qemu-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Qemu-devel] Re: [PATCH v9 1/8] pci: revise pci command register initial


From: Michael S. Tsirkin
Subject: [Qemu-devel] Re: [PATCH v9 1/8] pci: revise pci command register initialization
Date: Tue, 16 Nov 2010 12:50:19 +0200
User-agent: Mutt/1.5.21 (2010-09-15)

On Tue, Nov 16, 2010 at 05:26:05PM +0900, Isaku Yamahata wrote:
> This patch cleans up command register initialization with
> comments. It also fixes the initialization of io/memory bit of command 
> register.
> Those bits for type 1 device is RW.
> Those bits for type 0 device is
>   RO = 0 if it has no io/memory BAR
>   RW if it has io/memory BAR
> 
> Signed-off-by: Isaku Yamahata <address@hidden>

There's a bug here: you can not assume that device that has
no io BAR claims no io transactions.

Another bug is that migrating from qemu where a bit is writeable to one
where it's RO creates a situation where a RW bit becomes RO, or the
reverse, which might confuse guests.  So we will need a compatibility
flag and set it for old machine types.

> ---
> Changes v8 -> v9
> - patch squash
> ---
>  hw/pci.c |   58 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++-
>  1 files changed, 57 insertions(+), 1 deletions(-)
> 
> diff --git a/hw/pci.c b/hw/pci.c
> index 962886e..2fc8ab1 100644
> --- a/hw/pci.c
> +++ b/hw/pci.c
> @@ -544,8 +544,53 @@ static void pci_init_wmask(PCIDevice *dev)
>  
>      dev->wmask[PCI_CACHE_LINE_SIZE] = 0xff;
>      dev->wmask[PCI_INTERRUPT_LINE] = 0xff;
> +
> +    /*
> +     * bit 0: PCI_COMMAND_IO
> +     *        type 0: if IO BAR is used, RW
> +     *                This is handled by pci_register_bar()
> +     *        type 1: RW:
> +     *                This is fixed by pci_init_wmask_bridge()
> +     * bit 1: PCI_COMMAND_MEMORY
> +     *        type 0: if IO BAR is used, RW
> +     *                This is handled by pci_register_bar()
> +     *        type 1: RW
> +     *                This is fixed by pci_init_wmask_bridge()
> +     * bit 2: PCI_COMMAND_MASTER
> +     *        type 0: RW if bus master
> +     *        type 1: RW
> +     * bit 3: PCI_COMMAND_SPECIAL
> +     *        RO=0, optionally RW: Such device should set this bit itself
> +     * bit 4: PCI_COMMAND_INVALIDATE
> +     *        RO=0, optionally RW: Such device should set this bit itself
> +     * bit 5: PCI_COMMAND_VGA_PALETTE
> +     *        RO=0, optionally RW: Such device should set this bit itself
> +     * bit 6: PCI_COMMAND_PARITY
> +     *        RW with exceptions: Such device should clear this bit itself
> +     *        Given that qemu doesn't emulate pci bus cycles, so that there
> +     *        is no place to generate parity error. So just making this
> +     *        register RW is okay because there is no place which refers
> +     *        this bit.
> +     *        TODO: When device assignment tried to inject PERR# into qemu,
> +     *              some extra work would be needed.
> +     * bit 7: PCI_COMMAND_WAIT: reserved (PCI 3.0)
> +     *        RO=0
> +     * bit 8: PCI_COMMAND_SERR
> +     *        RW with exceptions: Such device should clear this bit itself
> +     *        Given that qemu doesn't emulate pci bus cycles, so that there
> +     *        is no place to generate system error. So just making this
> +     *        register RW is okay because there is no place which refers
> +     *        this bit.
> +     *        TODO: When device assignment tried to inject SERR# into qemu,
> +     *              some extra work would be needed.
> +     * bit 9: PCI_COMMAND_FAST_BACK
> +     *        RO=0, optionally RW: Such device should set this bit itself
> +     * bit 10: PCI_COMMAND_INTX_DISABLE
> +     *         RW
> +     * bit 11-15: reserved
> +     */

Let's document non-obvious things, like maybe COMMAND_PARITY/COMMAND_SERR.
I don't cherish writing each bit up in two places.

>      pci_set_word(dev->wmask + PCI_COMMAND,
> -                 PCI_COMMAND_IO | PCI_COMMAND_MEMORY | PCI_COMMAND_MASTER |
> +                 PCI_COMMAND_MASTER | PCI_COMMAND_PARITY | PCI_COMMAND_SERR |
>                   PCI_COMMAND_INTX_DISABLE);
>  
>      memset(dev->wmask + PCI_CONFIG_HEADER_SIZE, 0xff,
> @@ -554,6 +599,9 @@ static void pci_init_wmask(PCIDevice *dev)
>  
>  static void pci_init_wmask_bridge(PCIDevice *d)
>  {
> +    pci_word_test_and_set_mask(d->wmask + PCI_COMMAND,
> +                               PCI_COMMAND_IO | PCI_COMMAND_MEMORY);
> +
>      /* PCI_PRIMARY_BUS, PCI_SECONDARY_BUS, PCI_SUBORDINATE_BUS and
>         PCI_SEC_LETENCY_TIMER */
>      memset(d->wmask + PCI_PRIMARY_BUS, 0xff, 4);
> @@ -791,6 +839,14 @@ void pci_register_bar(PCIDevice *pci_dev, int region_num,
>      if (region_num == PCI_ROM_SLOT) {
>          /* ROM enable bit is writeable */
>          wmask |= PCI_ROM_ADDRESS_ENABLE;
> +    } else {
> +        if (r->type & PCI_BASE_ADDRESS_SPACE_IO) {
> +            pci_word_test_and_set_mask(pci_dev->wmask + PCI_COMMAND,
> +                                       PCI_COMMAND_IO);
> +        } else {
> +            pci_word_test_and_set_mask(pci_dev->wmask + PCI_COMMAND,
> +                                       PCI_COMMAND_MEMORY);
> +        }
>      }
>      pci_set_long(pci_dev->config + addr, type);
>      if (!(r->type & PCI_BASE_ADDRESS_SPACE_IO) &&
> -- 
> 1.7.1.1



reply via email to

[Prev in Thread] Current Thread [Next in Thread]