From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from us-smtp-delivery-1.mimecast.com (us-smtp-delivery-1.mimecast.com [207.211.31.81])
 by mx.groups.io with SMTP id smtpd.web10.10550.1574769558764963161
 for <devel@edk2.groups.io>;
 Tue, 26 Nov 2019 03:59:19 -0800
Authentication-Results: mx.groups.io;
 dkim=pass header.i=@redhat.com header.s=mimecast20190719 header.b=QgGO3r1O;
 spf=pass (domain: redhat.com, ip: 207.211.31.81, mailfrom: philmd@redhat.com)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=redhat.com;
	s=mimecast20190719; t=1574769557;
	h=from:from:reply-to:subject:subject:date:date:message-id:message-id:
	 to:to:cc:cc:mime-version:mime-version:content-type:content-type:
	 content-transfer-encoding:content-transfer-encoding:
	 in-reply-to:in-reply-to:references:references;
	bh=ZTUatFnEMTHUGpn0gdMRPNWidaonKaZ4LkOt45QerwQ=;
	b=QgGO3r1O+jorbDGjfUyuiart7MJumAxMZA+tNcstuABoYHygnZ3blPtZV7Ql254WOuHcRe
	b7SSIKdBN7PcqGpKRGd5JV/+VjNtSCDlXaKVx/+h8QGb9xIJXGVW+A3UqKLAc7xk72Qv8M
	+HuQH8/q/FOGOtW/xMiND8qEgIsiaVo=
Received: from mail-wr1-f69.google.com (mail-wr1-f69.google.com
 [209.85.221.69]) (Using TLS) by relay.mimecast.com with ESMTP id
 us-mta-145-GX2WxiJ5PVexcT1hfpDZ6A-1; Tue, 26 Nov 2019 06:59:16 -0500
Received: by mail-wr1-f69.google.com with SMTP id q6so10414924wrv.11
        for <devel@edk2.groups.io>; Tue, 26 Nov 2019 03:59:16 -0800 (PST)
X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed;
        d=1e100.net; s=20161025;
        h=x-gm-message-state:subject:from:to:cc:references:message-id:date
         :user-agent:mime-version:in-reply-to:content-language
         :content-transfer-encoding;
        bh=UyuWvPXcn/PKDhKvhcDjGHzOU3mmkQ1lvs/uEwVbqIw=;
        b=XlC+g1oxAIhrygTAkQWdZD/FiIgReFRUiLJkxTxLjIhAI/S4e3M7iynp2f2Iazmuv5
         N+pNUZ77e3fBiCKVXwiBZBMTgF1Y66FUMR241amKWQ6vGG6LXtgoojcILDXnKaLD+Nlr
         mTSjBG9KEXexJL/GFkseLeBn2TTfvgRdFnRGUhN/DG+Fr5UhpELeG31tNr82ZRpdgMDj
         KyEFewVqVRzKZniDHf3+SJrBiE9ENaUZ8asUlZZu5NfN26ifBEBFfKF1lc7zEAU7qei9
         OnWj1aEvbEJowQZ3CUKmFcoo5F/jbHWhcB90FYOW3mhx2jlm6G/BMhQTuspjREQxwGpi
         QKAg==
X-Gm-Message-State: APjAAAUjBPU1IgEadlTgsVkRQeaZaXpNeyMYk1AB2QFfzBMOA0m6/I7G
	eJlOZfJl8iKniIxcQHbEhp/fTfY55LU5zTVdSVKC7D7flHAMEmPG4aEYWZsgS61MWmWoMYwVnAD
	OarflKogVVLtLwA==
X-Received: by 2002:a5d:444a:: with SMTP id x10mr36730761wrr.228.1574769554878;
        Tue, 26 Nov 2019 03:59:14 -0800 (PST)
X-Google-Smtp-Source: APXvYqxXw3PaLJpOMtjzEdxRFtmT5t+ljusK/pYucoeF30exZpP8x2joQ6E8LHY9Hr51dAZbv+D+Gw==
X-Received: by 2002:a5d:444a:: with SMTP id x10mr36730736wrr.228.1574769554496;
        Tue, 26 Nov 2019 03:59:14 -0800 (PST)
Return-Path: <philmd@redhat.com>
Received: from [192.168.1.35] (182.red-88-21-103.staticip.rima-tde.net. [88.21.103.182])
        by smtp.gmail.com with ESMTPSA id k1sm15049126wrp.29.2019.11.26.03.59.13
        (version=TLS1_3 cipher=TLS_AES_128_GCM_SHA256 bits=128/128);
        Tue, 26 Nov 2019 03:59:13 -0800 (PST)
Subject: Re: [edk2-devel] [PATCH v2 1/2] EmbeddedPkg/NonCoherentDmaLib: implement support for DMA range limits
From: =?UTF-8?B?UGhpbGlwcGUgTWF0aGlldS1EYXVkw6k=?= <philmd@redhat.com>
To: devel@edk2.groups.io, ard.biesheuvel@linaro.org
Cc: leif.lindholm@linaro.org
References: <20191125231242.12193-1-ard.biesheuvel@linaro.org>
 <20191125231242.12193-2-ard.biesheuvel@linaro.org>
 <ee6f5062-eeb1-fe89-ec8a-5ab2e6367b7a@redhat.com>
Message-ID: <08876497-a2ce-4b55-6f50-f332ffd45ddd@redhat.com>
Date: Tue, 26 Nov 2019 12:59:12 +0100
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:68.0) Gecko/20100101
 Thunderbird/68.1.1
MIME-Version: 1.0
In-Reply-To: <ee6f5062-eeb1-fe89-ec8a-5ab2e6367b7a@redhat.com>
X-MC-Unique: GX2WxiJ5PVexcT1hfpDZ6A-1
X-Mimecast-Spam-Score: 0
Content-Language: en-US
Content-Type: text/plain; charset=WINDOWS-1252; format=flowed
Content-Transfer-Encoding: quoted-printable

On 11/26/19 11:44 AM, Philippe Mathieu-Daud=E9 wrote:
> On 11/26/19 12:12 AM, Ard Biesheuvel via Groups.Io wrote:
>> Implement support for driving peripherals with limited DMA ranges to
>> NonCoherentDmaLib, by adding a device address limit, and taking it,
>> along with the device offset, into account when allocating or mapping
>> DMA buffers.
>>
>> Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
>> ---
>> =A0 EmbeddedPkg/EmbeddedPkg.dec=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=
=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 |=A0=A0 6 +
>> =A0 EmbeddedPkg/Library/NonCoherentDmaLib/NonCoherentDmaLib.c=A0=A0 | 16=
5=20
>> ++++++++++++++++++--
>> =A0 EmbeddedPkg/Library/NonCoherentDmaLib/NonCoherentDmaLib.inf |=A0=A0 =
1 +
>> =A0 3 files changed, 160 insertions(+), 12 deletions(-)
>>
>> diff --git a/EmbeddedPkg/EmbeddedPkg.dec b/EmbeddedPkg/EmbeddedPkg.dec
>> index 8812a6db7c30..69922802f473 100644
>> --- a/EmbeddedPkg/EmbeddedPkg.dec
>> +++ b/EmbeddedPkg/EmbeddedPkg.dec
>> @@ -186,6 +186,12 @@ [PcdsFixedAtBuild.common, PcdsDynamic.common]
>> =A0=A0=A0 #
>> =A0=A0=A0 gEmbeddedTokenSpaceGuid.PcdDmaDeviceOffset|0x0|UINT64|0x000005=
8
>> +=A0 #
>> +=A0 # Highest address value supported by the device for DMA addressing.=
=20
>> Note
>> +=A0 # that this value should be strictly greater than PcdDmaDeviceOffse=
t.
>> +=A0 #
>> + =20
>> gEmbeddedTokenSpaceGuid.PcdDmaDeviceLimit|0xFFFFFFFFFFFFFFFF|UINT64|0x00=
0005A=20
>>
>> +
>> =A0=A0=A0 #
>> =A0=A0=A0 # Selection between DT and ACPI as a default
>> =A0=A0=A0 #
>> diff --git a/EmbeddedPkg/Library/NonCoherentDmaLib/NonCoherentDmaLib.c=
=20
>> b/EmbeddedPkg/Library/NonCoherentDmaLib/NonCoherentDmaLib.c
>> index 78220f6358aa..115345765435 100644
>> --- a/EmbeddedPkg/Library/NonCoherentDmaLib/NonCoherentDmaLib.c
>> +++ b/EmbeddedPkg/Library/NonCoherentDmaLib/NonCoherentDmaLib.c
>> @@ -40,6 +40,8 @@ typedef struct {
>> =A0 STATIC EFI_CPU_ARCH_PROTOCOL=A0=A0=A0=A0=A0 *mCpu;
>> =A0 STATIC LIST_ENTRY=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 Un=
cachedAllocationList;
>> +STATIC PHYSICAL_ADDRESS=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 mDmaHostAddressLi=
mit;
>> +
>> =A0 STATIC
>> =A0 PHYSICAL_ADDRESS
>> =A0 HostToDeviceAddress (
>> @@ -49,6 +51,102 @@ HostToDeviceAddress (
>> =A0=A0=A0 return (PHYSICAL_ADDRESS)(UINTN)Address + PcdGet64=20
>> (PcdDmaDeviceOffset);
>> =A0 }
>> +/**
>> +=A0 Allocates one or more 4KB pages of a certain memory type at a=20
>> specified
>> +=A0 alignment.
>> +
>> +=A0 Allocates the number of 4KB pages specified by Pages of a certain=
=20
>> memory type
>> +=A0 with an alignment specified by Alignment. The allocated buffer is=
=20
>> returned.
>> +=A0 If Pages is 0, then NULL is returned. If there is not enough memory=
=20
>> at the
>> +=A0 specified alignment remaining to satisfy the request, then NULL is=
=20
>> returned.
>> +=A0 If Alignment is not a power of two and Alignment is not zero, then=
=20
>> ASSERT().
>> +=A0 If Pages plus EFI_SIZE_TO_PAGES (Alignment) overflows, then ASSERT(=
).
>> +
>> +=A0 @param=A0 MemoryType=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 The type of m=
emory to allocate.
>> +=A0 @param=A0 Pages=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 The=
 number of 4 KB pages to allocate.
>> +=A0 @param=A0 Alignment=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 The request=
ed alignment of the=20
>> allocation.
>> +=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=
=A0=A0=A0=A0=A0=A0=A0=A0 Must be a power of two.
>> +=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=
=A0=A0=A0=A0=A0=A0=A0=A0 If Alignment is zero, then byte=20
>> alignment is
>> +=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=
=A0=A0=A0=A0=A0=A0=A0=A0 used.
>> +
>> +=A0 @return A pointer to the allocated buffer or NULL if allocation fai=
ls.
>> +
>> +**/
>> +STATIC
>> +VOID *
>> +InternalAllocateAlignedPages (
>> +=A0 IN EFI_MEMORY_TYPE=A0 MemoryType,
>> +=A0 IN UINTN=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 Pages,
>> +=A0 IN UINTN=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 Alignment
>> +=A0 )
>> +{
>> +=A0 EFI_STATUS=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 Status;
>> +=A0 EFI_PHYSICAL_ADDRESS=A0 Memory;
>> +=A0 UINTN=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 AlignedMemory=
;
>> +=A0 UINTN=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 AlignmentMask=
;
>> +=A0 UINTN=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 UnalignedPage=
s;
>> +=A0 UINTN=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 RealPages;
>> +
>> +=A0 //
>> +=A0 // Alignment must be a power of two or zero.
>> +=A0 //
>> +=A0 ASSERT ((Alignment & (Alignment - 1)) =3D=3D 0);
>> +
>> +=A0 if (Pages =3D=3D 0) {
>> +=A0=A0=A0 return NULL;
>> +=A0 }
>> +=A0 if (Alignment > EFI_PAGE_SIZE) {
>> +=A0=A0=A0 //
>> +=A0=A0=A0 // Calculate the total number of pages since alignment is lar=
ger=20
>> than page
>> +=A0=A0=A0 // size.
>> +=A0=A0=A0 //
>> +=A0=A0=A0 AlignmentMask=A0 =3D Alignment - 1;
>> +=A0=A0=A0 RealPages=A0=A0=A0=A0=A0 =3D Pages + EFI_SIZE_TO_PAGES (Align=
ment);
>> +=A0=A0=A0 //
>> +=A0=A0=A0 // Make sure that Pages plus EFI_SIZE_TO_PAGES (Alignment) do=
es not
>> +=A0=A0=A0 // overflow.
>> +=A0=A0=A0 //
>> +=A0=A0=A0 ASSERT (RealPages > Pages);
>> +
>> +=A0=A0=A0 Memory =3D mDmaHostAddressLimit;
>> +=A0=A0=A0 Status =3D gBS->AllocatePages (AllocateMaxAddress, MemoryType=
,=20
>> RealPages,
>> +=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 &Memory);
>> +=A0=A0=A0 if (EFI_ERROR (Status)) {
>> +=A0=A0=A0=A0=A0 return NULL;
>> +=A0=A0=A0 }
>> +=A0=A0=A0 AlignedMemory=A0 =3D ((UINTN)Memory + AlignmentMask) & ~Align=
mentMask;
>> +=A0=A0=A0 UnalignedPages =3D EFI_SIZE_TO_PAGES (AlignedMemory - (UINTN)=
Memory);

The previous line made me think twice (due to the cast), but I can't=20
find a simpler way to write the same code.

>> +=A0=A0=A0 if (UnalignedPages > 0) {
>> +=A0=A0=A0=A0=A0 //
>> +=A0=A0=A0=A0=A0 // Free first unaligned page(s).
>> +=A0=A0=A0=A0=A0 //
>> +=A0=A0=A0=A0=A0 Status =3D gBS->FreePages (Memory, UnalignedPages);
>> +=A0=A0=A0=A0=A0 ASSERT_EFI_ERROR (Status);
>> +=A0=A0=A0 }
>> +=A0=A0=A0 Memory=A0=A0=A0=A0=A0=A0=A0=A0 =3D AlignedMemory + EFI_PAGES_=
TO_SIZE (Pages);
>> +=A0=A0=A0 UnalignedPages =3D RealPages - Pages - UnalignedPages;
>> +=A0=A0=A0 if (UnalignedPages > 0) {
>> +=A0=A0=A0=A0=A0 //
>> +=A0=A0=A0=A0=A0 // Free last unaligned page(s).
>> +=A0=A0=A0=A0=A0 //
>> +=A0=A0=A0=A0=A0 Status =3D gBS->FreePages (Memory, UnalignedPages);
>> +=A0=A0=A0=A0=A0 ASSERT_EFI_ERROR (Status);
>> +=A0=A0=A0 }
>> +=A0 } else {
>> +=A0=A0=A0 //
>> +=A0=A0=A0 // Do not over-allocate pages in this case.
>> +=A0=A0=A0 //
>> +=A0=A0=A0 Memory =3D mDmaHostAddressLimit;
>> +=A0=A0=A0 Status =3D gBS->AllocatePages (AllocateMaxAddress, MemoryType=
, Pages,
>> +=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 &Memory);
>> +=A0=A0=A0 if (EFI_ERROR (Status)) {
>> +=A0=A0=A0=A0=A0 return NULL;
>> +=A0=A0=A0 }
>> +=A0=A0=A0 AlignedMemory =3D (UINTN)Memory;
>> +=A0 }
>> +=A0 return (VOID *)AlignedMemory;
>> +}
>> +
>> =A0 /**
>> =A0=A0=A0 Provides the DMA controller-specific addresses needed to acces=
s=20
>> system memory.
>> @@ -111,7 +209,30 @@ DmaMap (
>> =A0=A0=A0=A0=A0 return=A0 EFI_OUT_OF_RESOURCES;
>> =A0=A0=A0 }
>> -=A0 if (Operation !=3D MapOperationBusMasterRead &&
>> +=A0 if (((UINTN)HostAddress + *NumberOfBytes) > mDmaHostAddressLimit) {
>> +
>> +=A0=A0=A0 if (Operation =3D=3D MapOperationBusMasterCommonBuffer) {
>> +=A0=A0=A0=A0=A0 goto CommonBufferError;
>> +=A0=A0=A0 }
>> +
>> +=A0=A0=A0 AllocSize =3D ALIGN_VALUE (*NumberOfBytes, mCpu->DmaBufferAli=
gnment);
>> +=A0=A0=A0 Map->BufferAddress =3D InternalAllocateAlignedPages=20
>> (EfiBootServicesData,
>> +=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=
=A0=A0=A0 EFI_SIZE_TO_PAGES (AllocSize),
>> +=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=
=A0=A0=A0 mCpu->DmaBufferAlignment);
>> +=A0=A0=A0 if (Map->BufferAddress =3D=3D NULL) {
>> +=A0=A0=A0=A0=A0 Status =3D EFI_OUT_OF_RESOURCES;
>> +=A0=A0=A0=A0=A0 goto FreeMapInfo;
>> +=A0=A0=A0 }
>> +
>> +=A0=A0=A0 if (Map->Operation =3D=3D MapOperationBusMasterRead) {
>> +=A0=A0=A0=A0=A0 CopyMem (Map->BufferAddress, (VOID *)(UINTN)Map->HostAd=
dress,
>> +=A0=A0=A0=A0=A0=A0=A0 *NumberOfBytes);
>> +=A0=A0=A0 }
>> +=A0=A0=A0 mCpu->FlushDataCache (mCpu, (UINTN)Map->BufferAddress, AllocS=
ize,
>> +=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 EfiCpuFlushTypeWriteBack);
>> +
>> +=A0=A0=A0 *DeviceAddress =3D HostToDeviceAddress (Map->BufferAddress);
>> +=A0 } else if (Operation !=3D MapOperationBusMasterRead &&
>> =A0=A0=A0=A0=A0=A0=A0 ((((UINTN)HostAddress & (mCpu->DmaBufferAlignment =
- 1)) !=3D 0) ||
>> =A0=A0=A0=A0=A0=A0=A0=A0 ((*NumberOfBytes & (mCpu->DmaBufferAlignment - =
1)) !=3D 0))) {
>> @@ -128,12 +249,7 @@ DmaMap (
>> =A0=A0=A0=A0=A0=A0=A0 // on uncached buffers.
>> =A0=A0=A0=A0=A0=A0=A0 //
>> =A0=A0=A0=A0=A0=A0=A0 if (Operation =3D=3D MapOperationBusMasterCommonBu=
ffer) {
>> -=A0=A0=A0=A0=A0=A0=A0 DEBUG ((DEBUG_ERROR,
>> -=A0=A0=A0=A0=A0=A0=A0=A0=A0 "%a: Operation type 'MapOperationBusMasterC=
ommonBuffer' is=20
>> only "
>> -=A0=A0=A0=A0=A0=A0=A0=A0=A0 "supported\non memory regions that were all=
ocated using "
>> -=A0=A0=A0=A0=A0=A0=A0=A0=A0 "DmaAllocateBuffer ()\n", __FUNCTION__));
>> -=A0=A0=A0=A0=A0=A0=A0 Status =3D EFI_UNSUPPORTED;
>> -=A0=A0=A0=A0=A0=A0=A0 goto FreeMapInfo;
>> +=A0=A0=A0=A0=A0=A0=A0 goto CommonBufferError;
>> =A0=A0=A0=A0=A0=A0=A0 }
>> =A0=A0=A0=A0=A0=A0=A0 //
>> @@ -199,6 +315,12 @@ DmaMap (
>> =A0=A0=A0 return EFI_SUCCESS;
>> +CommonBufferError:
>> +=A0 DEBUG ((DEBUG_ERROR,
>> +=A0=A0=A0 "%a: Operation type 'MapOperationBusMasterCommonBuffer' is on=
ly "
>> +=A0=A0=A0 "supported\non memory regions that were allocated using "
>> +=A0=A0=A0 "DmaAllocateBuffer ()\n", __FUNCTION__));
>> +=A0 Status =3D EFI_UNSUPPORTED;
>> =A0 FreeMapInfo:
>> =A0=A0=A0 FreePool (Map);
>> @@ -229,6 +351,7 @@ DmaUnmap (
>> =A0=A0=A0 MAP_INFO_INSTANCE *Map;
>> =A0=A0=A0 EFI_STATUS=A0=A0=A0=A0=A0=A0=A0 Status;
>> =A0=A0=A0 VOID=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 *Buffer;
>> +=A0 UINTN=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 AllocSize;
>> =A0=A0=A0 if (Mapping =3D=3D NULL) {
>> =A0=A0=A0=A0=A0 ASSERT (FALSE);
>> @@ -238,7 +361,17 @@ DmaUnmap (
>> =A0=A0=A0 Map =3D (MAP_INFO_INSTANCE *)Mapping;
>> =A0=A0=A0 Status =3D EFI_SUCCESS;
>> -=A0 if (Map->DoubleBuffer) {
>> +=A0 if (((UINTN)Map->HostAddress + Map->NumberOfBytes) >=20
>> mDmaHostAddressLimit) {
>> +=A0=A0=A0 AllocSize =3D ALIGN_VALUE (Map->NumberOfBytes,=20
>> mCpu->DmaBufferAlignment);
>> +=A0=A0=A0 if (Map->Operation =3D=3D MapOperationBusMasterWrite) {
>> +=A0=A0=A0=A0=A0 mCpu->FlushDataCache (mCpu, (UINTN)Map->BufferAddress, =
AllocSize,
>> +=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 EfiCpuFlushTypeInvalidate);
>> +=A0=A0=A0=A0=A0 CopyMem ((VOID *)(UINTN)Map->HostAddress, Map->BufferAd=
dress,
>> +=A0=A0=A0=A0=A0=A0=A0 Map->NumberOfBytes);
>> +=A0=A0=A0 }
>> +=A0=A0=A0 FreePages (Map->BufferAddress, EFI_SIZE_TO_PAGES (AllocSize))=
;
>> +=A0 } else if (Map->DoubleBuffer) {
>> +
>> =A0=A0=A0=A0=A0 ASSERT (Map->Operation =3D=3D MapOperationBusMasterWrite=
);
>> =A0=A0=A0=A0=A0 if (Map->Operation !=3D MapOperationBusMasterWrite) {
>> @@ -335,10 +468,9 @@ DmaAllocateAlignedBuffer (
>> =A0=A0=A0=A0=A0 return EFI_INVALID_PARAMETER;
>> =A0=A0=A0 }
>> -=A0 if (MemoryType =3D=3D EfiBootServicesData) {
>> -=A0=A0=A0 Allocation =3D AllocateAlignedPages (Pages, Alignment);
>> -=A0 } else if (MemoryType =3D=3D EfiRuntimeServicesData) {
>> -=A0=A0=A0 Allocation =3D AllocateAlignedRuntimePages (Pages, Alignment)=
;
>> +=A0 if (MemoryType =3D=3D EfiBootServicesData ||
>> +=A0=A0=A0=A0=A0 MemoryType =3D=3D EfiRuntimeServicesData) {
>> +=A0=A0=A0 Allocation =3D InternalAllocateAlignedPages (MemoryType, Page=
s,=20
>> Alignment);
>> =A0=A0=A0 } else {
>> =A0=A0=A0=A0=A0 return EFI_INVALID_PARAMETER;
>> =A0=A0=A0 }
>> @@ -479,6 +611,15 @@ NonCoherentDmaLibConstructor (
>> =A0 {
>> =A0=A0=A0 InitializeListHead (&UncachedAllocationList);
>> +=A0 //
>> +=A0 // Ensure that the combination of DMA addressing offset and limit=
=20
>> produces
>> +=A0 // a sane value.
>> +=A0 //
>> +=A0 ASSERT (PcdGet64 (PcdDmaDeviceLimit) > PcdGet64 (PcdDmaDeviceOffset=
));
>> +
>> +=A0 mDmaHostAddressLimit =3D PcdGet64 (PcdDmaDeviceLimit) -
>> +=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=
=A0 PcdGet64 (PcdDmaDeviceOffset);
>> +
>> =A0=A0=A0 // Get the Cpu protocol for later use
>> =A0=A0=A0 return gBS->LocateProtocol (&gEfiCpuArchProtocolGuid, NULL, (V=
OID=20
>> **)&mCpu);
>> =A0 }
>> diff --git=20
>> a/EmbeddedPkg/Library/NonCoherentDmaLib/NonCoherentDmaLib.inf=20
>> b/EmbeddedPkg/Library/NonCoherentDmaLib/NonCoherentDmaLib.inf
>> index 2db751afee91..1a21cfe4ff23 100644
>> --- a/EmbeddedPkg/Library/NonCoherentDmaLib/NonCoherentDmaLib.inf
>> +++ b/EmbeddedPkg/Library/NonCoherentDmaLib/NonCoherentDmaLib.inf
>> @@ -38,6 +38,7 @@ [Protocols]
>> =A0 [Pcd]
>> =A0=A0=A0 gEmbeddedTokenSpaceGuid.PcdDmaDeviceOffset
>> +=A0 gEmbeddedTokenSpaceGuid.PcdDmaDeviceLimit
>> =A0 [Depex]
>> =A0=A0=A0 gEfiCpuArchProtocolGuid
>>
>=20
> For the 82% of the patch I understood:
> Reviewed-by: Philippe Mathieu-Daude <philmd@redhat.com>

Leif explained me that "I've reviewed it, but don't fully understand it,=20
but I think it's valuable" translate into a Acked-by on EDK2, so please=20
replace the previous Reviewed-by by:
Acked-by: Philippe Mathieu-Daude <philmd@redhat.com>

Thanks,

Phil.