From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:36646) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1aW92f-00014P-R8 for qemu-devel@nongnu.org; Wed, 17 Feb 2016 15:49:46 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1aW92b-00026r-Q0 for qemu-devel@nongnu.org; Wed, 17 Feb 2016 15:49:45 -0500 Received: from mx1.redhat.com ([209.132.183.28]:34433) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1aW92b-00026l-Ht for qemu-devel@nongnu.org; Wed, 17 Feb 2016 15:49:41 -0500 Received: from int-mx09.intmail.prod.int.phx2.redhat.com (int-mx09.intmail.prod.int.phx2.redhat.com [10.5.11.22]) by mx1.redhat.com (Postfix) with ESMTPS id 1A18E7AE89 for ; Wed, 17 Feb 2016 20:49:41 +0000 (UTC) Date: Wed, 17 Feb 2016 22:49:37 +0200 From: "Michael S. Tsirkin" Message-ID: <20160217224718-mutt-send-email-mst@redhat.com> References: <20160118230413.2140.8336.stgit@gimli.home> <20160217132812.7e9c16bf@t450s.home> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20160217132812.7e9c16bf@t450s.home> Subject: Re: [Qemu-devel] [RE-RESEND PATCH] pci: Adjust PCI config limit based on bus topology List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Alex Williamson Cc: qemu-devel@nongnu.org On Wed, Feb 17, 2016 at 01:28:12PM -0700, Alex Williamson wrote: > > Hi Michael, > > Would you like me to send this for a 4th time or are you OK with me > pulling this through my tree with Marcel's R-b? Thanks, > > Alex Sorry, keep forgetting about it. > > On Mon, 18 Jan 2016 16:06:03 -0700 > Alex Williamson wrote: > > > A conventional PCI bus does not support config space accesses above > > the standard 256 byte configuration space. PCIe-to-PCI bridges are > > not permitted to forward transactions if the extended register address > > field is non-zero and must handle it as an unsupported request (PCIe > > bridge spec rev 1.0, 4.1.3, 4.1.4). Therefore, we should not support > > extended config space if there is a conventional bus anywhere on the > > path to a device. > > > > Signed-off-by: Alex Williamson > > --- > > Previous postings: > > https://lists.gnu.org/archive/html/qemu-devel/2015-10/msg05384.html > > https://lists.gnu.org/archive/html/qemu-devel/2015-11/msg02422.html > > > > hw/pci/pci_host.c | 26 ++++++++++++++++++++++++++ > > 1 file changed, 26 insertions(+) > > > > diff --git a/hw/pci/pci_host.c b/hw/pci/pci_host.c > > index 49f59a5..3a3e294 100644 > > --- a/hw/pci/pci_host.c > > +++ b/hw/pci/pci_host.c > > @@ -19,6 +19,7 @@ > > */ > > > > #include "hw/pci/pci.h" > > +#include "hw/pci/pci_bridge.h" > > #include "hw/pci/pci_host.h" > > #include "hw/pci/pci_bus.h" > > #include "trace.h" > > @@ -49,9 +50,29 @@ static inline PCIDevice *pci_dev_find_by_addr(PCIBus *bus, uint32_t addr) > > return pci_find_device(bus, bus_num, devfn); > > } > > > > +static void pci_adjust_config_limit(PCIBus *bus, uint32_t *limit) > > +{ > > + if (*limit > PCI_CONFIG_SPACE_SIZE) { > > + if (!pci_bus_is_express(bus)) { > > + *limit = PCI_CONFIG_SPACE_SIZE; > > + return; > > + } > > + > > + if (!pci_bus_is_root(bus)) { > > + PCIDevice *bridge = pci_bridge_get_device(bus); > > + pci_adjust_config_limit(bridge->bus, limit); > > + } > > + } > > +} > > + > > void pci_host_config_write_common(PCIDevice *pci_dev, uint32_t addr, > > uint32_t limit, uint32_t val, uint32_t len) > > { > > + pci_adjust_config_limit(pci_dev->bus, &limit); > > + if (limit <= addr) { > > + return; > > + } > > + > > assert(len <= 4); > > /* non-zero functions are only exposed when function 0 is present, > > * allowing direct removal of unexposed functions. > > @@ -70,6 +91,11 @@ uint32_t pci_host_config_read_common(PCIDevice *pci_dev, uint32_t addr, > > { > > uint32_t ret; > > > > + pci_adjust_config_limit(pci_dev->bus, &limit); > > + if (limit <= addr) { > > + return ~0x0; > > + } > > + > > assert(len <= 4); > > /* non-zero functions are only exposed when function 0 is present, > > * allowing direct removal of unexposed functions. > > It's kind of nasty that we do it on each access ... How about storing the size in device itself? -- MST