From mboxrd@z Thu Jan 1 00:00:00 1970 Authentication-Results: mx.groups.io; dkim=missing; spf=pass (domain: redhat.com, ip: 209.132.183.28, mailfrom: lersek@redhat.com) Received: from mx1.redhat.com (mx1.redhat.com [209.132.183.28]) by groups.io with SMTP; Fri, 26 Apr 2019 11:15:13 -0700 Received: from smtp.corp.redhat.com (int-mx04.intmail.prod.int.phx2.redhat.com [10.5.11.14]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by mx1.redhat.com (Postfix) with ESMTPS id D314A2FE55A; Fri, 26 Apr 2019 18:15:11 +0000 (UTC) Received: from lacos-laptop-7.usersys.redhat.com (ovpn-121-104.rdu2.redhat.com [10.10.121.104]) by smtp.corp.redhat.com (Postfix) with ESMTP id A54955DE82; Fri, 26 Apr 2019 18:15:08 +0000 (UTC) Subject: Re: [PATCH v3 3/3] OvmfPkg/XenSupport: turn off address decoding before BAR sizing To: Igor Druzhinin , devel@edk2.groups.io Cc: jordan.l.justen@intel.com, ard.biesheuvel@linaro.org, anthony.perard@citrix.com, julien.grall@arm.com, xen-devel@lists.xenproject.org References: <1556223838-5176-1-git-send-email-igor.druzhinin@citrix.com> <1556223838-5176-4-git-send-email-igor.druzhinin@citrix.com> From: "Laszlo Ersek" Message-ID: Date: Fri, 26 Apr 2019 20:15:07 +0200 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:52.0) Gecko/20100101 Thunderbird/52.9.1 MIME-Version: 1.0 In-Reply-To: <1556223838-5176-4-git-send-email-igor.druzhinin@citrix.com> X-Scanned-By: MIMEDefang 2.79 on 10.5.11.14 X-Greylist: Sender IP whitelisted, not delayed by milter-greylist-4.5.16 (mx1.redhat.com [10.5.110.29]); Fri, 26 Apr 2019 18:15:12 +0000 (UTC) Content-Type: text/plain; charset=utf-8 Content-Language: en-US Content-Transfer-Encoding: 7bit On 04/25/19 22:23, Igor Druzhinin wrote: > On Xen, hvmloader firmware leaves address decoding enabled for > enumerated PCI device before jumping into OVMF. OVMF seems to > expect it to be disabled and tries to size PCI BARs in several places > without disabling it which causes BAR64, for example, being > incorrectly placed by QEMU. > > Fix it by disabling PCI address decoding explicitly before the > first attempt to size BARs on Xen. > > Signed-off-by: Igor Druzhinin > --- > Changes in v3: > - dropped unused arguments and rename PcatPciRootBridgeDecoding function > - make mask application more clear as suggested > --- > OvmfPkg/Library/PciHostBridgeLib/XenSupport.c | 21 +++++++++++++++++++++ > 1 file changed, 21 insertions(+) > > diff --git a/OvmfPkg/Library/PciHostBridgeLib/XenSupport.c b/OvmfPkg/Library/PciHostBridgeLib/XenSupport.c > index 76fca53..b41bd6b 100644 > --- a/OvmfPkg/Library/PciHostBridgeLib/XenSupport.c > +++ b/OvmfPkg/Library/PciHostBridgeLib/XenSupport.c > @@ -49,6 +49,22 @@ PcatPciRootBridgeBarExisted ( > EnableInterrupts (); > } > > +#define PCI_COMMAND_DECODE ((UINT16)(EFI_PCI_COMMAND_IO_SPACE | \ > + EFI_PCI_COMMAND_MEMORY_SPACE)) > +STATIC > +VOID > +PcatPciRootBridgeDecodingDisable ( > + IN UINTN Address > + ) > +{ > + UINT16 Value; > + > + Value = PciRead16 (Address); > + if (Value & PCI_COMMAND_DECODE) { > + PciWrite16 (Address, Value & ~(UINT32)PCI_COMMAND_DECODE); > + } > +} > + > STATIC > VOID > PcatPciRootBridgeParseBars ( > @@ -75,6 +91,11 @@ PcatPciRootBridgeParseBars ( > UINT64 Limit; > PCI_ROOT_BRIDGE_APERTURE *MemAperture; > > + // Disable address decoding for every device before OVMF starts sizing it > + PcatPciRootBridgeDecodingDisable ( > + PCI_LIB_ADDRESS (Bus, Device, Function, PCI_COMMAND_OFFSET) > + ); > + > for (Offset = BarOffsetBase; Offset < BarOffsetEnd; Offset += sizeof (UINT32)) { > PcatPciRootBridgeBarExisted ( > PCI_LIB_ADDRESS (Bus, Device, Function, Offset), > Acked-by: Laszlo Ersek Someone from the Xen community please ACK this too, and then we can push the series. Thanks, Laszlo