From: David Gibson <david@gibson.dropbear.id.au>
To: Laurent Vivier <lvivier@redhat.com>
Cc: pbonzini@redhat.com, qemu-devel@nongnu.org, qemu-ppc@nongnu.org,
agraf@suse.de, stefanha@redhat.com, mst@redhat.com,
aik@ozlabs.ru, mdroth@linux.vnet.ibm.com, groug@kaod.org,
thuth@redhat.com
Subject: Re: [Qemu-devel] [PATCH 3/8] libqos: Move BAR assignment to common code
Date: Wed, 19 Oct 2016 14:07:15 +1100 [thread overview]
Message-ID: <20161019030715.GG11140@umbus.fritz.box> (raw)
In-Reply-To: <1dc5eb87-ebc5-40db-25ef-5e61bd488a7e@redhat.com>
[-- Attachment #1: Type: text/plain, Size: 9487 bytes --]
On Tue, Oct 18, 2016 at 05:00:08PM +0200, Laurent Vivier wrote:
>
>
> On 18/10/2016 12:52, David Gibson wrote:
> > The PCI backends in libqos each supply an iomap() and iounmap() function
> > which is used to set up a specified PCI BAR. But PCI BAR allocation takes
> > place entirely within PCI space, so doesn't really need per-backend
> > versions. For example, Linux includes generic BAR allocation code used on
> > platforms where that isn't done by firmware.
> >
> > This patch merges the BAR allocation from the two existing backends into a
> > single simplified copy. The back ends just need to set up some parameters
> > describing the window of PCI IO and PCI memory addresses which are
> > available for allocation. Like both the existing versions the new one uses
> > a simple bump allocator.
> >
> > Note that (again like the existing versions) this doesn't really handle
> > 64-bit memory BARs properly. It is actually used for such a BAR by the
> > ivshmem test, and apparently the 32-bit MMIO BAR logic is close enough to
> > work, as long as the BAR isn't too big. Fixing that to properly handle
> > 64-bit BAR allocation is a problem for another time.
> >
> > Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
> > ---
> > tests/libqos/pci-pc.c | 79 ++--------------------------------------------
> > tests/libqos/pci-spapr.c | 81 ++----------------------------------------------
> > tests/libqos/pci.c | 56 +++++++++++++++++++++++++++++++--
> > tests/libqos/pci.h | 7 ++---
> > 4 files changed, 63 insertions(+), 160 deletions(-)
> >
> > diff --git a/tests/libqos/pci-pc.c b/tests/libqos/pci-pc.c
> > index 51dff8a..b087d13 100644
> > --- a/tests/libqos/pci-pc.c
> > +++ b/tests/libqos/pci-pc.c
> > @@ -17,7 +17,6 @@
> > #include "hw/pci/pci_regs.h"
> >
> > #include "qemu-common.h"
> > -#include "qemu/host-utils.h"
> >
> >
> > #define ACPI_PCIHP_ADDR 0xae00
> > @@ -132,71 +131,6 @@ static void qpci_pc_config_writel(QPCIBus *bus, int devfn, uint8_t offset, uint3
> > outl(0xcfc, value);
> > }
> >
> > -static void *qpci_pc_iomap(QPCIBus *bus, QPCIDevice *dev, int barno, uint64_t *sizeptr)
> > -{
> > - QPCIBusPC *s = container_of(bus, QPCIBusPC, bus);
> > - static const int bar_reg_map[] = {
> > - PCI_BASE_ADDRESS_0, PCI_BASE_ADDRESS_1, PCI_BASE_ADDRESS_2,
> > - PCI_BASE_ADDRESS_3, PCI_BASE_ADDRESS_4, PCI_BASE_ADDRESS_5,
> > - };
> > - int bar_reg;
> > - uint32_t addr;
> > - uint64_t size;
> > - uint32_t io_type;
> > -
> > - g_assert(barno >= 0 && barno <= 5);
> > - bar_reg = bar_reg_map[barno];
> > -
> > - qpci_config_writel(dev, bar_reg, 0xFFFFFFFF);
> > - addr = qpci_config_readl(dev, bar_reg);
> > -
> > - io_type = addr & PCI_BASE_ADDRESS_SPACE;
> > - if (io_type == PCI_BASE_ADDRESS_SPACE_IO) {
> > - addr &= PCI_BASE_ADDRESS_IO_MASK;
> > - } else {
> > - addr &= PCI_BASE_ADDRESS_MEM_MASK;
> > - }
> > -
> > - size = (1ULL << ctzl(addr));
> > - if (size == 0) {
> > - return NULL;
> > - }
> > - if (sizeptr) {
> > - *sizeptr = size;
> > - }
> > -
> > - if (io_type == PCI_BASE_ADDRESS_SPACE_IO) {
> > - uint16_t loc;
> > -
> > - g_assert(QEMU_ALIGN_UP(s->pci_iohole_alloc, size) + size
> > - <= s->pci_iohole_size);
> > - s->pci_iohole_alloc = QEMU_ALIGN_UP(s->pci_iohole_alloc, size);
> > - loc = s->pci_iohole_start + s->pci_iohole_alloc;
> > - s->pci_iohole_alloc += size;
> > -
> > - qpci_config_writel(dev, bar_reg, loc | PCI_BASE_ADDRESS_SPACE_IO);
> > -
> > - return (void *)(intptr_t)loc;
> > - } else {
> > - uint64_t loc;
> > -
> > - g_assert(QEMU_ALIGN_UP(s->pci_hole_alloc, size) + size
> > - <= s->pci_hole_size);
> > - s->pci_hole_alloc = QEMU_ALIGN_UP(s->pci_hole_alloc, size);
> > - loc = s->pci_hole_start + s->pci_hole_alloc;
> > - s->pci_hole_alloc += size;
> > -
> > - qpci_config_writel(dev, bar_reg, loc);
> > -
> > - return (void *)(intptr_t)loc;
> > - }
> > -}
> > -
> > -static void qpci_pc_iounmap(QPCIBus *bus, void *data)
> > -{
> > - /* FIXME */
> > -}
> > -
> > QPCIBus *qpci_init_pc(QGuestAllocator *alloc)
> > {
> > QPCIBusPC *ret;
> > @@ -227,16 +161,9 @@ QPCIBus *qpci_init_pc(QGuestAllocator *alloc)
> > ret->bus.config_writew = qpci_pc_config_writew;
> > ret->bus.config_writel = qpci_pc_config_writel;
> >
> > - ret->bus.iomap = qpci_pc_iomap;
> > - ret->bus.iounmap = qpci_pc_iounmap;
> > -
> > - ret->pci_hole_start = 0xE0000000;
> > - ret->pci_hole_size = 0x20000000;
> > - ret->pci_hole_alloc = 0;
> > -
> > - ret->pci_iohole_start = 0xc000;
> > - ret->pci_iohole_size = 0x4000;
> > - ret->pci_iohole_alloc = 0;
>
> I think you can remove all these fields (pci_hole_.., pci_iohole_...)
> from QPCIBusPC.
>
> > + ret->bus.pio_alloc_ptr = 0xc000;
> > + ret->bus.mmio_alloc_ptr = 0xE0000000;
> > + ret->bus.mmio_limit = 0x100000000ULL;
> >
> > return &ret->bus;
> > }
> > diff --git a/tests/libqos/pci-spapr.c b/tests/libqos/pci-spapr.c
> > index 2d26a94..9a18d8a 100644
> > --- a/tests/libqos/pci-spapr.c
> > +++ b/tests/libqos/pci-spapr.c
> > @@ -167,72 +167,6 @@ static void qpci_spapr_config_writel(QPCIBus *bus, int devfn, uint8_t offset,
> > qrtas_ibm_write_pci_config(s->alloc, s->buid, config_addr, 4, value);
> > }
> >
> > -static void *qpci_spapr_iomap(QPCIBus *bus, QPCIDevice *dev, int barno,
> > - uint64_t *sizeptr)
> > -{
> > - QPCIBusSPAPR *s = container_of(bus, QPCIBusSPAPR, bus);
> > - static const int bar_reg_map[] = {
> > - PCI_BASE_ADDRESS_0, PCI_BASE_ADDRESS_1, PCI_BASE_ADDRESS_2,
> > - PCI_BASE_ADDRESS_3, PCI_BASE_ADDRESS_4, PCI_BASE_ADDRESS_5,
> > - };
> > - int bar_reg;
> > - uint32_t addr;
> > - uint64_t size;
> > - uint32_t io_type;
> > -
> > - g_assert(barno >= 0 && barno <= 5);
> > - bar_reg = bar_reg_map[barno];
> > -
> > - qpci_config_writel(dev, bar_reg, 0xFFFFFFFF);
> > - addr = qpci_config_readl(dev, bar_reg);
> > -
> > - io_type = addr & PCI_BASE_ADDRESS_SPACE;
> > - if (io_type == PCI_BASE_ADDRESS_SPACE_IO) {
> > - addr &= PCI_BASE_ADDRESS_IO_MASK;
> > - } else {
> > - addr &= PCI_BASE_ADDRESS_MEM_MASK;
> > - }
> > -
> > - size = (1ULL << ctzl(addr));
> > - if (size == 0) {
> > - return NULL;
> > - }
> > - if (sizeptr) {
> > - *sizeptr = size;
> > - }
> > -
> > - if (io_type == PCI_BASE_ADDRESS_SPACE_IO) {
> > - uint16_t loc;
> > -
> > - g_assert(QEMU_ALIGN_UP(s->pci_iohole_alloc, size) + size
> > - <= s->pci_iohole_size);
> > - s->pci_iohole_alloc = QEMU_ALIGN_UP(s->pci_iohole_alloc, size);
> > - loc = s->pci_iohole_start + s->pci_iohole_alloc;
> > - s->pci_iohole_alloc += size;
> > -
> > - qpci_config_writel(dev, bar_reg, loc | PCI_BASE_ADDRESS_SPACE_IO);
> > -
> > - return (void *)(unsigned long)loc;
> > - } else {
> > - uint64_t loc;
> > -
> > - g_assert(QEMU_ALIGN_UP(s->pci_hole_alloc, size) + size
> > - <= s->pci_hole_size);
> > - s->pci_hole_alloc = QEMU_ALIGN_UP(s->pci_hole_alloc, size);
> > - loc = s->pci_hole_start + s->pci_hole_alloc;
> > - s->pci_hole_alloc += size;
> > -
> > - qpci_config_writel(dev, bar_reg, loc);
> > -
> > - return (void *)(unsigned long)loc;
> > - }
> > -}
> > -
> > -static void qpci_spapr_iounmap(QPCIBus *bus, void *data)
> > -{
> > - /* FIXME */
> > -}
> > -
> > #define SPAPR_PCI_BASE (1ULL << 45)
> >
> > #define SPAPR_PCI_MMIO32_WIN_SIZE 0x80000000 /* 2 GiB */
> > @@ -270,9 +204,6 @@ QPCIBus *qpci_init_spapr(QGuestAllocator *alloc)
> > ret->bus.config_writew = qpci_spapr_config_writew;
> > ret->bus.config_writel = qpci_spapr_config_writel;
> >
> > - ret->bus.iomap = qpci_spapr_iomap;
> > - ret->bus.iounmap = qpci_spapr_iounmap;
> > -
> > /* FIXME: We assume the default location of the PHB for now.
> > * Ideally we'd parse the device tree deposited in the guest to
> > * get the window locations */
> > @@ -287,15 +218,9 @@ QPCIBus *qpci_init_spapr(QGuestAllocator *alloc)
> > ret->mmio32.pci_base = 0x80000000; /* 2 GiB */
> > ret->mmio32.size = SPAPR_PCI_MMIO32_WIN_SIZE;
> >
> > - ret->pci_hole_start = 0xC0000000;
> > - ret->pci_hole_size =
> > - ret->mmio32.pci_base + ret->mmio32.size - ret->pci_hole_start;
> > - ret->pci_hole_alloc = 0;
> > -
> > - ret->pci_iohole_start = 0xc000;
> > - ret->pci_iohole_size =
> > - ret->pio.pci_base + ret->pio.size - ret->pci_iohole_start;
> > - ret->pci_iohole_alloc = 0;
>
> I think you can remove all these fields (pci_hole_.., pci_iohole_...)
> from QPCIBusSPAPR.
Ah, yes, meant to remove them from both PC and spapr, but forgot.
Done for the next spin.
--
David Gibson | I'll have my music baroque, and my code
david AT gibson.dropbear.id.au | minimalist, thank you. NOT _the_ _other_
| _way_ _around_!
http://www.ozlabs.org/~dgibson
[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 819 bytes --]
next prev parent reply other threads:[~2016-10-19 4:49 UTC|newest]
Thread overview: 28+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-10-18 10:52 [Qemu-devel] [PATCH 0/8] Cleanups to qtest PCI handling David Gibson
2016-10-18 10:52 ` [Qemu-devel] [PATCH 1/8] libqos: Give qvirtio_config_read*() consistent semantics David Gibson
2016-10-18 12:41 ` Laurent Vivier
2016-10-18 13:27 ` Greg Kurz
2016-10-19 2:41 ` David Gibson
2016-10-18 10:52 ` [Qemu-devel] [PATCH 2/8] libqos: Handle PCI IO de-multiplexing in common code David Gibson
2016-10-18 13:28 ` Laurent Vivier
2016-10-19 2:59 ` David Gibson
2016-10-18 10:52 ` [Qemu-devel] [PATCH 3/8] libqos: Move BAR assignment to " David Gibson
2016-10-18 15:00 ` Laurent Vivier
2016-10-19 3:07 ` David Gibson [this message]
2016-10-18 10:52 ` [Qemu-devel] [PATCH 4/8] tests: Better handle legacy IO addresses in tco-test David Gibson
2016-10-18 15:14 ` Laurent Vivier
2016-10-18 16:28 ` Laurent Vivier
2016-10-19 3:19 ` David Gibson
2016-10-19 3:09 ` David Gibson
2016-10-18 10:52 ` [Qemu-devel] [PATCH 5/8] libqos: Add streaming accessors for PCI MMIO David Gibson
2016-10-18 15:17 ` Laurent Vivier
2016-10-18 10:52 ` [Qemu-devel] [PATCH 6/8] libqos: Implement mmio accessors in terms of mem{read, write} David Gibson
2016-10-18 16:04 ` Laurent Vivier
2016-10-18 10:52 ` [Qemu-devel] [PATCH 7/8] tests: Use qpci_mem{read, write} in ivshmem-test David Gibson
2016-10-18 16:14 ` Laurent Vivier
2016-10-19 4:13 ` David Gibson
2016-10-18 10:52 ` [Qemu-devel] [PATCH 8/8] libqos: Change PCI accessors to take opaque BAR handle David Gibson
2016-10-18 16:48 ` Laurent Vivier
2016-10-19 4:06 ` David Gibson
2016-10-18 11:56 ` [Qemu-devel] [PATCH 0/8] Cleanups to qtest PCI handling Laurent Vivier
2016-10-19 1:11 ` David Gibson
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=20161019030715.GG11140@umbus.fritz.box \
--to=david@gibson.dropbear.id.au \
--cc=agraf@suse.de \
--cc=aik@ozlabs.ru \
--cc=groug@kaod.org \
--cc=lvivier@redhat.com \
--cc=mdroth@linux.vnet.ibm.com \
--cc=mst@redhat.com \
--cc=pbonzini@redhat.com \
--cc=qemu-devel@nongnu.org \
--cc=qemu-ppc@nongnu.org \
--cc=stefanha@redhat.com \
--cc=thuth@redhat.com \
/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).