qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: "Michael S. Tsirkin" <mst@redhat.com>
To: Isaku Yamahata <yamahata@valinux.co.jp>
Cc: qemu-devel@nongnu.org
Subject: [Qemu-devel] Re: [PATCH V5 19/29] pci: split out ioport address parsing from pci configuration access logic.
Date: Tue, 13 Oct 2009 16:49:26 +0200	[thread overview]
Message-ID: <20091013144926.GC17895@redhat.com> (raw)
In-Reply-To: <20091013141411.GG2306%yamahata@valinux.co.jp>

On Tue, Oct 13, 2009 at 11:14:11PM +0900, Isaku Yamahata wrote:
> On Sat, Oct 10, 2009 at 09:45:00PM +0200, Michael S. Tsirkin wrote:
> > On Fri, Oct 09, 2009 at 03:28:52PM +0900, Isaku Yamahata wrote:
> > > Split out pci_data_{write, read} into two part, one is the code which
> > > convert ioport address into pci device and the other is the code
> > > which access PCI configuration space.
> > > 
> > > Later PCI express code will use the access code.
> > > PCI express addressing scheme is different so that the address
> > > must be parsed differently.
> > > 
> > > PCI:
> > >  0- 7 bit: offset in the configuration space (256bytes)
> > >  7-15 bit: devfn
> > > 16-24 bit: bus
> > > 
> > > PCI express:
> > >  0-11 bit: offset in the configuration space (4KBytes)
> > > 12-19 bit: devfn
> > > 20-28 bit: bus
> > > 
> > > Signed-off-by: Isaku Yamahata <yamahata@valinux.co.jp>
> > 
> > So I think I said this already: the only useful thing that
> > pci_data_read/write does is address conversion which is different for
> > express.  PCI Express should do its own thing and then call
> > pci_dev->config_read / pci_dev->config_write directly.
> > 
> > I think this will even make it possible to put express code in a
> > separate file from pci.c.
> 
> Yes, I did it. And with the next patch, they will moved to pci_host.c
> too as you noticed it.
> Hmm, you seem to be saying that pci_dev_{write, read}_config() aren't
> needed, however they are used by both pci_host.c and pcie_host.c
> as follows.

Yes, but if you look at them they do not do antyhing useful,
just call dev->config_write. So open-code and be done with it.

> So introducing pci_dev_{write, read}_config() makes sense.
> 
> 
> void pci_data_write(PCIBus *s, uint32_t addr, uint32_t val, int len)
> {
>     pci_dev_write_config(pci_addr_to_dev(s, addr), pci_addr_to_config(addr),
>                          val, len);
> }
> It parses pci address and pass it to lower function.
> 
> 
> With the patch of "pci: pcie host and mmcfg support.",
> pcie counter part is as follows.
> static void pcie_mmcfg_data_write(PCIBus *s,
>                                   uint32_t mmcfg_addr, uint32_t val, int len)
> {
>     pci_dev_write_config(pcie_mmcfg_addr_to_dev(s, mmcfg_addr),
>                          PCIE_MMCFG_CONFOFFSET(mmcfg_addr),
>                          val, len);
> }
> It parses pcie mmconfig address and pass it to lower function.

Yes but the lower function does nothing, so just opencode it.

So that will be
void pci_data_write(PCIBus *s,
		    uint32_t addr, uint32_t val, int len)
{
	PCIDevice *dev = pci_addr_to_dev(s, addr);
	if (!dev)
		return;
	dev->config_write(dev, pci_addr_to_config(addr), val, len);
}

and separately
static void pcie_mmcfg_data_write(PCIBus *s,
                                  uint32_t mmcfg_addr, uint32_t val, int len)
{
	PCIDevice *dev = pcie_mmcfg_addr_to_dev(s, mmcfg_addr);
	if (!dev)
		return;
	dev->config_write(dev, PCIE_MMCFG_CONFOFFSET(mmcfg_addr), val, len);
}

> > 
> > > ---
> > >  hw/pci.c |   80 ++++++++++++++++++++++++++++++-------------------------------
> > >  1 files changed, 39 insertions(+), 41 deletions(-)
> > > 
> > > diff --git a/hw/pci.c b/hw/pci.c
> > > index 0021c96..d472b58 100644
> > > --- a/hw/pci.c
> > > +++ b/hw/pci.c
> > > @@ -654,46 +654,24 @@ void pci_default_write_config(PCIDevice *d, uint32_t addr, uint32_t val, int l)
> > >          pci_update_mappings(d);
> > >  }
> > >  
> > > -static inline PCIDevice *pci_addr_to_dev(PCIBus *bus, uint32_t addr)
> > > -{
> > > -    uint8_t bus_num = (addr >> 16) & 0xff;
> > > -    uint8_t devfn = (addr >> 8) & 0xff;
> > > -    return pci_find_device(bus, bus_num, PCI_SLOT(devfn), PCI_FUNC(devfn));
> > > -}
> > > -
> > > -static inline int pci_addr_to_config(uint32_t addr)
> > > -{
> > > -    return addr & 0xff;
> > > -}
> > > -
> > > -void pci_data_write(void *opaque, uint32_t addr, uint32_t val, int len)
> > > +static void pci_dev_write_config(PCIDevice *pci_dev,
> > > +                                 uint32_t config_addr, uint32_t val, int len)
> > >  {
> > > -    PCIBus *s = opaque;
> > > -    PCIDevice *pci_dev;
> > > -    int config_addr;
> > > -
> > > -#if 0
> > > -    PCI_DPRINTF("pci_data_write: addr=%08x val=%08x len=%d\n",
> > > -                addr, val, len);
> > > -#endif
> > > -    pci_dev = pci_addr_to_dev(s, addr);
> > > +    assert(len == 1 || len == 2 || len == 4);
> > >      if (!pci_dev)
> > >          return;
> > > -    config_addr = addr & 0xff;
> > > -    config_addr = pci_addr_to_config(addr);
> > > -    PCI_DPRINTF("pci_config_write: %s: addr=%02x val=%08x len=%d\n",
> > > -                pci_dev->name, config_addr, val, len);
> > > +
> > > +    PCI_DPRINTF("%s: %s: addr=%02"PRIx32" val=%08"PRI32x" len=%d\n",
> > > +                __func__, pci_dev->name, config_addr, val, len);
> > >      pci_dev->config_write(pci_dev, config_addr, val, len);
> > >  }
> > >  
> > > -uint32_t pci_data_read(void *opaque, uint32_t addr, int len)
> > > +static uint32_t pci_dev_read_config(PCIDevice *pci_dev,
> > > +                                    uint32_t config_addr, int len)
> > >  {
> > > -    PCIBus *s = opaque;
> > > -    PCIDevice *pci_dev;
> > > -    int config_addr;
> > >      uint32_t val;
> > >  
> > > -    pci_dev = pci_addr_to_dev(s, addr);
> > > +    assert(len == 1 || len == 2 || len == 4);
> > >      if (!pci_dev) {
> > >          switch(len) {
> > >          case 1:
> > > @@ -707,20 +685,40 @@ uint32_t pci_data_read(void *opaque, uint32_t addr, int len)
> > >              val = 0xffffffff;
> > >              break;
> > >          }
> > > -        goto the_end;
> > > +    } else {
> > > +        val = pci_dev->config_read(pci_dev, config_addr, len);
> > > +        PCI_DPRINTF("%s: %s: addr=%02"PRIx32" val=%08"PRIx32" len=%d\n",
> > > +                    __func__, pci_dev->name, config_addr, val, len);
> > >      }
> > > -    config_addr = pci_addr_to_config(addr);
> > > -    val = pci_dev->config_read(pci_dev, config_addr, len);
> > > -    PCI_DPRINTF("pci_config_read: %s: addr=%02x val=%08x len=%d\n",
> > > -                pci_dev->name, config_addr, val, len);
> > > - the_end:
> > > -#if 0
> > > -    PCI_DPRINTF("pci_data_read: addr=%08x val=%08x len=%d\n",
> > > -                addr, val, len);
> > > -#endif
> > >      return val;
> > >  }
> > >  
> > > +static inline PCIDevice *pci_addr_to_dev(PCIBus *bus, uint32_t addr)
> > > +{
> > > +    uint8_t bus_num = (addr >> 16) & 0xff;
> > > +    uint8_t devfn = (addr >> 8) & 0xff;
> > > +    return pci_find_device(bus, bus_num, PCI_SLOT(devfn), PCI_FUNC(devfn));
> > > +}
> > > +
> > > +static inline int pci_addr_to_config(uint32_t addr)
> > > +{
> > > +    return addr & 0xff;
> > > +}
> > > +
> > > +void pci_data_write(void *opaque, uint32_t addr, uint32_t val, int len)
> > > +{
> > > +    PCIBus *s = opaque;
> > > +    pci_dev_write_config(pci_addr_to_dev(s, addr), pci_addr_to_config(addr),
> > > +                         val, len);
> > > +}
> > > +
> > > +uint32_t pci_data_read(void *opaque, uint32_t addr, int len)
> > > +{
> > > +    PCIBus *s = opaque;
> > > +    return pci_dev_read_config(pci_addr_to_dev(s, addr),
> > > +                               pci_addr_to_config(addr), len);
> > > +}
> > > +
> > >  /***********************************************************/
> > >  /* generic PCI irq support */
> > >  
> > 
> 
> -- 
> yamahata

  reply	other threads:[~2009-10-13 14:51 UTC|newest]

Thread overview: 68+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2009-10-09  6:28 [Qemu-devel] [PATCH V5 00/29] pci: various pci clean up and pci express support Isaku Yamahata
2009-10-09  6:28 ` [Qemu-devel] [PATCH V5 01/29] pci: fix PCI_DPRINTF() wrt variadic macro Isaku Yamahata
2009-10-09  6:28 ` [Qemu-devel] [PATCH V5 02/29] pci: introduce constant PCI_NUM_PINS for the number of interrupt pins, 4 Isaku Yamahata
2009-10-09  6:28 ` [Qemu-devel] [PATCH V5 03/29] pci: use PCI_SLOT() and PCI_FUNC() Isaku Yamahata
2009-10-09  6:28 ` [Qemu-devel] [PATCH V5 04/29] pci: define a constant to represent a unmapped bar and use it Isaku Yamahata
2009-10-09  6:28 ` [Qemu-devel] [PATCH V5 05/29] pci: helper functions to access PCIDevice::config Isaku Yamahata
2009-10-09  6:28 ` [Qemu-devel] [PATCH V5 06/29] pci: use helper functions to access pci config space Isaku Yamahata
2009-10-09  6:28 ` [Qemu-devel] [PATCH V5 07/29] pci/bridge: clean up of pci_bridge_initfn() Isaku Yamahata
2009-10-09  6:53   ` [Qemu-devel] " Michael S. Tsirkin
2009-10-13 13:20     ` Isaku Yamahata
2009-10-13 14:17       ` Michael S. Tsirkin
2009-10-13 15:12       ` Blue Swirl
2009-10-13 15:26         ` Michael S. Tsirkin
2009-10-13 16:32           ` Blue Swirl
2009-10-09  6:54   ` Michael S. Tsirkin
2009-10-09  6:28 ` [Qemu-devel] [PATCH V5 08/29] pci: s/PCI_ADDRESS_SPACE_/PCI_BASE_ADDRESS_SPACE_/ to match pci_regs.h Isaku Yamahata
2009-10-09  6:57   ` [Qemu-devel] " Michael S. Tsirkin
2009-10-13 13:21     ` Isaku Yamahata
2009-10-09  6:28 ` [Qemu-devel] [PATCH V5 09/29] pci: clean up of pci_default_read_config Isaku Yamahata
2009-10-09  6:50   ` [Qemu-devel] " Michael S. Tsirkin
2009-10-09  6:58   ` Michael S. Tsirkin
2009-10-09  6:28 ` [Qemu-devel] [PATCH V5 10/29] pci: make pci_bar() aware of header type 1 Isaku Yamahata
2009-10-09  7:06   ` [Qemu-devel] " Michael S. Tsirkin
2009-10-10 19:29   ` Michael S. Tsirkin
2009-10-09  6:28 ` [Qemu-devel] [PATCH V5 11/29] pci_host.h: move functions in pci_host.h into .c file Isaku Yamahata
2009-10-09  6:28 ` [Qemu-devel] [PATCH V5 12/29] pci_host: consolidate pci config address access Isaku Yamahata
2009-10-09  6:28 ` [Qemu-devel] [PATCH V5 13/29] pci: introduce pcibus_t to represent pci bus address/size instead of uint32_t Isaku Yamahata
2009-10-09  6:28 ` [Qemu-devel] [PATCH V5 14/29] pci: introduce FMT_PCIBUS for printf format for pcibus_t Isaku Yamahata
2009-10-10 19:32   ` [Qemu-devel] " Michael S. Tsirkin
2009-10-09  6:28 ` [Qemu-devel] [PATCH V5 15/29] pci: typedef pcibus_t as uint64_t instead of uint32_t Isaku Yamahata
2009-10-11 10:43   ` [Qemu-devel] " Michael S. Tsirkin
2009-10-13 13:31     ` Isaku Yamahata
2009-10-13 14:39       ` Michael S. Tsirkin
2009-10-14  4:35         ` Isaku Yamahata
2009-10-14  8:55           ` Michael S. Tsirkin
2009-10-09  6:28 ` [Qemu-devel] [PATCH V5 16/29] pci: 64bit bar support Isaku Yamahata
2009-10-10 19:39   ` [Qemu-devel] " Michael S. Tsirkin
2009-10-13 13:52     ` Isaku Yamahata
2009-10-13 15:00       ` Michael S. Tsirkin
2009-10-09  6:28 ` [Qemu-devel] [PATCH V5 17/29] pci: make pci configuration transaction more accurate Isaku Yamahata
2009-10-09 12:52   ` [Qemu-devel] " Michael S. Tsirkin
2009-10-09  6:28 ` [Qemu-devel] [PATCH V5 18/29] pci: factor out the conversion logic from io port address into pci device Isaku Yamahata
2009-10-10 19:41   ` [Qemu-devel] " Michael S. Tsirkin
2009-10-09  6:28 ` [Qemu-devel] [PATCH V5 19/29] pci: split out ioport address parsing from pci configuration access logic Isaku Yamahata
2009-10-10 19:45   ` [Qemu-devel] " Michael S. Tsirkin
2009-10-13 14:14     ` Isaku Yamahata
2009-10-13 14:49       ` Michael S. Tsirkin [this message]
2009-10-09  6:28 ` [Qemu-devel] [PATCH V5 20/29] pci: move pci host stuff from pci.c to pci_host.c Isaku Yamahata
2009-10-10 19:46   ` [Qemu-devel] " Michael S. Tsirkin
2009-10-09  6:28 ` [Qemu-devel] [PATCH V5 21/29] pci_host: change the signature of pci_data_{read, write} Isaku Yamahata
2009-10-09 12:02   ` [Qemu-devel] " Michael S. Tsirkin
2009-10-09  6:28 ` [Qemu-devel] [PATCH V5 22/29] vmstate: add VMSTATE_ARRAY_POINTER for pointer to array Isaku Yamahata
2009-10-11 10:37   ` [Qemu-devel] " Michael S. Tsirkin
2009-10-09  6:28 ` [Qemu-devel] [PATCH V5 23/29] pci: pcie host and mmcfg support Isaku Yamahata
2009-10-11 10:26   ` [Qemu-devel] " Michael S. Tsirkin
2009-10-09  6:28 ` [Qemu-devel] [PATCH V5 24/29] pci: fix pci_default_write_config() Isaku Yamahata
2009-10-09  6:28 ` [Qemu-devel] [PATCH V5 25/29] pci: add helper functions for pci config write function Isaku Yamahata
2009-10-10 19:58   ` [Qemu-devel] " Michael S. Tsirkin
2009-10-09  6:28 ` [Qemu-devel] [PATCH V5 26/29] pci: use helper function in pci_default_write_config() Isaku Yamahata
2009-10-09  6:29 ` [Qemu-devel] [PATCH V5 27/29] pci/bridge: don't update bar mapping when bar2-5 is changed Isaku Yamahata
2009-10-09 10:35   ` [Qemu-devel] " Michael S. Tsirkin
2009-10-09  6:29 ` [Qemu-devel] [PATCH V5 28/29] pci: initialize pci config headers depending it pci header type Isaku Yamahata
2009-10-09 10:42   ` [Qemu-devel] " Michael S. Tsirkin
2009-10-13 14:31     ` Isaku Yamahata
2009-10-13 14:52       ` Michael S. Tsirkin
2009-10-13 15:06       ` Michael S. Tsirkin
2009-10-09  6:29 ` [Qemu-devel] [PATCH V5 29/29] pci/monitor: print out bridge's filtering values and so on Isaku Yamahata
2009-10-10 20:05   ` [Qemu-devel] " Michael S. Tsirkin

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20091013144926.GC17895@redhat.com \
    --to=mst@redhat.com \
    --cc=qemu-devel@nongnu.org \
    --cc=yamahata@valinux.co.jp \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).