define FW_CFG_FILE_SLOTS       0x10
-#define FW_CFG_MAX_ENTRY        (FW_CFG_FILE_FIRST+FW_CFG_FILE_SLOTS)
+#define FW_CFG_FILE_LAST_SLOT   (FW_CFG_FILE_FIRST+FW_CFG_FILE_SLOTS)
+#define FW_CFG_BOOTINDEX        (FW_CFG_FILE_LAST_SLOT + 1)
+#define FW_CFG_MAX_ENTRY        FW_CFG_BOOTINDEX

This should be
#define FW_CFG_MAX_ENTRY        (FW_CFG_BOOTINDEX + 1) because the check is like this:
    if ((key & FW_CFG_ENTRY_MASK) >= FW_CFG_MAX_ENTRY) { s->cur_entry = FW_CFG_INVALID;

Yeah, will fix.

Quoted 9 lines

With that change, I got the bootindex passed to OpenBIOS:
OpenBIOS for Sparc64
Configuration device id QEMU version 1 machine id 0 kernel cmdline
CPUs: 1 x SUNW,UltraSPARC-IIi
UUID: 00000000-0000-0000-0000-000000000000 bootindex num_strings 1
bootindex /pbm@000001fe00000000/ide@5/drive@1/disk@0

Quoted 8 lines
Quoted 2 lines

The device path does not match exactly, but it's close:
/pci@1fe,0/pci-ata@5/ide1@600/disk@0

pbm->pci should be solvable by the patch at the end. Were in the spec it is allowed to abbreviate 1fe00000000 as 1fe,0? Spec allows to drop starting zeroes but TARGET_FMT_plx definition in targphys.h has 0 after %. I can define another one without leading zeroes. Can you suggest a name?

I think OpenBIOS for Sparc64 is not correct here, so it may be a bad reference architecture. OBP on a real Ultra-5 used a path like this:
/pci@1f,0/pci@1,1/ide@3/disk@0,0

pci@1f,0 specifies the PCI host bridge at UPA bus port ID of 0x1f. pci@1,1 specifies a PCI-PCI bridge.

Quoted 4 lines

 TARGET_FMT_lx is poisoned. As of ATA there is no open firmware binding spec for ATA, so everyone does what he pleases. I based my implementation on what open firmware showing when running on qemu x86. "pci-ata" should be "ide" according to PCI binding spec :)

Yes, for example there is no ATA in the Ultra-5 tree but in UltraAX it exists:
/pci@1f,4000/ide@3/ata@0,0/cmdk@0,0

hw/apb_pci.c static PCIDeviceInfo pbm_pci_host_info = { · line 456

Quoted 9 lines
hw/apb_pci.c +1 -0
index c619112..643aa49 100644
@@ -453,6 +453,7 @@ static PCIDeviceInfo pbm_pci_host_info = {
 
  static SysBusDeviceInfo pbm_host_info = {
     .qdev.name = "pbm",
+    .qdev.fw_name = "pci",

Perhaps the FW path should use device class names if no name is specified.

I'll try Sparc32 to see how this fits there.

#31 lore raw

Re: [PATCHv4 15/15] Pass boot device list to firmware.

Quoted 90 lines · Blue Swirl

2010/11/15 Gleb Natapov <[email protected]>:

Quoted 80 lines · Blue Swirl
Quoted 55 lines · Gleb Natapov
Quoted 54 lines

---
 hw/fw_cfg.c |   14 ++++++++++++++
 hw/fw_cfg.h |    4 +++-
 sysemu.h    |    1 +
 vl.c        |   51 +++++++++++++++++++++++++++++++++++++++++++++++++++

 4 files changed, 69 insertions(+), 1 deletions(-)
hw/fw_cfg.c +14 -0
index 7b9434f..f6a67db 100644
@@ -53,6 +53,7 @@ struct FWCfgState {
     FWCfgFiles *files;
     uint16_t cur_entry;
     uint32_t cur_offset;
+    Notifier machine_ready;
  };
 
  static void fw_cfg_write(FWCfgState *s, uint8_t value)
@@ -315,6 +316,15 @@ int fw_cfg_add_file(FWCfgState *s,  const char *filename, uint8_t *data,
     return 1;
  }
 
+static void fw_cfg_machine_ready(struct Notifier* n)
+{
+    uint32_t len;
+    char *bootindex = get_boot_devices_list(&len);
+
+    fw_cfg_add_bytes(container_of(n, FWCfgState, machine_ready),
+                     FW_CFG_BOOTINDEX, (uint8_t*)bootindex, len);
+}
+
  FWCfgState *fw_cfg_init(uint32_t ctl_port, uint32_t data_port,
                         target_phys_addr_t ctl_addr, target_phys_addr_t data_addr)
  {
@@ -343,6 +353,10 @@ FWCfgState *fw_cfg_init(uint32_t ctl_port, uint32_t data_port,
     fw_cfg_add_i16(s, FW_CFG_MAX_CPUS, (uint16_t)max_cpus);
     fw_cfg_add_i16(s, FW_CFG_BOOT_MENU, (uint16_t)boot_menu);
 
+
+    s->machine_ready.notify = fw_cfg_machine_ready;
+    qemu_add_machine_init_done_notifier(&s->machine_ready);
+
     return s;
  }
 
hw/fw_cfg.h +0 -0
index 856bf91..4d61410 100644
@@ -30,7 +30,9 @@
 
Quoted 13 lines

hw/fw_cfg.h line 35

Quoted 6 lines
hw/fw_cfg.h +3 -1
  #define FW_CFG_FILE_FIRST       0x20
  #define FW_CFG_FILE_SLOTS       0x10
-#define FW_CFG_MAX_ENTRY        (FW_CFG_FILE_FIRST+FW_CFG_FILE_SLOTS)
+#define FW_CFG_FILE_LAST_SLOT   (FW_CFG_FILE_FIRST+FW_CFG_FILE_SLOTS)
+#define FW_CFG_BOOTINDEX        (FW_CFG_FILE_LAST_SLOT + 1)
+#define FW_CFG_MAX_ENTRY        FW_CFG_BOOTINDEX

This should be
#define FW_CFG_MAX_ENTRY        (FW_CFG_BOOTINDEX + 1) because the check is like this:
    if ((key & FW_CFG_ENTRY_MASK) >= FW_CFG_MAX_ENTRY) { s->cur_entry = FW_CFG_INVALID;

Yeah, will fix.

Quoted 9 lines

With that change, I got the bootindex passed to OpenBIOS:
OpenBIOS for Sparc64
Configuration device id QEMU version 1 machine id 0 kernel cmdline
CPUs: 1 x SUNW,UltraSPARC-IIi
UUID: 00000000-0000-0000-0000-000000000000 bootindex num_strings 1
bootindex /pbm@000001fe00000000/ide@5/drive@1/disk@0

Quoted 8 lines
Quoted 2 lines

The device path does not match exactly, but it's close:
/pci@1fe,0/pci-ata@5/ide1@600/disk@0

pbm->pci should be solvable by the patch at the end. Were in the spec it is allowed to abbreviate 1fe00000000 as 1fe,0? Spec allows to drop starting zeroes but TARGET_FMT_plx definition in targphys.h has 0 after %. I can define another one without leading zeroes. Can you suggest a name?

Quoted 5 lines

I think OpenBIOS for Sparc64 is not correct here, so it may be a bad reference architecture. OBP on a real Ultra-5 used a path like this:
/pci@1f,0/pci@1,1/ide@3/disk@0,0

pci@1f,0 specifies the PCI host bridge at UPA bus port ID of 0x1f.

According to device name qemu creates pci controller is memory mapped at address 1fe00000000 and by looking at the code I can see that this is indeed the case. How is UPA naming works?

Quoted 16 lines

pci@1,1 specifies a PCI-PCI bridge.

Quoted 4 lines

 TARGET_FMT_lx is poisoned. As of ATA there is no open firmware binding spec for ATA, so everyone does what he pleases. I based my implementation on what open firmware showing when running on qemu x86. "pci-ata" should be "ide" according to PCI binding spec :)

Yes, for example there is no ATA in the Ultra-5 tree but in UltraAX it exists:
/pci@1f,4000/ide@3/ata@0,0/cmdk@0,0

Quoted 6 lines
hw/apb_pci.c +0 -0
index c619112..643aa49 100644
@@ -453,6 +453,7 @@ static PCIDeviceInfo pbm_pci_host_info = {
 
Quoted 5 lines

hw/apb_pci.c static PCIDeviceInfo pbm_pci_host_info = { · line 456

Quoted 3 lines
hw/apb_pci.c +1 -0
  static SysBusDeviceInfo pbm_host_info = {
     .qdev.name = "pbm",
+    .qdev.fw_name = "pci",

Perhaps the FW path should use device class names if no name is specified.

What do you mean by "device class name". We can do something like this:
if (dev->child_bus.lh_first)

	return dev->child_bus.lh_first->info->name;

i.e if there is child bus use its bus name as fw name. This will make all pci devices to have "pci" as fw name automatically. The problem is that theoretically same device can provide different buses.

Quoted 2 lines

I'll try Sparc32 to see how this fits there.

--
Gleb.

#32 lore raw

Re: [PATCHv4 15/15] Pass boot device list to firmware.

2010/11/16 Gleb Natapov <[email protected]>:

Quoted 91 lines · Blue Swirl
Quoted 90 lines

2010/11/15 Gleb Natapov <[email protected]>:

Quoted 88 lines · Blue Swirl
Quoted 55 lines · Gleb Natapov
Quoted 54 lines

---
 hw/fw_cfg.c |   14 ++++++++++++++
 hw/fw_cfg.h |    4 +++-
 sysemu.h    |    1 +
 vl.c        |   51 +++++++++++++++++++++++++++++++++++++++++++++++++++

 4 files changed, 69 insertions(+), 1 deletions(-)
hw/fw_cfg.c +14 -0
index 7b9434f..f6a67db 100644
@@ -53,6 +53,7 @@ struct FWCfgState {
     FWCfgFiles *files;
     uint16_t cur_entry;
     uint32_t cur_offset;
+    Notifier machine_ready;
  };
 
  static void fw_cfg_write(FWCfgState *s, uint8_t value)
@@ -315,6 +316,15 @@ int fw_cfg_add_file(FWCfgState *s,  const char *filename, uint8_t *data,
     return 1;
  }
 
+static void fw_cfg_machine_ready(struct Notifier* n)
+{
+    uint32_t len;
+    char *bootindex = get_boot_devices_list(&len);
+
+    fw_cfg_add_bytes(container_of(n, FWCfgState, machine_ready),
+                     FW_CFG_BOOTINDEX, (uint8_t*)bootindex, len);
+}
+
  FWCfgState *fw_cfg_init(uint32_t ctl_port, uint32_t data_port,
                         target_phys_addr_t ctl_addr, target_phys_addr_t data_addr)
  {
@@ -343,6 +353,10 @@ FWCfgState *fw_cfg_init(uint32_t ctl_port, uint32_t data_port,
     fw_cfg_add_i16(s, FW_CFG_MAX_CPUS, (uint16_t)max_cpus);
     fw_cfg_add_i16(s, FW_CFG_BOOT_MENU, (uint16_t)boot_menu);
 
+
+    s->machine_ready.notify = fw_cfg_machine_ready;
+    qemu_add_machine_init_done_notifier(&s->machine_ready);
+
     return s;
  }
 
hw/fw_cfg.h +0 -0
index 856bf91..4d61410 100644
@@ -30,7 +30,9 @@
 
Quoted 13 lines

hw/fw_cfg.h line 35

Quoted 6 lines
hw/fw_cfg.h +3 -1
  #define FW_CFG_FILE_FIRST       0x20
  #define FW_CFG_FILE_SLOTS       0x10
-#define FW_CFG_MAX_ENTRY        (FW_CFG_FILE_FIRST+FW_CFG_FILE_SLOTS)
+#define FW_CFG_FILE_LAST_SLOT   (FW_CFG_FILE_FIRST+FW_CFG_FILE_SLOTS)
+#define FW_CFG_BOOTINDEX        (FW_CFG_FILE_LAST_SLOT + 1)
+#define FW_CFG_MAX_ENTRY        FW_CFG_BOOTINDEX

This should be
#define FW_CFG_MAX_ENTRY        (FW_CFG_BOOTINDEX + 1) because the check is like this:
    if ((key & FW_CFG_ENTRY_MASK) >= FW_CFG_MAX_ENTRY) { s->cur_entry = FW_CFG_INVALID;

Yeah, will fix.

Quoted 11 lines

With that change, I got the bootindex passed to OpenBIOS:
OpenBIOS for Sparc64
Configuration device id QEMU version 1 machine id 0 kernel cmdline
CPUs: 1 x SUNW,UltraSPARC-IIi
UUID: 00000000-0000-0000-0000-000000000000 bootindex num_strings 1
bootindex /pbm@000001fe00000000/ide@5/drive@1/disk@0

The device path does not match exactly, but it's close:
/pci@1fe,0/pci-ata@5/ide1@600/disk@0

pbm->pci should be solvable by the patch at the end. Were in the spec it is allowed to abbreviate 1fe00000000 as 1fe,0? Spec allows to drop starting zeroes but TARGET_FMT_plx definition in targphys.h has 0 after %. I can define another one without leading zeroes. Can you suggest a name?

Quoted 8 lines
Quoted 5 lines

I think OpenBIOS for Sparc64 is not correct here, so it may be a bad reference architecture. OBP on a real Ultra-5 used a path like this:
/pci@1f,0/pci@1,1/ide@3/disk@0,0

pci@1f,0 specifies the PCI host bridge at UPA bus port ID of 0x1f.

According to device name qemu creates pci controller is memory mapped at address 1fe00000000 and by looking at the code I can see that this is indeed the case. How is UPA naming works?

No idea.

Quoted 21 lines
Quoted 16 lines

pci@1,1 specifies a PCI-PCI bridge.

Quoted 4 lines

 TARGET_FMT_lx is poisoned. As of ATA there is no open firmware binding spec for ATA, so everyone does what he pleases. I based my implementation on what open firmware showing when running on qemu x86. "pci-ata" should be "ide" according to PCI binding spec :)

Yes, for example there is no ATA in the Ultra-5 tree but in UltraAX it exists:
/pci@1f,4000/ide@3/ata@0,0/cmdk@0,0

Quoted 6 lines
hw/apb_pci.c +0 -0
index c619112..643aa49 100644
@@ -453,6 +453,7 @@ static PCIDeviceInfo pbm_pci_host_info = {
 
Quoted 5 lines

hw/apb_pci.c static PCIDeviceInfo pbm_pci_host_info = { · line 456

Quoted 3 lines
hw/apb_pci.c +1 -0
  static SysBusDeviceInfo pbm_host_info = {
     .qdev.name = "pbm",
+    .qdev.fw_name = "pci",

Perhaps the FW path should use device class names if no name is specified.

Quoted 7 lines

What do you mean by "device class name". We can do something like this:
if (dev->child_bus.lh_first)
       return dev->child_bus.lh_first->info->name;

i.e if there is child bus use its bus name as fw name. This will make all pci devices to have "pci" as fw name automatically. The problem is that theoretically same device can provide different buses.

I meant PCI class name, like "display" for display controllers, "network" for NICs etc.

Quoted 1 line
Quoted 1 line

I'll try Sparc32 to see how this fits there.

Except bootindex is not implemented for SCSI.

#33 lore raw

Re: [PATCHv4 15/15] Pass boot device list to firmware.

Quoted 5 lines · Blue Swirl
Quoted 5 lines
Quoted 1 line

Perhaps the FW path should use device class names if no name is specified.

What do you mean by "device class name". We can do something like this:
if (dev->child_bus.lh_first)
       return dev->child_bus.lh_first->info->name;

Quoted 7 lines
Quoted 3 lines

i.e if there is child bus use its bus name as fw name. This will make all pci devices to have "pci" as fw name automatically. The problem is that theoretically same device can provide different buses.

I meant PCI class name, like "display" for display controllers, "network" for NICs etc.

That is what my pci bus related patch is doing already.

Quoted 3 lines
Quoted 1 line
Quoted 1 line

I'll try Sparc32 to see how this fits there.

Except bootindex is not implemented for SCSI.

Will look into adding it.

--
Gleb.

#34 lore raw

Re: [PATCHv4 15/15] Pass boot device list to firmware.

2010/11/16 Gleb Natapov <[email protected]>:

Quoted 13 lines · Blue Swirl
Quoted 5 lines
Quoted 5 lines
Quoted 1 line

Perhaps the FW path should use device class names if no name is specified.

What do you mean by "device class name". We can do something like this:
if (dev->child_bus.lh_first)
       return dev->child_bus.lh_first->info->name;

Quoted 7 lines
Quoted 3 lines

i.e if there is child bus use its bus name as fw name. This will make all pci devices to have "pci" as fw name automatically. The problem is that theoretically same device can provide different buses.

I meant PCI class name, like "display" for display controllers, "network" for NICs etc.

Quoted 6 lines

That is what my pci bus related patch is doing already.

Quoted 3 lines
Quoted 1 line
Quoted 1 line

I'll try Sparc32 to see how this fits there.

Except bootindex is not implemented for SCSI.

Will look into adding it.

Thanks. The bootindex on Sparc32 looks like this:
bootindex /esp@0000000078800000/disk@1,0
/ethernet@ffffffffffffffff/ethernet-phy@0

I don't think I got Lance setup right.

OF paths for the devices would be:
/iommu@0,10000000/sbus@0,10001000/espdma@5,8400000/esp@5,8800000/sd@1,0 /iommu@0,10000000/sbus@0,10001000/ledma@5,8400010/le@5,8c00000

The logic for ESP is that ESP (registers at 0x78800000, slot offset 0x880000) is handled by the DMA controller (registers at 0x78400000, slot offset 0x840000), they are in a SBus slot #5, and SBus (registers at 0x10001000) is in turn handled by IOMMU (registers at 0x10000000). Lance should be handled the same way.

This hierarchy is partly known by QEMU because DMA accesses use this flow, but not otherwise. There is no concept of SBus slots, DMA talks to IOMMU directly. Though in this case both ESP, Lance and their DMA controllers are on board devices in a MACIO chip. It may be possible to add the hierarchy information at each stage.

It should also be possible for BIOS to determine the device just from the physical address if we ignored OF compatibility.

#35 lore raw

Re: [PATCHv4 15/15] Pass boot device list to firmware.

Quoted 16 lines · Blue Swirl

2010/11/16 Gleb Natapov <[email protected]>:

Quoted 15 lines · Blue Swirl
Quoted 5 lines
Quoted 5 lines
Quoted 1 line

Perhaps the FW path should use device class names if no name is specified.

What do you mean by "device class name". We can do something like this:
if (dev->child_bus.lh_first)
       return dev->child_bus.lh_first->info->name;

Quoted 7 lines
Quoted 3 lines

i.e if there is child bus use its bus name as fw name. This will make all pci devices to have "pci" as fw name automatically. The problem is that theoretically same device can provide different buses.

I meant PCI class name, like "display" for display controllers, "network" for NICs etc.

That is what my pci bus related patch is doing already.

Quoted 9 lines
Quoted 4 lines
Quoted 3 lines
Quoted 1 line
Quoted 1 line

I'll try Sparc32 to see how this fits there.

Except bootindex is not implemented for SCSI.

Will look into adding it.

Thanks. The bootindex on Sparc32 looks like this:
bootindex /esp@0000000078800000/disk@1,0
/ethernet@ffffffffffffffff/ethernet-phy@0

For arches other then x86 there is a lot of work left to be done :) For starter exotic sparc buses should get their own get_fw_dev_path() implementation.

Quoted 5 lines

I don't think I got Lance setup right.

OF paths for the devices would be:
/iommu@0,10000000/sbus@0,10001000/espdma@5,8400000/esp@5,8800000/sd@1,0 /iommu@0,10000000/sbus@0,10001000/ledma@5,8400010/le@5,8c00000

If qdev hierarchy does not correspond to real HW there is no much we can do expect for fixing qdev.

Quoted 7 lines

The logic for ESP is that ESP (registers at 0x78800000, slot offset 0x880000) is handled by the DMA controller (registers at 0x78400000, slot offset 0x840000), they are in a SBus slot #5, and SBus (registers at 0x10001000) is in turn handled by IOMMU (registers at 0x10000000). Lance should be handled the same way.

Quoted 8 lines

This hierarchy is partly known by QEMU because DMA accesses use this flow, but not otherwise. There is no concept of SBus slots, DMA talks to IOMMU directly. Though in this case both ESP, Lance and their DMA controllers are on board devices in a MACIO chip. It may be possible to add the hierarchy information at each stage.

It should also be possible for BIOS to determine the device just from the physical address if we ignored OF compatibility.

It would be nice to be OF compatible at least at some level. Of course OF spec is not strict enough to have two different implementations produce exactly same device path that can be compared by strcpy. Can we apply the series now? At least for x86 it provides useful paths and work can be continue for other arches by interested parties.

--
Gleb.

#36 lore raw

Re: [PATCHv4 15/15] Pass boot device list to firmware.

Quoted 30 lines · Gleb Natapov
Quoted 16 lines · Blue Swirl

2010/11/16 Gleb Natapov <[email protected]>:

Quoted 15 lines · Blue Swirl
Quoted 5 lines
Quoted 5 lines
Quoted 1 line

Perhaps the FW path should use device class names if no name is specified.

What do you mean by "device class name". We can do something like this:
if (dev->child_bus.lh_first)
       return dev->child_bus.lh_first->info->name;

Quoted 7 lines
Quoted 3 lines

i.e if there is child bus use its bus name as fw name. This will make all pci devices to have "pci" as fw name automatically. The problem is that theoretically same device can provide different buses.

I meant PCI class name, like "display" for display controllers, "network" for NICs etc.

That is what my pci bus related patch is doing already.

Quoted 9 lines
Quoted 4 lines
Quoted 3 lines
Quoted 1 line
Quoted 1 line

I'll try Sparc32 to see how this fits there.

Except bootindex is not implemented for SCSI.

Will look into adding it.

Thanks. The bootindex on Sparc32 looks like this:
bootindex /esp@0000000078800000/disk@1,0
/ethernet@ffffffffffffffff/ethernet-phy@0

For arches other then x86 there is a lot of work left to be done :) For starter exotic sparc buses should get their own get_fw_dev_path() implementation.

Quoted 7 lines
Quoted 5 lines

I don't think I got Lance setup right.

OF paths for the devices would be:
/iommu@0,10000000/sbus@0,10001000/espdma@5,8400000/esp@5,8800000/sd@1,0 /iommu@0,10000000/sbus@0,10001000/ledma@5,8400010/le@5,8c00000

If qdev hierarchy does not correspond to real HW there is no much we can do expect for fixing qdev.

That's bad. This raises a concern: if these paths expose qdev internals, any attempt to fix this will break migration.

Quoted 15 lines
Quoted 7 lines

The logic for ESP is that ESP (registers at 0x78800000, slot offset 0x880000) is handled by the DMA controller (registers at 0x78400000, slot offset 0x840000), they are in a SBus slot #5, and SBus (registers at 0x10001000) is in turn handled by IOMMU (registers at 0x10000000). Lance should be handled the same way.

Quoted 8 lines

This hierarchy is partly known by QEMU because DMA accesses use this flow, but not otherwise. There is no concept of SBus slots, DMA talks to IOMMU directly. Though in this case both ESP, Lance and their DMA controllers are on board devices in a MACIO chip. It may be possible to add the hierarchy information at each stage.

It should also be possible for BIOS to determine the device just from the physical address if we ignored OF compatibility.

Quoted 8 lines

It would be nice to be OF compatible at least at some level. Of course OF spec is not strict enough to have two different implementations produce exactly same device path that can be compared by strcpy. Can we apply the series now? At least for x86 it provides useful paths and work can be continue for other arches by interested parties.

--
Gleb.

Something I only now realized is that we commit
to never changing the paths for any architecture
that supports migration.

Signature · 1 line
MST
#37 lore raw

Re: [PATCHv4 15/15] Pass boot device list to firmware.

Quoted 33 lines · Michael S. Tsirkin
Quoted 32 lines · Gleb Natapov
Quoted 16 lines · Blue Swirl

2010/11/16 Gleb Natapov <[email protected]>:

Quoted 15 lines · Blue Swirl
Quoted 5 lines
Quoted 5 lines
Quoted 1 line

Perhaps the FW path should use device class names if no name is specified.

What do you mean by "device class name". We can do something like this:
if (dev->child_bus.lh_first)
       return dev->child_bus.lh_first->info->name;

Quoted 7 lines
Quoted 3 lines

i.e if there is child bus use its bus name as fw name. This will make all pci devices to have "pci" as fw name automatically. The problem is that theoretically same device can provide different buses.

I meant PCI class name, like "display" for display controllers, "network" for NICs etc.

That is what my pci bus related patch is doing already.

Quoted 9 lines
Quoted 4 lines
Quoted 3 lines
Quoted 1 line
Quoted 1 line

I'll try Sparc32 to see how this fits there.

Except bootindex is not implemented for SCSI.

Will look into adding it.

Thanks. The bootindex on Sparc32 looks like this:
bootindex /esp@0000000078800000/disk@1,0
/ethernet@ffffffffffffffff/ethernet-phy@0

For arches other then x86 there is a lot of work left to be done :) For starter exotic sparc buses should get their own get_fw_dev_path() implementation.

Quoted 2 lines

I don't think I got Lance setup right.

Quoted 9 lines
Quoted 5 lines
Quoted 3 lines

OF paths for the devices would be:
/iommu@0,10000000/sbus@0,10001000/espdma@5,8400000/esp@5,8800000/sd@1,0 /iommu@0,10000000/sbus@0,10001000/ledma@5,8400010/le@5,8c00000

If qdev hierarchy does not correspond to real HW there is no much we can do expect for fixing qdev.

That's bad. This raises a concern: if these paths expose qdev internals, any attempt to fix this will break migration.

The path expose internal HW hierarchy. It is designed to do so. Qdev designed to do the same: describe HW hierarchy. If qdev fails to do so it is broken. I do not see connection to migration at all since the path is not used in migration code.

Quoted 21 lines
Quoted 21 lines
Quoted 7 lines

The logic for ESP is that ESP (registers at 0x78800000, slot offset 0x880000) is handled by the DMA controller (registers at 0x78400000, slot offset 0x840000), they are in a SBus slot #5, and SBus (registers at 0x10001000) is in turn handled by IOMMU (registers at 0x10000000). Lance should be handled the same way.

Quoted 8 lines

This hierarchy is partly known by QEMU because DMA accesses use this flow, but not otherwise. There is no concept of SBus slots, DMA talks to IOMMU directly. Though in this case both ESP, Lance and their DMA controllers are on board devices in a MACIO chip. It may be possible to add the hierarchy information at each stage.

It should also be possible for BIOS to determine the device just from the physical address if we ignored OF compatibility.

It would be nice to be OF compatible at least at some level. Of course OF spec is not strict enough to have two different implementations produce exactly same device path that can be compared by strcpy. Can we apply the series now? At least for x86 it provides useful paths and work can be continue for other arches by interested parties.

Quoted 7 lines
Quoted 2 lines

--
Gleb.

Something I only now realized is that we commit
to never changing the paths for any architecture
that supports migration.

No connection to migration whatsoever.

--
Gleb.

#38 lore raw

Re: [PATCHv4 15/15] Pass boot device list to firmware.

Quoted 40 lines · Gleb Natapov
Quoted 39 lines · Michael S. Tsirkin
Quoted 37 lines · Gleb Natapov
Quoted 16 lines · Blue Swirl

2010/11/16 Gleb Natapov <[email protected]>:

Quoted 15 lines · Blue Swirl
Quoted 5 lines
Quoted 5 lines
Quoted 1 line

Perhaps the FW path should use device class names if no name is specified.

What do you mean by "device class name". We can do something like this:
if (dev->child_bus.lh_first)
       return dev->child_bus.lh_first->info->name;

Quoted 7 lines
Quoted 3 lines

i.e if there is child bus use its bus name as fw name. This will make all pci devices to have "pci" as fw name automatically. The problem is that theoretically same device can provide different buses.

I meant PCI class name, like "display" for display controllers, "network" for NICs etc.

That is what my pci bus related patch is doing already.

Quoted 9 lines
Quoted 4 lines
Quoted 3 lines
Quoted 1 line
Quoted 1 line

I'll try Sparc32 to see how this fits there.

Except bootindex is not implemented for SCSI.

Will look into adding it.

Thanks. The bootindex on Sparc32 looks like this:
bootindex /esp@0000000078800000/disk@1,0
/ethernet@ffffffffffffffff/ethernet-phy@0

For arches other then x86 there is a lot of work left to be done :) For starter exotic sparc buses should get their own get_fw_dev_path() implementation.

Quoted 5 lines

I don't think I got Lance setup right.

OF paths for the devices would be:
/iommu@0,10000000/sbus@0,10001000/espdma@5,8400000/esp@5,8800000/sd@1,0 /iommu@0,10000000/sbus@0,10001000/ledma@5,8400010/le@5,8c00000

If qdev hierarchy does not correspond to real HW there is no much we can do expect for fixing qdev.

Quoted 6 lines
Quoted 3 lines

That's bad. This raises a concern: if these paths expose qdev internals, any attempt to fix this will break migration.

The path expose internal HW hierarchy. It is designed to do so. Qdev designed to do the same: describe HW hierarchy. If qdev fails to do so it is broken.

Yes. But since you use qdev to build up the path, a broken qdev will give you a broken path.

Quoted 2 lines

I do not see connection to migration at all since the path is not used in migration code.

The connection is that if we pass the list with path 1 which you define as broken to BIOS, then migrate to a machine with an updated qemu which has a correct path, BIOS won't be able to complete the boot. Right? Same in reverse direction.
As solution could be a fuzzy matching
of paths that wiull let us recover.

Quoted 21 lines
Quoted 21 lines
Quoted 21 lines
Quoted 7 lines

The logic for ESP is that ESP (registers at 0x78800000, slot offset 0x880000) is handled by the DMA controller (registers at 0x78400000, slot offset 0x840000), they are in a SBus slot #5, and SBus (registers at 0x10001000) is in turn handled by IOMMU (registers at 0x10000000). Lance should be handled the same way.

Quoted 8 lines

This hierarchy is partly known by QEMU because DMA accesses use this flow, but not otherwise. There is no concept of SBus slots, DMA talks to IOMMU directly. Though in this case both ESP, Lance and their DMA controllers are on board devices in a MACIO chip. It may be possible to add the hierarchy information at each stage.

It should also be possible for BIOS to determine the device just from the physical address if we ignored OF compatibility.

It would be nice to be OF compatible at least at some level. Of course OF spec is not strict enough to have two different implementations produce exactly same device path that can be compared by strcpy. Can we apply the series now? At least for x86 it provides useful paths and work can be continue for other arches by interested parties.

Quoted 8 lines
Quoted 7 lines
Quoted 2 lines

--
Gleb.

Something I only now realized is that we commit
to never changing the paths for any architecture
that supports migration.

No connection to migration whatsoever.

It just seems silly to use different paths for the same thing.

Besides the connection above, I was hoping to use these paths for section names in migration. If we can't guarantee they are stable, we'll have to roll our own, and if we do this, with stability guarantees required for migration format, maybe use it for other things like BIOS as well?

Quoted 2 lines

--
Gleb.

#39 lore raw

Re: [PATCHv4 15/15] Pass boot device list to firmware.

Quoted 44 lines · Michael S. Tsirkin
Quoted 43 lines · Gleb Natapov
Quoted 42 lines · Michael S. Tsirkin
Quoted 30 lines · Gleb Natapov
Quoted 16 lines · Blue Swirl

2010/11/16 Gleb Natapov <[email protected]>:

Quoted 15 lines · Blue Swirl
Quoted 5 lines
Quoted 5 lines
Quoted 1 line

Perhaps the FW path should use device class names if no name is specified.

What do you mean by "device class name". We can do something like this:
if (dev->child_bus.lh_first)
       return dev->child_bus.lh_first->info->name;

Quoted 7 lines
Quoted 3 lines

i.e if there is child bus use its bus name as fw name. This will make all pci devices to have "pci" as fw name automatically. The problem is that theoretically same device can provide different buses.

I meant PCI class name, like "display" for display controllers, "network" for NICs etc.

That is what my pci bus related patch is doing already.

Quoted 9 lines
Quoted 4 lines
Quoted 3 lines
Quoted 1 line
Quoted 1 line

I'll try Sparc32 to see how this fits there.

Except bootindex is not implemented for SCSI.

Will look into adding it.

Thanks. The bootindex on Sparc32 looks like this:
bootindex /esp@0000000078800000/disk@1,0
/ethernet@ffffffffffffffff/ethernet-phy@0

For arches other then x86 there is a lot of work left to be done :) For starter exotic sparc buses should get their own get_fw_dev_path() implementation.

Quoted 7 lines
Quoted 5 lines

I don't think I got Lance setup right.

OF paths for the devices would be:
/iommu@0,10000000/sbus@0,10001000/espdma@5,8400000/esp@5,8800000/sd@1,0 /iommu@0,10000000/sbus@0,10001000/ledma@5,8400010/le@5,8c00000

If qdev hierarchy does not correspond to real HW there is no much we can do expect for fixing qdev.

That's bad. This raises a concern: if these paths expose qdev internals, any attempt to fix this will break migration.

Quoted 7 lines
Quoted 3 lines

The path expose internal HW hierarchy. It is designed to do so. Qdev designed to do the same: describe HW hierarchy. If qdev fails to do so it is broken.

Yes. But since you use qdev to build up the path, a broken qdev will give you a broken path.

Qdev bug. Fix it like any other bug. The nice is that when you compare device path produced by qdev and real HW you can see when qdev is wrong.

Quoted 6 lines
Quoted 2 lines

I do not see connection to migration at all since the path is not used in migration code.

The connection is that if we pass the list with path 1 which you define as broken to BIOS, then migrate to a machine with an updated qemu which has a correct path, BIOS won't be able to complete the boot.

You solve it like you solve all such issue with -M machine type. But the problem exists only if migration happens in a short window between start of the boot process and BIOS reading boot order string. After reboot new qemu should have new BIOS.

Quoted 1 line

Right? Same in reverse direction.

Reverse direction is not and never was supported.

Quoted 3 lines

As solution could be a fuzzy matching
of paths that wiull let us recover.

Firmware can try its best of course, but nothing is guarantied.

Quoted 30 lines
Quoted 21 lines
Quoted 21 lines
Quoted 21 lines
Quoted 7 lines

The logic for ESP is that ESP (registers at 0x78800000, slot offset 0x880000) is handled by the DMA controller (registers at 0x78400000, slot offset 0x840000), they are in a SBus slot #5, and SBus (registers at 0x10001000) is in turn handled by IOMMU (registers at 0x10000000). Lance should be handled the same way.

Quoted 8 lines

This hierarchy is partly known by QEMU because DMA accesses use this flow, but not otherwise. There is no concept of SBus slots, DMA talks to IOMMU directly. Though in this case both ESP, Lance and their DMA controllers are on board devices in a MACIO chip. It may be possible to add the hierarchy information at each stage.

It should also be possible for BIOS to determine the device just from the physical address if we ignored OF compatibility.

It would be nice to be OF compatible at least at some level. Of course OF spec is not strict enough to have two different implementations produce exactly same device path that can be compared by strcpy. Can we apply the series now? At least for x86 it provides useful paths and work can be continue for other arches by interested parties.

Quoted 8 lines
Quoted 7 lines
Quoted 2 lines

--
Gleb.

Something I only now realized is that we commit
to never changing the paths for any architecture
that supports migration.

No connection to migration whatsoever.

Quoted 8 lines

It just seems silly to use different paths for the same thing.

Besides the connection above, I was hoping to use these paths for section names in migration. If we can't guarantee they are stable, we'll have to roll our own, and if we do this, with stability guarantees required for migration format, maybe use it for other things like BIOS as well?

It doesn't matter what do you use for migration purposes as long as it depend on qdev hierarchy it will have problem when qdev hierarchy changes and if it doesn't you can't produce unique names reliably.

--
Gleb.

#40 lore raw

Re: [PATCHv4 15/15] Pass boot device list to firmware.

Quoted 55 lines · Gleb Natapov
Quoted 44 lines · Michael S. Tsirkin
Quoted 43 lines · Gleb Natapov
Quoted 42 lines · Michael S. Tsirkin
Quoted 30 lines · Gleb Natapov
Quoted 16 lines · Blue Swirl

2010/11/16 Gleb Natapov <[email protected]>:

Quoted 15 lines · Blue Swirl
Quoted 5 lines
Quoted 5 lines
Quoted 1 line

Perhaps the FW path should use device class names if no name is specified.

What do you mean by "device class name". We can do something like this:
if (dev->child_bus.lh_first)
       return dev->child_bus.lh_first->info->name;

Quoted 7 lines
Quoted 3 lines

i.e if there is child bus use its bus name as fw name. This will make all pci devices to have "pci" as fw name automatically. The problem is that theoretically same device can provide different buses.

I meant PCI class name, like "display" for display controllers, "network" for NICs etc.

That is what my pci bus related patch is doing already.

Quoted 9 lines
Quoted 4 lines
Quoted 3 lines
Quoted 1 line
Quoted 1 line

I'll try Sparc32 to see how this fits there.

Except bootindex is not implemented for SCSI.

Will look into adding it.

Thanks. The bootindex on Sparc32 looks like this:
bootindex /esp@0000000078800000/disk@1,0
/ethernet@ffffffffffffffff/ethernet-phy@0

For arches other then x86 there is a lot of work left to be done :) For starter exotic sparc buses should get their own get_fw_dev_path() implementation.

Quoted 7 lines
Quoted 5 lines

I don't think I got Lance setup right.

OF paths for the devices would be:
/iommu@0,10000000/sbus@0,10001000/espdma@5,8400000/esp@5,8800000/sd@1,0 /iommu@0,10000000/sbus@0,10001000/ledma@5,8400010/le@5,8c00000

If qdev hierarchy does not correspond to real HW there is no much we can do expect for fixing qdev.

That's bad. This raises a concern: if these paths expose qdev internals, any attempt to fix this will break migration.

Quoted 7 lines
Quoted 3 lines

The path expose internal HW hierarchy. It is designed to do so. Qdev designed to do the same: describe HW hierarchy. If qdev fails to do so it is broken.

Yes. But since you use qdev to build up the path, a broken qdev will give you a broken path.

Qdev bug. Fix it like any other bug. The nice is that when you compare device path produced by qdev and real HW you can see when qdev is wrong.

Quoted 7 lines
Quoted 6 lines
Quoted 2 lines

I do not see connection to migration at all since the path is not used in migration code.

The connection is that if we pass the list with path 1 which you define as broken to BIOS, then migrate to a machine with an updated qemu which has a correct path, BIOS won't be able to complete the boot.

You solve it like you solve all such issue with -M machine type.

So that's unavoidable if we think paths are correct. But if we know they are wrong, we are better off correcting them first IMO.

Quoted 3 lines

But the problem exists only if migration happens in a short window between start of the boot process and BIOS reading boot order string. After reboot new qemu should have new BIOS.

That makes it even more nasty, doesn't it?

Quoted 7 lines
Quoted 1 line

Right? Same in reverse direction.

Reverse direction is not and never was supported.

Quoted 3 lines

As solution could be a fuzzy matching
of paths that wiull let us recover.

Firmware can try its best of course, but nothing is guarantied.

No I mean qemu could do matching fuzzily. This way if we get a path from the old BIOS we can survive.

Quoted 38 lines
Quoted 30 lines
Quoted 21 lines
Quoted 21 lines
Quoted 21 lines
Quoted 7 lines

The logic for ESP is that ESP (registers at 0x78800000, slot offset 0x880000) is handled by the DMA controller (registers at 0x78400000, slot offset 0x840000), they are in a SBus slot #5, and SBus (registers at 0x10001000) is in turn handled by IOMMU (registers at 0x10000000). Lance should be handled the same way.

Quoted 8 lines

This hierarchy is partly known by QEMU because DMA accesses use this flow, but not otherwise. There is no concept of SBus slots, DMA talks to IOMMU directly. Though in this case both ESP, Lance and their DMA controllers are on board devices in a MACIO chip. It may be possible to add the hierarchy information at each stage.

It should also be possible for BIOS to determine the device just from the physical address if we ignored OF compatibility.

It would be nice to be OF compatible at least at some level. Of course OF spec is not strict enough to have two different implementations produce exactly same device path that can be compared by strcpy. Can we apply the series now? At least for x86 it provides useful paths and work can be continue for other arches by interested parties.

Quoted 8 lines
Quoted 7 lines
Quoted 2 lines

--
Gleb.

Something I only now realized is that we commit
to never changing the paths for any architecture
that supports migration.

No connection to migration whatsoever.

Quoted 8 lines

It just seems silly to use different paths for the same thing.

Besides the connection above, I was hoping to use these paths for section names in migration. If we can't guarantee they are stable, we'll have to roll our own, and if we do this, with stability guarantees required for migration format, maybe use it for other things like BIOS as well?

Quoted 6 lines

It doesn't matter what do you use for migration purposes as long as it depend on qdev hierarchy it will have problem when qdev hierarchy changes and if it doesn't you can't produce unique names reliably.

--
Gleb.

We can, it's not like OF is the only way to enumerate.
We could have driver-specific paths for example, exactly like we currently have. I.e. paths don't have to be globally unique because each driver has it's own domain.

It seems cleaner to use an existing spec but we must figure out how it will not become a support issue.

Signature · 1 line
MST
#41 lore raw

Re: [PATCHv4 15/15] Pass boot device list to firmware.

Quoted 59 lines · Michael S. Tsirkin
Quoted 58 lines · Gleb Natapov
Quoted 44 lines · Michael S. Tsirkin
Quoted 43 lines · Gleb Natapov
Quoted 42 lines · Michael S. Tsirkin
Quoted 30 lines · Gleb Natapov
Quoted 16 lines · Blue Swirl

2010/11/16 Gleb Natapov <[email protected]>:

Quoted 15 lines · Blue Swirl
Quoted 5 lines
Quoted 5 lines
Quoted 1 line

Perhaps the FW path should use device class names if no name is specified.

What do you mean by "device class name". We can do something like this:
if (dev->child_bus.lh_first)
       return dev->child_bus.lh_first->info->name;

Quoted 7 lines
Quoted 3 lines

i.e if there is child bus use its bus name as fw name. This will make all pci devices to have "pci" as fw name automatically. The problem is that theoretically same device can provide different buses.

I meant PCI class name, like "display" for display controllers, "network" for NICs etc.

That is what my pci bus related patch is doing already.

Quoted 9 lines
Quoted 4 lines
Quoted 3 lines
Quoted 1 line
Quoted 1 line

I'll try Sparc32 to see how this fits there.

Except bootindex is not implemented for SCSI.

Will look into adding it.

Thanks. The bootindex on Sparc32 looks like this:
bootindex /esp@0000000078800000/disk@1,0
/ethernet@ffffffffffffffff/ethernet-phy@0

For arches other then x86 there is a lot of work left to be done :) For starter exotic sparc buses should get their own get_fw_dev_path() implementation.

Quoted 7 lines
Quoted 5 lines

I don't think I got Lance setup right.

OF paths for the devices would be:
/iommu@0,10000000/sbus@0,10001000/espdma@5,8400000/esp@5,8800000/sd@1,0 /iommu@0,10000000/sbus@0,10001000/ledma@5,8400010/le@5,8c00000

If qdev hierarchy does not correspond to real HW there is no much we can do expect for fixing qdev.

That's bad. This raises a concern: if these paths expose qdev internals, any attempt to fix this will break migration.

Quoted 7 lines
Quoted 3 lines

The path expose internal HW hierarchy. It is designed to do so. Qdev designed to do the same: describe HW hierarchy. If qdev fails to do so it is broken.

Yes. But since you use qdev to build up the path, a broken qdev will give you a broken path.

Qdev bug. Fix it like any other bug. The nice is that when you compare device path produced by qdev and real HW you can see when qdev is wrong.

Quoted 3 lines
Quoted 2 lines

I do not see connection to migration at all since the path is not used in migration code.

Quoted 9 lines
Quoted 4 lines
Quoted 3 lines

The connection is that if we pass the list with path 1 which you define as broken to BIOS, then migrate to a machine with an updated qemu which has a correct path, BIOS won't be able to complete the boot.

You solve it like you solve all such issue with -M machine type.

So that's unavoidable if we think paths are correct. But if we know they are wrong, we are better off correcting them first IMO.

They are correct for x86. My patch set does not even tries to cover all HW. If sparc want to use them to it better be fixed. Or if there is enough info in the path to determine device it may choose to use it as is.

Quoted 5 lines
Quoted 3 lines

But the problem exists only if migration happens in a short window between start of the boot process and BIOS reading boot order string. After reboot new qemu should have new BIOS.

That makes it even more nasty, doesn't it?

No.

Quoted 4 lines
Quoted 4 lines
Quoted 2 lines

Right? Same in reverse direction.

Reverse direction is not and never was supported.

Quoted 8 lines
Quoted 4 lines
Quoted 3 lines

As solution could be a fuzzy matching
of paths that wiull let us recover.

Firmware can try its best of course, but nothing is guarantied.

No I mean qemu could do matching fuzzily. This way if we get a path from the old BIOS we can survive.

Qemu does not take paths from BIOS so I don't know what are you talking about here.

Quoted 43 lines
Quoted 43 lines
Quoted 31 lines
Quoted 22 lines
Quoted 22 lines
Quoted 22 lines
Quoted 8 lines

The logic for ESP is that ESP (registers at 0x78800000, slot offset 0x880000) is handled by the DMA controller (registers at 0x78400000, slot offset 0x840000), they are in a SBus slot #5, and SBus (registers at 0x10001000) is in turn handled by IOMMU (registers at 0x10000000). Lance should be handled the same way.

Quoted 8 lines

This hierarchy is partly known by QEMU because DMA accesses use this flow, but not otherwise. There is no concept of SBus slots, DMA talks to IOMMU directly. Though in this case both ESP, Lance and their DMA controllers are on board devices in a MACIO chip. It may be possible to add the hierarchy information at each stage.

It should also be possible for BIOS to determine the device just from the physical address if we ignored OF compatibility.

It would be nice to be OF compatible at least at some level. Of course OF spec is not strict enough to have two different implementations produce exactly same device path that can be compared by strcpy. Can we apply the series now? At least for x86 it provides useful paths and work can be continue for other arches by interested parties.

Quoted 8 lines
Quoted 7 lines
Quoted 2 lines

--
Gleb.

Something I only now realized is that we commit
to never changing the paths for any architecture
that supports migration.

No connection to migration whatsoever.

Quoted 8 lines

It just seems silly to use different paths for the same thing.

Besides the connection above, I was hoping to use these paths for section names in migration. If we can't guarantee they are stable, we'll have to roll our own, and if we do this, with stability guarantees required for migration format, maybe use it for other things like BIOS as well?

It doesn't matter what do you use for migration purposes as long as it depend on qdev hierarchy it will have problem when qdev hierarchy changes and if it doesn't you can't produce unique names reliably.

Quoted 8 lines
Quoted 2 lines

--
Gleb.

We can, it's not like OF is the only way to enumerate.
We could have driver-specific paths for example, exactly like we currently have. I.e. paths don't have to be globally unique because each driver has it's own domain.

What we have currently is not even close to be correct. It happens to work since it is implemented only for one bus type and we can have only one of this bus right now. And of course it is not suitable for passing boot path to FW since it does not describe path to device.

Quoted 3 lines

It seems cleaner to use an existing spec but we must figure out how it will not become a support issue.

If you think you can figure out how to describe device path (or even give globally unique name to device) and to not depend on internal qdev implementation go ahead and do that for migration.

--
Gleb.

#42 lore raw

Re: [PATCHv4 15/15] Pass boot device list to firmware.

Quoted 7 lines · Gleb Natapov
Quoted 4 lines

So that's unavoidable if we think paths are correct. But if we know they are wrong, we are better off correcting them first IMO.

They are correct for x86. My patch set does not even tries to cover all HW. If sparc want to use them to it better be fixed. Or if there is enough info in the path to determine device it may choose to use it as is.

Fair enough I guess.

Quoted 6 lines
Quoted 5 lines
Quoted 3 lines

But the problem exists only if migration happens in a short window between start of the boot process and BIOS reading boot order string. After reboot new qemu should have new BIOS.

That makes it even more nasty, doesn't it?

No.

Nasty as in hard to reproduce.

Signature · 1 line
MST
#43 lore raw

Re: [PATCHv4 15/15] Pass boot device list to firmware.

Quoted 11 lines · Michael S. Tsirkin
Quoted 7 lines · Gleb Natapov
Quoted 4 lines

So that's unavoidable if we think paths are correct. But if we know they are wrong, we are better off correcting them first IMO.

They are correct for x86. My patch set does not even tries to cover all HW. If sparc want to use them to it better be fixed. Or if there is enough info in the path to determine device it may choose to use it as is.

Fair enough I guess.

Quoted 9 lines
Quoted 6 lines
Quoted 5 lines
Quoted 3 lines

But the problem exists only if migration happens in a short window between start of the boot process and BIOS reading boot order string. After reboot new qemu should have new BIOS.

That makes it even more nasty, doesn't it?

No.

Nasty as in hard to reproduce.

It is very easy to reproduce if you know what you are looking for :). Just stick sleep() in correct place in the BIOS.

--
Gleb.

#44 lore raw

Re: [PATCHv4 15/15] Pass boot device list to firmware.

In reply to Gleb Natapov
Quoted 11 lines · Gleb Natapov

+/*
+ * This function returns device list as an array in a below format:
+ * +-----+-----+---------------+-----+---------------+--
+ * | n | l1 | devpath1 | l2 | devpath2 | ...
+ * +-----+-----+---------------+-----+---------------+--
+ * where:

+ *   n - a number of devise pathes (one byte)
+ *   l - length of following device path string (one byte)
+ *   devpath - non-null terminated string of length l representing

+ * one device path
+ */

Why not just return a newline separated list that is null terminated?

-Kevin

#45 lore raw

Re: [PATCHv4 15/15] Pass boot device list to firmware.

Quoted 13 lines · Kevin O'Connor
Quoted 11 lines · Gleb Natapov

+/*
+ * This function returns device list as an array in a below format:
+ * +-----+-----+---------------+-----+---------------+--
+ * | n | l1 | devpath1 | l2 | devpath2 | ...
+ * +-----+-----+---------------+-----+---------------+--
+ * where:

+ *   n - a number of devise pathes (one byte)
+ *   l - length of following device path string (one byte)
+ *   devpath - non-null terminated string of length l representing

+ * one device path
+ */

Quoted 2 lines

Why not just return a newline separated list that is null terminated?

Doing it like this will needlessly complicate firmware side. How do you know how much memory to allocate before reading device list? Doing it like Blue suggest (have BOOTINDEX_LEN and BOOTINDEX_STRING) solves this. To create nice array from bootindex string you firmware will still have to do additional pass on it though. With format like above the code would look like that:

qemu_cfg_read(&n, 1);
arr = alloc(n);
for (i=0; i<n; i++) {
qemu_cfg_read(&l, 1);
arr[i] = zalloc(l+1);
qemu_cfg_read(arr[i], l);
}

--
Gleb.

#46 lore raw

Re: [PATCHv4 15/15] Pass boot device list to firmware.

Quoted 14 lines · Gleb Natapov
Quoted 13 lines · Kevin O'Connor
Quoted 11 lines · Gleb Natapov

+/*
+ * This function returns device list as an array in a below format:
+ * +-----+-----+---------------+-----+---------------+--
+ * | n | l1 | devpath1 | l2 | devpath2 | ...
+ * +-----+-----+---------------+-----+---------------+--
+ * where:

+ *   n - a number of devise pathes (one byte)
+ *   l - length of following device path string (one byte)
+ *   devpath - non-null terminated string of length l representing

+ * one device path
+ */

Quoted 4 lines
Quoted 2 lines

Why not just return a newline separated list that is null terminated?

Doing it like this will needlessly complicate firmware side. How do you know how much memory to allocate before reading device list?

Do a memory scan, count newlines until you reach 0?

Quoted 4 lines

Doing it
like Blue suggest (have BOOTINDEX_LEN and BOOTINDEX_STRING) solves this. To create nice array from bootindex string you firmware will still have to do additional pass on it though.

Why is this a problem? Pass over memory is cheap, isn't it?

Quoted 12 lines

With format like above the code
would look like that:

qemu_cfg_read(&n, 1);
arr = alloc(n);
for (i=0; i<n; i++) {
qemu_cfg_read(&l, 1);
arr[i] = zalloc(l+1);
qemu_cfg_read(arr[i], l);
}

Quoted 2 lines

--
Gleb.

At this point I don't care about format.
But I would like one without 1-byte-length limitations, just so we can cover whatever pci can through at us.

Signature · 1 line
MST
#47 lore raw

Re: [PATCHv4 15/15] Pass boot device list to firmware.

Quoted 15 lines · Michael S. Tsirkin
Quoted 14 lines · Gleb Natapov
Quoted 13 lines · Kevin O'Connor
Quoted 11 lines · Gleb Natapov

+/*
+ * This function returns device list as an array in a below format:
+ * +-----+-----+---------------+-----+---------------+--
+ * | n | l1 | devpath1 | l2 | devpath2 | ...
+ * +-----+-----+---------------+-----+---------------+--
+ * where:

+ *   n - a number of devise pathes (one byte)
+ *   l - length of following device path string (one byte)
+ *   devpath - non-null terminated string of length l representing

+ * one device path
+ */

Quoted 7 lines
Quoted 4 lines
Quoted 2 lines

Why not just return a newline separated list that is null terminated?

Doing it like this will needlessly complicate firmware side. How do you know how much memory to allocate before reading device list?

Do a memory scan, count newlines until you reach 0?

To do memory scan you need to read it into memory first. To read it into memory you need to know how much memory to allocate to know how much memory to allocate you meed to do memory scan... Notice pattern here :) Of course you can scan IO space too discarding everything you read first time, but why introduce broken interface in the first place?

Quoted 7 lines
Quoted 4 lines

Doing it
like Blue suggest (have BOOTINDEX_LEN and BOOTINDEX_STRING) solves this. To create nice array from bootindex string you firmware will still have to do additional pass on it though.

Why is this a problem? Pass over memory is cheap, isn't it?

More code, each line of code potentially introduce bug. But I will go with Blue suggestion anyway since we already use it for other things.

Quoted 12 lines
Quoted 12 lines

With format like above the code
would look like that:

qemu_cfg_read(&n, 1);
arr = alloc(n);
for (i=0; i<n; i++) {
qemu_cfg_read(&l, 1);
arr[i] = zalloc(l+1);
qemu_cfg_read(arr[i], l);
}

Quoted 5 lines
Quoted 2 lines

--
Gleb.

At this point I don't care about format.

I do.

Quoted 3 lines

But I would like one without 1-byte-length limitations, just so we can cover whatever pci can through at us.

I agree. 1-byte for one device string may be to limiting. It is still more then 15 PCI bridges on a PC and if you have your pci bus go that deep you are doing something very wrong. But according to spec device name can be 32 byte long and device address may be 64bit physical address and that makes length of one device element to be 50 byte.

--
Gleb.

#48 lore raw

Re: [PATCHv4 15/15] Pass boot device list to firmware.

In reply to Gleb Natapov
Quoted 5 lines · Gleb Natapov
Quoted 2 lines · Kevin O'Connor

Why not just return a newline separated list that is null terminated?

Doing it like this will needlessly complicate firmware side. How do you know how much memory to allocate before reading device list?

My preference would be for the size to be exposed via the QEMU_CFG_FILE_DIR selector. (My preference would be for all objects in fw_cfg to have entries in QEMU_CFG_FILE_DIR describing their size in a reliable manner.)

-Kevin

#49 lore raw

Re: [PATCHv4 15/15] Pass boot device list to firmware.

Quoted 4 lines · Kevin O'Connor
Quoted 3 lines · Gleb Natapov
Quoted 2 lines · Kevin O'Connor

Why not just return a newline separated list that is null terminated?

Quoted 8 lines
Quoted 2 lines

Doing it like this will needlessly complicate firmware side. How do you know how much memory to allocate before reading device list?

My preference would be for the size to be exposed via the QEMU_CFG_FILE_DIR selector. (My preference would be for all objects in fw_cfg to have entries in QEMU_CFG_FILE_DIR describing their size in a reliable manner.)

Will interface suggested by Blue will be good for you? The one with two fw_cfg ids. BOOTINDEX_LEN for len and BOOTINDEX_DATA for device list. I already changed my implementation to this one. Using FILE_DIR requires us to generate synthetic name. Hmm BTW I do not see proper endianness handling in FILE_DIR.

--
Gleb.

#50 lore raw

Re: [PATCHv4 15/15] Pass boot device list to firmware.

Quoted 3 lines · Gleb Natapov

Hmm BTW I do not see proper endianness
handling in FILE_DIR.

That's just me. Everything it OK there with endianness.

--
Gleb.

#51 lore raw

Re: [PATCHv4 15/15] Pass boot device list to firmware.

In reply to Gleb Natapov
Quoted 8 lines · Gleb Natapov
Quoted 7 lines · Kevin O'Connor
Quoted 5 lines · Gleb Natapov
Quoted 2 lines · Kevin O'Connor

Why not just return a newline separated list that is null terminated?

Doing it like this will needlessly complicate firmware side. How do you know how much memory to allocate before reading device list?

Quoted 7 lines
Quoted 5 lines

My preference would be for the size to be exposed via the QEMU_CFG_FILE_DIR selector. (My preference would be for all objects in fw_cfg to have entries in QEMU_CFG_FILE_DIR describing their size in a reliable manner.)

Will interface suggested by Blue will be good for you? The one with two fw_cfg ids. BOOTINDEX_LEN for len and BOOTINDEX_DATA for device list. I

I dislike how different fw_cfg objects pass the length in different ways (eg, QEMU_CFG_E820_TABLE passes length as first 4 bytes). This is a common problem - I'd prefer if we could adopt one uniform way of passing length. I think QEMU_CFG_FILE_DIR solves this problem well.

I also have an ulterior motive here. If the boot order is exposed as a newline separated list via an entry in QEMU_CFG_FILE_DIR, then this becomes free for coreboot users as well. (On coreboot, the boot order could be placed in a "file" in flash with no change to the seabios code.)

-Kevin

#52 lore raw

Re: [PATCHv4 15/15] Pass boot device list to firmware.

Quoted 14 lines · Kevin O'Connor
Quoted 13 lines · Gleb Natapov
Quoted 12 lines · Kevin O'Connor
Quoted 5 lines · Gleb Natapov
Quoted 2 lines · Kevin O'Connor

Why not just return a newline separated list that is null terminated?

Doing it like this will needlessly complicate firmware side. How do you know how much memory to allocate before reading device list?

My preference would be for the size to be exposed via the QEMU_CFG_FILE_DIR selector. (My preference would be for all objects in fw_cfg to have entries in QEMU_CFG_FILE_DIR describing their size in a reliable manner.)

Quoted 8 lines
Quoted 2 lines

Will interface suggested by Blue will be good for you? The one with two fw_cfg ids. BOOTINDEX_LEN for len and BOOTINDEX_DATA for device list. I

I dislike how different fw_cfg objects pass the length in different ways (eg, QEMU_CFG_E820_TABLE passes length as first 4 bytes). This is a common problem - I'd prefer if we could adopt one uniform way of passing length. I think QEMU_CFG_FILE_DIR solves this problem well.

Looking at available fw cfg option I see that _SIZE _DATA is also a common pattern. The problem with QEMU_CFG_FILE_DIR is that we have very little available slots right now. If we a going to require everything to use it we better grow number of available slots considerably now while it is easily done (no option defined above file slots yet).

I personally do not have preferences one way or the other. Blue are you OK with using QEMU_CFG_FILE_DIR?

Quoted 6 lines

I also have an ulterior motive here. If the boot order is exposed as a newline separated list via an entry in QEMU_CFG_FILE_DIR, then this becomes free for coreboot users as well. (On coreboot, the boot order could be placed in a "file" in flash with no change to the seabios code.)

You can define get_boot_order() function and implement it differently for qemu and coreboot. For coreboot it will be one linear. Just call cbfs_copyfile("bootorder"). BTW why newline separation is important?

--
Gleb.

#53 lore raw

Re: [PATCHv4 15/15] Pass boot device list to firmware.

Quoted 10 lines · Gleb Natapov
Quoted 6 lines · Kevin O'Connor

I also have an ulterior motive here. If the boot order is exposed as a newline separated list via an entry in QEMU_CFG_FILE_DIR, then this becomes free for coreboot users as well. (On coreboot, the boot order could be placed in a "file" in flash with no change to the seabios code.)

You can define get_boot_order() function and implement it differently for qemu and coreboot. For coreboot it will be one linear. Just call cbfs_copyfile("bootorder"). BTW why newline separation is important?

Sure, but it'd be nice to just use romfile_copy("bootorder").

Using newline separated just makes it easier for users to vi and/or cat the file.

-Kevin

#54 lore raw

Re: [PATCHv4 15/15] Pass boot device list to firmware.

In reply to Gleb Natapov
Quoted 23 lines · Gleb Natapov
Quoted 14 lines · Kevin O'Connor
Quoted 13 lines · Gleb Natapov
Quoted 12 lines · Kevin O'Connor
Quoted 5 lines · Gleb Natapov
Quoted 2 lines · Kevin O'Connor

Why not just return a newline separated list that is null terminated?

Doing it like this will needlessly complicate firmware side. How do you know how much memory to allocate before reading device list?

My preference would be for the size to be exposed via the QEMU_CFG_FILE_DIR selector.  (My preference would be for all objects in fw_cfg to have entries in QEMU_CFG_FILE_DIR describing their size in a reliable manner.)

Quoted 8 lines
Quoted 2 lines

Will interface suggested by Blue will be good for you? The one with two fw_cfg ids. BOOTINDEX_LEN for len and BOOTINDEX_DATA for device list. I

I dislike how different fw_cfg objects pass the length in different ways (eg, QEMU_CFG_E820_TABLE passes length as first 4 bytes).  This is a common problem - I'd prefer if we could adopt one uniform way of passing length.  I think QEMU_CFG_FILE_DIR solves this problem well.

Quoted 5 lines

Looking at available fw cfg option I see that _SIZE _DATA is also a common pattern. The problem with QEMU_CFG_FILE_DIR is that we have very little available slots right now. If we a going to require everything to use it we better grow number of available slots considerably now while it is easily done (no option defined above file slots yet).

FW_CFG_FILE_DIR seems to be a bit poorly designed. Maybe we should deprecate it and design a more scalable model. There are also string variables passed to BIOS (-prom-env for Sparc/PPC) which could then use this new model instead of NVRAM.

Quoted 2 lines

I personally do not have preferences one way or the other. Blue are you OK with using QEMU_CFG_FILE_DIR?

That would also work.

Quoted 9 lines
Quoted 6 lines

I also have an ulterior motive here.  If the boot order is exposed as a newline separated list via an entry in QEMU_CFG_FILE_DIR, then this becomes free for coreboot users as well.  (On coreboot, the boot order could be placed in a "file" in flash with no change to the seabios code.)

You can define get_boot_order() function and implement it differently for qemu and coreboot. For coreboot it will be one linear. Just call cbfs_copyfile("bootorder"). BTW why newline separation is important?

Newline and zero are both OK since neither can appear inside a valid boot path.