* [PATCH v3 0/6] Add ARM support for VS2017 @ 2017-12-08 14:06 Pete Batard 2017-12-08 14:06 ` [PATCH v3 1/6] MdePkg: Disable some Level 4 warnings for VS2017/ARM Pete Batard ` (6 more replies) 0 siblings, 7 replies; 12+ messages in thread From: Pete Batard @ 2017-12-08 14:06 UTC (permalink / raw) To: edk2-devel; +Cc: liming.gao This v3 is the same as v2, except for the following comment added in tools_def for VS2017: # Building of XIP firmware images for ARM is not currently supported (only applications). # /FILEALIGN:4096 and other changes are needed for ARM firmware builds. The following series adds ARM compilation support for the VS2017 toolchain. * PATCH 1 targets the disabling of VS Level 4 warnings. The disabled warnings for ARM are now aligned with IA32 and X64. * PATCH 2 adds a NULL handler for the base stack check, since this is a GCC functionality. * PATCH 3 updates MdePkg/Library/BaseLib so that the RVCT assembly sources are also used for MSFT. * PATCH 4 adds the required compiler intrinsics replacements for division, shift and memset/memcpy. * PATCH 5 adds variable argument handlers for print output. Note that this is done without relying on any external headers, with the VA_ARG macro having been reverse engineered from MSFT ARM assembly output. * PATCH 6 enables the selection of ARM in the conf templates. With these patches, VS2017 toolchain users should be able to compile regular UEFI ARM applications using EDK2. Note that, unlike ARM64 support, ARM support does not require a specific update of Visual Studio 2017, as the ARM toolchain has been available from the very first release. Additional notes: We tested compiling and running the full UEFI Shell with this series, as well as a small set of applications and drivers, and found no issues. With an additional patch [1], it is also possible to use this proposal to compile a complete QEMU ARM firmware. As the patch shows, the changes that need to be applied to the EDK2 sources to achieve this are actually very minimal. However, the generated firmware does not currently boot, possibly because of the following warnings being generated by the MS compiler: - ArmCpuDxe.dll : warning LNK4072: section count 118 exceeds max (96); image may not run - UiApp.dll : warning LNK4072: section count 113 exceeds max (96); image may not run As far as I could see, the section count max is hardcoded so a workaround would be needed to address those. Also, because the VS2017 ARM compiler forces a section alignment of 4096 bytes (which in turn forces use to use /FILEALIGN:4096 as a linker option for the firmware generation), the generated firmware exceeds 2MB and we had to double its size to 4MB. At this stage, since the goal of this series is to allow users to compile regular ARM UEFI applications using the VS2017 toolchain, I have non plans to spend more time on the QEMU firmware issues, especially as I suspect that reducing the firmware size back to 2 MB may not be achievable without Microsoft altering their compiler. I am however hopeful that ARM specialists can take this matter over eventually... Regards, /Pete [1] https://github.com/pbatard/edk2/commit/c4ce41094a46f4f3dc7ccc64a90604813f037b13 Pete Batard (6): MdePkg: Disable some Level 4 warnings for VS2017/ARM MdePkg/Library/BaseStackCheckLib: Add Null handler for VS2017/ARM MdePkg/Library/BaseLib: Enable VS2017/ARM builds ArmPkg/Library/CompilerIntrinsicsLib: Enable VS2017/ARM builds MdePkg/Include: Add VA list support for VS2017/ARM BaseTools/Conf: Add VS2017/ARM support ArmPkg/Library/CompilerIntrinsicsLib/Arm/rtdiv.asm | 255 ++++++++++++++++++++ ArmPkg/Library/CompilerIntrinsicsLib/Arm/rtsrsh.asm | 45 ++++ ArmPkg/Library/CompilerIntrinsicsLib/CompilerIntrinsicsLib.inf | 13 +- ArmPkg/Library/CompilerIntrinsicsLib/memcpy_ms.c | 34 +++ ArmPkg/Library/CompilerIntrinsicsLib/memset_ms.c | 33 +++ BaseTools/Conf/build_rule.template | 31 ++- BaseTools/Conf/tools_def.template | 31 +++ MdePkg/Include/Arm/ProcessorBind.h | 96 ++++++-- MdePkg/Include/Base.h | 13 + MdePkg/Library/BaseLib/Arm/CpuBreakpoint.asm | 5 +- MdePkg/Library/BaseLib/BaseLib.inf | 16 +- MdePkg/Library/BaseStackCheckLib/BaseStackCheckLib.inf | 5 +- MdePkg/Library/BaseStackCheckLib/BaseStackCheckNull.c | 18 ++ 13 files changed, 565 insertions(+), 30 deletions(-) create mode 100644 ArmPkg/Library/CompilerIntrinsicsLib/Arm/rtdiv.asm create mode 100644 ArmPkg/Library/CompilerIntrinsicsLib/Arm/rtsrsh.asm create mode 100644 ArmPkg/Library/CompilerIntrinsicsLib/memcpy_ms.c create mode 100644 ArmPkg/Library/CompilerIntrinsicsLib/memset_ms.c create mode 100644 MdePkg/Library/BaseStackCheckLib/BaseStackCheckNull.c -- 2.9.3.windows.2 ^ permalink raw reply [flat|nested] 12+ messages in thread
* [PATCH v3 1/6] MdePkg: Disable some Level 4 warnings for VS2017/ARM 2017-12-08 14:06 [PATCH v3 0/6] Add ARM support for VS2017 Pete Batard @ 2017-12-08 14:06 ` Pete Batard 2017-12-08 14:06 ` [PATCH v3 2/6] MdePkg/Library/BaseStackCheckLib: Add Null handler " Pete Batard ` (5 subsequent siblings) 6 siblings, 0 replies; 12+ messages in thread From: Pete Batard @ 2017-12-08 14:06 UTC (permalink / raw) To: edk2-devel; +Cc: liming.gao We disable the exact same warnings as IA32 and X64. Also create a dummy macro for PRESERVE8, as this is not supported by the Microsoft ARM assembler. Contributed-under: TianoCore Contribution Agreement 1.1 Signed-off-by: Pete Batard <pete@akeo.ie> --- MdePkg/Include/Arm/ProcessorBind.h | 96 +++++++++++++++----- 1 file changed, 75 insertions(+), 21 deletions(-) diff --git a/MdePkg/Include/Arm/ProcessorBind.h b/MdePkg/Include/Arm/ProcessorBind.h index dde1fd1152ba..0700e4a4c50a 100644 --- a/MdePkg/Include/Arm/ProcessorBind.h +++ b/MdePkg/Include/Arm/ProcessorBind.h @@ -1,15 +1,15 @@ /** @file Processor or Compiler specific defines and types for ARM. - Copyright (c) 2006 - 2013, Intel Corporation. All rights reserved.<BR> + Copyright (c) 2006 - 2017, Intel Corporation. All rights reserved.<BR> Portions copyright (c) 2008 - 2009, Apple Inc. All rights reserved.<BR> - This program and the accompanying materials - are licensed and made available under the terms and conditions of the BSD License - which accompanies this distribution. The full text of the license may be found at - http://opensource.org/licenses/bsd-license.php + This program and the accompanying materials + are licensed and made available under the terms and conditions of the BSD License + which accompanies this distribution. The full text of the license may be found at + http://opensource.org/licenses/bsd-license.php - THE PROGRAM IS DISTRIBUTED UNDER THE BSD LICENSE ON AN "AS IS" BASIS, - WITHOUT WARRANTIES OR REPRESENTATIONS OF ANY KIND, EITHER EXPRESS OR IMPLIED. + THE PROGRAM IS DISTRIBUTED UNDER THE BSD LICENSE ON AN "AS IS" BASIS, + WITHOUT WARRANTIES OR REPRESENTATIONS OF ANY KIND, EITHER EXPRESS OR IMPLIED. **/ @@ -28,14 +28,63 @@ #pragma pack() #endif +#if defined(_MSC_EXTENSIONS) + // -// RVCT does not support the __builtin_unreachable() macro +// Disable some level 4 compilation warnings (same as IA32 and X64) // -#ifdef __ARMCC_VERSION + +// +// Disabling bitfield type checking warnings. +// +#pragma warning ( disable : 4214 ) + +// +// Disabling the unreferenced formal parameter warnings. +// +#pragma warning ( disable : 4100 ) + +// +// Disable slightly different base types warning as CHAR8 * can not be set +// to a constant string. +// +#pragma warning ( disable : 4057 ) + +// +// ASSERT(FALSE) or while (TRUE) are legal constructs so suppress this warning +// +#pragma warning ( disable : 4127 ) + +// +// This warning is caused by functions defined but not used. For precompiled header only. +// +#pragma warning ( disable : 4505 ) + +// +// This warning is caused by empty (after preprocessing) source file. For precompiled header only. +// +#pragma warning ( disable : 4206 ) + +// +// Disable 'potentially uninitialized local variable X used' warnings +// +#pragma warning ( disable : 4701 ) + +// +// Disable 'potentially uninitialized local pointer variable X used' warnings +// +#pragma warning ( disable : 4703 ) + +#endif + +// +// RVCT and MSFT don't support the __builtin_unreachable() macro +// +#if defined(__ARMCC_VERSION) || defined(_MSC_EXTENSIONS) #define UNREACHABLE() #endif -#if _MSC_EXTENSIONS +#if defined(_MSC_EXTENSIONS) // // use Microsoft* C compiler dependent integer width types // @@ -52,7 +101,7 @@ typedef signed char INT8; #else // - // Assume standard ARM alignment. + // Assume standard ARM alignment. // Need to check portability of long long // typedef unsigned long long UINT64; @@ -121,7 +170,7 @@ typedef INT32 INTN; // use the correct C calling convention. All protocol member functions and // EFI intrinsics are required to modify their member functions with EFIAPI. // -#define EFIAPI +#define EFIAPI // When compiling with Clang, we still use GNU as for the assembler, so we still // need to define the GCC_ASM* macros. @@ -142,34 +191,39 @@ typedef INT32 INTN; #define GCC_ASM_EXPORT(func__) \ .global _CONCATENATE (__USER_LABEL_PREFIX__, func__) ;\ - .type ASM_PFX(func__), %function + .type ASM_PFX(func__), %function #define GCC_ASM_IMPORT(func__) \ .extern _CONCATENATE (__USER_LABEL_PREFIX__, func__) - + #else // - // .type not supported by Apple Xcode tools + // .type not supported by Apple Xcode tools // - #define INTERWORK_FUNC(func__) + #define INTERWORK_FUNC(func__) #define GCC_ASM_EXPORT(func__) \ .globl _CONCATENATE (__USER_LABEL_PREFIX__, func__) \ - - #define GCC_ASM_IMPORT(name) + + #define GCC_ASM_IMPORT(name) #endif +#elif defined(_MSC_EXTENSIONS) + // + // PRESERVE8 is not supported by the MSFT assembler. + // + #define PRESERVE8 #endif /** Return the pointer to the first instruction of a function given a function pointer. - On ARM CPU architectures, these two pointer values are the same, + On ARM CPU architectures, these two pointer values are the same, so the implementation of this macro is very simple. - + @param FunctionPointer A pointer to a function. @return The pointer to the first instruction of a function given a function pointer. - + **/ #define FUNCTION_ENTRY_POINT(FunctionPointer) (VOID *)(UINTN)(FunctionPointer) -- 2.9.3.windows.2 ^ permalink raw reply related [flat|nested] 12+ messages in thread
* [PATCH v3 2/6] MdePkg/Library/BaseStackCheckLib: Add Null handler for VS2017/ARM 2017-12-08 14:06 [PATCH v3 0/6] Add ARM support for VS2017 Pete Batard 2017-12-08 14:06 ` [PATCH v3 1/6] MdePkg: Disable some Level 4 warnings for VS2017/ARM Pete Batard @ 2017-12-08 14:06 ` Pete Batard 2017-12-08 14:06 ` [PATCH v3 3/6] MdePkg/Library/BaseLib: Enable VS2017/ARM builds Pete Batard ` (4 subsequent siblings) 6 siblings, 0 replies; 12+ messages in thread From: Pete Batard @ 2017-12-08 14:06 UTC (permalink / raw) To: edk2-devel; +Cc: liming.gao Contributed-under: TianoCore Contribution Agreement 1.1 Signed-off-by: Pete Batard <pete@akeo.ie> --- MdePkg/Library/BaseStackCheckLib/BaseStackCheckLib.inf | 5 +++-- MdePkg/Library/BaseStackCheckLib/BaseStackCheckNull.c | 18 ++++++++++++++++++ 2 files changed, 21 insertions(+), 2 deletions(-) diff --git a/MdePkg/Library/BaseStackCheckLib/BaseStackCheckLib.inf b/MdePkg/Library/BaseStackCheckLib/BaseStackCheckLib.inf index d02d97107b08..e280651b1199 100644 --- a/MdePkg/Library/BaseStackCheckLib/BaseStackCheckLib.inf +++ b/MdePkg/Library/BaseStackCheckLib/BaseStackCheckLib.inf @@ -30,8 +30,9 @@ [Defines] # [Sources] - BaseStackCheckGcc.c | GCC - BaseStackCheckGcc.c | RVCT + BaseStackCheckGcc.c | GCC + BaseStackCheckGcc.c | RVCT + BaseStackCheckNull.c | MSFT [Packages] MdePkg/MdePkg.dec diff --git a/MdePkg/Library/BaseStackCheckLib/BaseStackCheckNull.c b/MdePkg/Library/BaseStackCheckLib/BaseStackCheckNull.c new file mode 100644 index 000000000000..fb2f65929d3e --- /dev/null +++ b/MdePkg/Library/BaseStackCheckLib/BaseStackCheckNull.c @@ -0,0 +1,18 @@ +/*++ + + Copyright (c) 2017, Intel Corporation. All rights reserved.<BR> + This program and the accompanying materials + are licensed and made available under the terms and conditions of the BSD License + which accompanies this distribution. The full text of the license may be found at + http://opensource.org/licenses/bsd-license.php. + + THE PROGRAM IS DISTRIBUTED UNDER THE BSD LICENSE ON AN "AS IS" BASIS, + WITHOUT WARRANTIES OR REPRESENTATIONS OF ANY KIND, EITHER EXPRESS OR IMPLIED. + +Abstract: + + This file is purely empty as a work around for BaseStackCheck to pass MSVC build. + +**/ + +extern int __BaseStackCheckNull; -- 2.9.3.windows.2 ^ permalink raw reply related [flat|nested] 12+ messages in thread
* [PATCH v3 3/6] MdePkg/Library/BaseLib: Enable VS2017/ARM builds 2017-12-08 14:06 [PATCH v3 0/6] Add ARM support for VS2017 Pete Batard 2017-12-08 14:06 ` [PATCH v3 1/6] MdePkg: Disable some Level 4 warnings for VS2017/ARM Pete Batard 2017-12-08 14:06 ` [PATCH v3 2/6] MdePkg/Library/BaseStackCheckLib: Add Null handler " Pete Batard @ 2017-12-08 14:06 ` Pete Batard 2017-12-08 14:06 ` [PATCH v3 4/6] ArmPkg/Library/CompilerIntrinsicsLib: " Pete Batard ` (3 subsequent siblings) 6 siblings, 0 replies; 12+ messages in thread From: Pete Batard @ 2017-12-08 14:06 UTC (permalink / raw) To: edk2-devel; +Cc: liming.gao Most of the RVCT assembly can be reused as is for MSFT except for CpuBreakpoint.asm, which we need to force to Arm mode. Contributed-under: TianoCore Contribution Agreement 1.1 Signed-off-by: Pete Batard <pete@akeo.ie> --- MdePkg/Library/BaseLib/Arm/CpuBreakpoint.asm | 5 ++++- MdePkg/Library/BaseLib/BaseLib.inf | 16 +++++++++++++--- 2 files changed, 17 insertions(+), 4 deletions(-) diff --git a/MdePkg/Library/BaseLib/Arm/CpuBreakpoint.asm b/MdePkg/Library/BaseLib/Arm/CpuBreakpoint.asm index 8a8065159bf2..2e508d6f1ad8 100644 --- a/MdePkg/Library/BaseLib/Arm/CpuBreakpoint.asm +++ b/MdePkg/Library/BaseLib/Arm/CpuBreakpoint.asm @@ -16,7 +16,10 @@ EXPORT CpuBreakpoint - AREA Cpu_Breakpoint, CODE, READONLY +; Force ARM mode for this section, as MSFT assembler defaults to THUMB + AREA Cpu_Breakpoint, CODE, READONLY, ARM + + ARM ;/** ; Generates a breakpoint on the CPU. diff --git a/MdePkg/Library/BaseLib/BaseLib.inf b/MdePkg/Library/BaseLib/BaseLib.inf index 320ac457ea3d..a81d56f61421 100644 --- a/MdePkg/Library/BaseLib/BaseLib.inf +++ b/MdePkg/Library/BaseLib/BaseLib.inf @@ -821,8 +821,9 @@ [Sources.EBC] [Sources.ARM] Arm/InternalSwitchStack.c Arm/Unaligned.c - Math64.c | RVCT - + Math64.c | RVCT + Math64.c | MSFT + Arm/SwitchStack.asm | RVCT Arm/SetJumpLongJump.asm | RVCT Arm/DisableInterrupts.asm | RVCT @@ -831,7 +832,16 @@ [Sources.ARM] Arm/CpuPause.asm | RVCT Arm/CpuBreakpoint.asm | RVCT Arm/MemoryFence.asm | RVCT - + + Arm/SwitchStack.asm | MSFT + Arm/SetJumpLongJump.asm | MSFT + Arm/DisableInterrupts.asm | MSFT + Arm/EnableInterrupts.asm | MSFT + Arm/GetInterruptsState.asm | MSFT + Arm/CpuPause.asm | MSFT + Arm/CpuBreakpoint.asm | MSFT + Arm/MemoryFence.asm | MSFT + Arm/Math64.S | GCC Arm/SwitchStack.S | GCC Arm/EnableInterrupts.S | GCC -- 2.9.3.windows.2 ^ permalink raw reply related [flat|nested] 12+ messages in thread
* [PATCH v3 4/6] ArmPkg/Library/CompilerIntrinsicsLib: Enable VS2017/ARM builds 2017-12-08 14:06 [PATCH v3 0/6] Add ARM support for VS2017 Pete Batard ` (2 preceding siblings ...) 2017-12-08 14:06 ` [PATCH v3 3/6] MdePkg/Library/BaseLib: Enable VS2017/ARM builds Pete Batard @ 2017-12-08 14:06 ` Pete Batard 2017-12-10 13:30 ` Gao, Liming 2017-12-12 19:59 ` Ard Biesheuvel 2017-12-08 14:06 ` [PATCH v3 5/6] MdePkg/Include: Add VA list support for VS2017/ARM Pete Batard ` (2 subsequent siblings) 6 siblings, 2 replies; 12+ messages in thread From: Pete Batard @ 2017-12-08 14:06 UTC (permalink / raw) To: edk2-devel; +Cc: liming.gao Introduce CRT assembly replacements for __rt_sdiv, __rt_udiv, __rt_udiv64, __rt_sdiv64, __rt_srsh, memcpy and memset. Contributed-under: TianoCore Contribution Agreement 1.1 Signed-off-by: Pete Batard <pete@akeo.ie> --- ArmPkg/Library/CompilerIntrinsicsLib/Arm/rtdiv.asm | 255 ++++++++++++++++++++ ArmPkg/Library/CompilerIntrinsicsLib/Arm/rtsrsh.asm | 45 ++++ ArmPkg/Library/CompilerIntrinsicsLib/CompilerIntrinsicsLib.inf | 13 +- ArmPkg/Library/CompilerIntrinsicsLib/memcpy_ms.c | 34 +++ ArmPkg/Library/CompilerIntrinsicsLib/memset_ms.c | 33 +++ 5 files changed, 378 insertions(+), 2 deletions(-) diff --git a/ArmPkg/Library/CompilerIntrinsicsLib/Arm/rtdiv.asm b/ArmPkg/Library/CompilerIntrinsicsLib/Arm/rtdiv.asm new file mode 100644 index 000000000000..096dc6317318 --- /dev/null +++ b/ArmPkg/Library/CompilerIntrinsicsLib/Arm/rtdiv.asm @@ -0,0 +1,255 @@ +///** @file +// +// This code provides replacement for MSVC CRT division functions +// +// Copyright (c) 2017, Pete Batard. All rights reserved.<BR> +// Based on generated assembly of ReactOS' sdk/lib/crt/math/arm/__rt_###div.c, +// Copyright (c) Timo Kreuzer. All rights reserved.<BR> +// +// This program and the accompanying materials +// are licensed and made available under the terms and conditions of the BSD License +// which accompanies this distribution. The full text of the license may be found at +// http://opensource.org/licenses/bsd-license.php +// +// THE PROGRAM IS DISTRIBUTED UNDER THE BSD LICENSE ON AN "AS IS" BASIS, +// WITHOUT WARRANTIES OR REPRESENTATIONS OF ANY KIND, EITHER EXPRESS OR IMPLIED. +// +//**/ + + EXPORT _fltused + EXPORT __brkdiv0 + + EXPORT __rt_sdiv + EXPORT __rt_udiv + EXPORT __rt_udiv64 + EXPORT __rt_sdiv64 + + AREA Math, CODE, READONLY + +_fltused + dcd 0x9875 + +__brkdiv0 + udf #249 + +// +// uint64_t __rt_udiv(uint32_t divisor, uint32_t dividend) +// + +__rt_udiv + cmp r0, #0 + beq __brkdiv0 + push {r3-r5,lr} + mov r5,r0 + mov r4,r1 + cmp r5,r4 + it hi + movhi r0,#0 + bhi __rt_udiv_label3 + clz r2,r5 + clz r3,r4 + subs r3,r2,r3 + movs r1,#1 + lsl r2,r5,r3 + lsl r3,r1,r3 + movs r0,#0 +__rt_udiv_label1 + cmp r4,r2 + bcc __rt_udiv_label2 + orrs r0,r0,r3 + subs r4,r4,r2 +__rt_udiv_label2 + lsrs r2,r2,#1 + lsrs r3,r3,#1 + bne __rt_udiv_label1 +__rt_udiv_label3 + mov r1,r4 + pop {r3-r5,pc} + +// +// uint64_t __rt_sdiv(int32_t divisor, int32_t dividend) +// + +__rt_sdiv + cmp r0, #0 + beq __brkdiv0 + push {r4-r6,lr} + mov r4,r1 + ands r6,r0,#0x80000000 + it ne + rsbne r4,r4,#0 + mov r5,r0 + rsbs r5,r5,#0 + cmp r5,r4 + it hi + movhi r0,#0 + bhi __rt_sdiv_label3 + clz r2,r5 + clz r3,r4 + subs r3,r2,r3 + movs r1,#1 + lsl r2,r5,r3 + lsl r3,r1,r3 + movs r0,#0 +__rt_sdiv_label1 + cmp r4,r2 + bcc __rt_sdiv_label2 + orrs r0,r0,r3 + subs r4,r4,r2 +__rt_sdiv_label2 + lsrs r2,r2,#1 + lsrs r3,r3,#1 + bne __rt_sdiv_label1 +__rt_sdiv_label3 + cbz r6,__rt_sdiv_label4 + rsbs r4,r4,#0 +__rt_sdiv_label4 + mov r1,r4 + pop {r4-r6,pc} + +// +// typedef struct { +// uint64_t quotient; +// uint64_t modulus; +// } udiv64_result_t; +// +// void __rt_udiv64_internal(udiv64_result_t *result, uint64_t divisor, uint64_t dividend) +// + +__rt_udiv64_internal + orrs r1,r2,r3 + beq __brkdiv0 + push {r4-r8,lr} + mov r7,r3 + mov r6,r2 + mov r4,r0 + ldrd r0,r5,[sp,#0x18] + cmp r7,r5 + bcc __rt_udiv64_internal_label2 + bhi __rt_udiv64_internal_label1 + cmp r6,r0 + bls __rt_udiv64_internal_label2 +__rt_udiv64_internal_label1 + movs r3,#0 + strd r3,r3,[r4] + b __rt_udiv64_internal_label8 +__rt_udiv64_internal_label2 + clz r2,r7 + cmp r2,#0x20 + bne __rt_udiv64_internal_label3 + clz r3,r6 + add r2,r2,r3 +__rt_udiv64_internal_label3 + clz r1,r5 ; + cmp r1,#0x20 + bne __rt_udiv64_internal_label4 + clz r3,r0 + add r1,r1,r3 +__rt_udiv64_internal_label4 + subs r1,r2,r1 + rsb r3,r1,#0x20 + lsr r3,r6,r3 + lsl r2,r7,r1 + orrs r2,r2,r3 + sub r3,r1,#0x20 + lsl r3,r6,r3 + orrs r2,r2,r3 + lsl r7,r6,r1 + sub r3,r1,#0x20 + movs r6,#1 + lsls r6,r6,r3 + movs r3,#1 + mov lr,#0 + lsl r1,r3,r1 + mov r8,lr +__rt_udiv64_internal_label5 + cmp r5,r2 + bcc __rt_udiv64_internal_label7 + bhi __rt_udiv64_internal_label6 + cmp r0,r7 + bcc __rt_udiv64_internal_label7 +__rt_udiv64_internal_label6 + subs r0,r0,r7 + sbcs r5,r5,r2 + orr lr,lr,r1 + orr r8,r8,r6 +__rt_udiv64_internal_label7 + lsls r3,r2,#0x1F + orr r7,r3,r7,lsr #1 + lsls r3,r6,#0x1F + orr r1,r3,r1,lsr #1 + lsrs r6,r6,#1 + lsrs r2,r2,#1 + orrs r3,r1,r6 + bne __rt_udiv64_internal_label5 + strd lr,r8,[r4] +__rt_udiv64_internal_label8 + str r5,[r4,#0xC] + str r0,[r4,#8] + pop {r4-r8,pc} + +// +// {int64_t, int64_t} __rt_sdiv64(int64_t divisor, int64_t dividend) +// + +__rt_sdiv64 + push {r4-r6,lr} + sub sp,sp,#0x18 + and r6,r1,#0x80000000 + movs r4,r6 + mov r5,r0 + beq __rt_sdiv64_label1 + movs r0,#0 + rsbs r2,r2,#0 + sbc r3,r0,r3 +__rt_sdiv64_label1 + movs r4,r6 + beq __rt_sdiv64_label2 + movs r0,#0 + rsbs r5,r5,#0 + sbc r1,r0,r1 +__rt_sdiv64_label2 + str r2,[sp] + str r3,[sp,#4] + mov r3,r1 + mov r2,r5 + add r0,sp,#8 + bl __rt_udiv64_internal + movs r3,r6 + beq __rt_sdiv64_label3 + ldrd r3,r2,[sp,#0x10] + movs r1,#0 + rsbs r3,r3,#0 + sbcs r1,r1,r2 + b __rt_sdiv64_label4 +__rt_sdiv64_label3 + ldrd r3,r1,[sp,#0x10] +__rt_sdiv64_label4 + mov r2,r3 + ldr r0,[sp,#8] + mov r3,r1 + ldr r1,[sp,#0xC] + add sp,sp,#0x18 + pop {r4-r6,pc} + +// +// {uint64_t, uint64_t} __rt_udiv64(uint64_t divisor, uint64_t dividend) +// + +__rt_udiv64 + push {r4,r5,lr} + sub sp,sp,#0x1C + mov r4,r2 + mov r2,r0 + mov r5,r3 + add r0,sp,#8 + mov r3,r1 + str r4,[sp] + str r5,[sp,#4] + bl __rt_udiv64_internal + ldrd r2,r3,[sp,#0x10] + ldrd r0,r1,[sp,#8] + add sp,sp,#0x1C + pop {r4,r5,pc} + + END diff --git a/ArmPkg/Library/CompilerIntrinsicsLib/Arm/rtsrsh.asm b/ArmPkg/Library/CompilerIntrinsicsLib/Arm/rtsrsh.asm new file mode 100644 index 000000000000..2332dda823f2 --- /dev/null +++ b/ArmPkg/Library/CompilerIntrinsicsLib/Arm/rtsrsh.asm @@ -0,0 +1,45 @@ +///** @file +// +// This code provides replacement for MSVC CRT __rt_srsh +// +// Copyright (c) Timo Kreuzer. All rights reserved.<BR> +// +// This program and the accompanying materials +// are licensed and made available under the terms and conditions of the BSD License +// which accompanies this distribution. The full text of the license may be found at +// http://opensource.org/licenses/bsd-license.php +// +// THE PROGRAM IS DISTRIBUTED UNDER THE BSD LICENSE ON AN "AS IS" BASIS, +// WITHOUT WARRANTIES OR REPRESENTATIONS OF ANY KIND, EITHER EXPRESS OR IMPLIED. +// +//**/ + + EXPORT __rt_srsh + + AREA Math, CODE, READONLY + +// +// int64_t __rt_srsh(int64_t value, uint32_t shift); +// + +__rt_srsh + rsbs r3, r2, #32 + bmi __rt_srsh_label1 + lsr r0, r0, r2 + lsl r3, r1, r3 + orr r0, r0, r3 + asr r1, r1, r2 + bx lr +__rt_srsh_label1 + cmp r2, 64 + bhs __rt_srsh_label2 + sub r3, r2, #32 + asr r0, r1, r3 + asr r1, r1, #32 + bx lr +__rt_srsh_label2 + asr r1, r1, #32 + mov r0, r1 + bx lr + + END diff --git a/ArmPkg/Library/CompilerIntrinsicsLib/CompilerIntrinsicsLib.inf b/ArmPkg/Library/CompilerIntrinsicsLib/CompilerIntrinsicsLib.inf index 44333141a70a..0dacc5e5117d 100644 --- a/ArmPkg/Library/CompilerIntrinsicsLib/CompilerIntrinsicsLib.inf +++ b/ArmPkg/Library/CompilerIntrinsicsLib/CompilerIntrinsicsLib.inf @@ -23,8 +23,12 @@ [Defines] LIBRARY_CLASS = CompilerIntrinsicsLib [Sources] - memcpy.c - memset.c + memcpy.c | RVCT + memcpy.c | GCC + memcpy_ms.c | MSFT + memset.c | RVCT + memset.c | GCC + memset_ms.c | MSFT [Sources.ARM] Arm/mullu.asm | RVCT @@ -94,6 +98,8 @@ [Sources.ARM] Arm/llsr.S | GCC Arm/llsl.S | GCC + Arm/rtdiv.asm | MSFT + Arm/rtsrsh.asm | MSFT [Packages] MdePkg/MdePkg.dec @@ -101,3 +107,6 @@ [Packages] [LibraryClasses] +[BuildOptions] + MSFT:*_*_ARM_CC_FLAGS = /GL- + MSFT:*_*_AARCH64_CC_FLAGS = /GL- diff --git a/ArmPkg/Library/CompilerIntrinsicsLib/memcpy_ms.c b/ArmPkg/Library/CompilerIntrinsicsLib/memcpy_ms.c new file mode 100644 index 000000000000..90bbbb930d31 --- /dev/null +++ b/ArmPkg/Library/CompilerIntrinsicsLib/memcpy_ms.c @@ -0,0 +1,34 @@ +//------------------------------------------------------------------------------ +// +// Copyright (c) 2017, Pete Batard. All rights reserved.<BR> +// +// This program and the accompanying materials are licensed and made +// available under the terms and conditions of the BSD License which +// accompanies this distribution. The full text of the license may be +// found at http://opensource.org/licenses/bsd-license.php +// +// THE PROGRAM IS DISTRIBUTED UNDER THE BSD LICENSE ON AN "AS IS" BASIS, +// WITHOUT WARRANTIES OR REPRESENTATIONS OF ANY KIND, EITHER EXPRESS OR +// IMPLIED. +// +//------------------------------------------------------------------------------ + +#if defined(_M_ARM64) +typedef unsigned __int64 size_t; +#else +typedef unsigned __int32 size_t; +#endif + +void* memcpy(void *, const void *, size_t); +#pragma intrinsic(memcpy) +#pragma function(memcpy) +void* memcpy(void *dest, const void *src, size_t n) +{ + unsigned char *d = dest; + unsigned char const *s = src; + + while (n--) + *d++ = *s++; + + return dest; +} diff --git a/ArmPkg/Library/CompilerIntrinsicsLib/memset_ms.c b/ArmPkg/Library/CompilerIntrinsicsLib/memset_ms.c new file mode 100644 index 000000000000..64205e5d1012 --- /dev/null +++ b/ArmPkg/Library/CompilerIntrinsicsLib/memset_ms.c @@ -0,0 +1,33 @@ +//------------------------------------------------------------------------------ +// +// Copyright (c) 2017, Pete Batard. All rights reserved.<BR> +// +// This program and the accompanying materials are licensed and made +// available under the terms and conditions of the BSD License which +// accompanies this distribution. The full text of the license may be +// found at http://opensource.org/licenses/bsd-license.php +// +// THE PROGRAM IS DISTRIBUTED UNDER THE BSD LICENSE ON AN "AS IS" BASIS, +// WITHOUT WARRANTIES OR REPRESENTATIONS OF ANY KIND, EITHER EXPRESS OR +// IMPLIED. +// +//------------------------------------------------------------------------------ + +#if defined(_M_ARM64) +typedef unsigned __int64 size_t; +#else +typedef unsigned __int32 size_t; +#endif + +void* memset(void *, int, size_t); +#pragma intrinsic(memset) +#pragma function(memset) +void *memset(void *s, int c, size_t n) +{ + unsigned char *d = s; + + while (n--) + *d++ = (unsigned char)c; + + return s; +} -- 2.9.3.windows.2 ^ permalink raw reply related [flat|nested] 12+ messages in thread
* Re: [PATCH v3 4/6] ArmPkg/Library/CompilerIntrinsicsLib: Enable VS2017/ARM builds 2017-12-08 14:06 ` [PATCH v3 4/6] ArmPkg/Library/CompilerIntrinsicsLib: " Pete Batard @ 2017-12-10 13:30 ` Gao, Liming 2017-12-12 19:59 ` Ard Biesheuvel 1 sibling, 0 replies; 12+ messages in thread From: Gao, Liming @ 2017-12-10 13:30 UTC (permalink / raw) To: Pete Batard, edk2-devel@lists.01.org Cc: leif.lindholm@linaro.org, ard.biesheuvel@linaro.org Include ArmPkg maintainers for review. > -----Original Message----- > From: Pete Batard [mailto:pete@akeo.ie] > Sent: Friday, December 8, 2017 10:06 PM > To: edk2-devel@lists.01.org > Cc: Gao, Liming <liming.gao@intel.com> > Subject: [PATCH v3 4/6] ArmPkg/Library/CompilerIntrinsicsLib: Enable VS2017/ARM builds > > Introduce CRT assembly replacements for __rt_sdiv, __rt_udiv, > __rt_udiv64, __rt_sdiv64, __rt_srsh, memcpy and memset. > > Contributed-under: TianoCore Contribution Agreement 1.1 > Signed-off-by: Pete Batard <pete@akeo.ie> > --- > ArmPkg/Library/CompilerIntrinsicsLib/Arm/rtdiv.asm | 255 ++++++++++++++++++++ > ArmPkg/Library/CompilerIntrinsicsLib/Arm/rtsrsh.asm | 45 ++++ > ArmPkg/Library/CompilerIntrinsicsLib/CompilerIntrinsicsLib.inf | 13 +- > ArmPkg/Library/CompilerIntrinsicsLib/memcpy_ms.c | 34 +++ > ArmPkg/Library/CompilerIntrinsicsLib/memset_ms.c | 33 +++ > 5 files changed, 378 insertions(+), 2 deletions(-) > > diff --git a/ArmPkg/Library/CompilerIntrinsicsLib/Arm/rtdiv.asm b/ArmPkg/Library/CompilerIntrinsicsLib/Arm/rtdiv.asm > new file mode 100644 > index 000000000000..096dc6317318 > --- /dev/null > +++ b/ArmPkg/Library/CompilerIntrinsicsLib/Arm/rtdiv.asm > @@ -0,0 +1,255 @@ > +///** @file > +// > +// This code provides replacement for MSVC CRT division functions > +// > +// Copyright (c) 2017, Pete Batard. All rights reserved.<BR> > +// Based on generated assembly of ReactOS' sdk/lib/crt/math/arm/__rt_###div.c, > +// Copyright (c) Timo Kreuzer. All rights reserved.<BR> > +// > +// This program and the accompanying materials > +// are licensed and made available under the terms and conditions of the BSD License > +// which accompanies this distribution. The full text of the license may be found at > +// http://opensource.org/licenses/bsd-license.php > +// > +// THE PROGRAM IS DISTRIBUTED UNDER THE BSD LICENSE ON AN "AS IS" BASIS, > +// WITHOUT WARRANTIES OR REPRESENTATIONS OF ANY KIND, EITHER EXPRESS OR IMPLIED. > +// > +//**/ > + > + EXPORT _fltused > + EXPORT __brkdiv0 > + > + EXPORT __rt_sdiv > + EXPORT __rt_udiv > + EXPORT __rt_udiv64 > + EXPORT __rt_sdiv64 > + > + AREA Math, CODE, READONLY > + > +_fltused > + dcd 0x9875 > + > +__brkdiv0 > + udf #249 > + > +// > +// uint64_t __rt_udiv(uint32_t divisor, uint32_t dividend) > +// > + > +__rt_udiv > + cmp r0, #0 > + beq __brkdiv0 > + push {r3-r5,lr} > + mov r5,r0 > + mov r4,r1 > + cmp r5,r4 > + it hi > + movhi r0,#0 > + bhi __rt_udiv_label3 > + clz r2,r5 > + clz r3,r4 > + subs r3,r2,r3 > + movs r1,#1 > + lsl r2,r5,r3 > + lsl r3,r1,r3 > + movs r0,#0 > +__rt_udiv_label1 > + cmp r4,r2 > + bcc __rt_udiv_label2 > + orrs r0,r0,r3 > + subs r4,r4,r2 > +__rt_udiv_label2 > + lsrs r2,r2,#1 > + lsrs r3,r3,#1 > + bne __rt_udiv_label1 > +__rt_udiv_label3 > + mov r1,r4 > + pop {r3-r5,pc} > + > +// > +// uint64_t __rt_sdiv(int32_t divisor, int32_t dividend) > +// > + > +__rt_sdiv > + cmp r0, #0 > + beq __brkdiv0 > + push {r4-r6,lr} > + mov r4,r1 > + ands r6,r0,#0x80000000 > + it ne > + rsbne r4,r4,#0 > + mov r5,r0 > + rsbs r5,r5,#0 > + cmp r5,r4 > + it hi > + movhi r0,#0 > + bhi __rt_sdiv_label3 > + clz r2,r5 > + clz r3,r4 > + subs r3,r2,r3 > + movs r1,#1 > + lsl r2,r5,r3 > + lsl r3,r1,r3 > + movs r0,#0 > +__rt_sdiv_label1 > + cmp r4,r2 > + bcc __rt_sdiv_label2 > + orrs r0,r0,r3 > + subs r4,r4,r2 > +__rt_sdiv_label2 > + lsrs r2,r2,#1 > + lsrs r3,r3,#1 > + bne __rt_sdiv_label1 > +__rt_sdiv_label3 > + cbz r6,__rt_sdiv_label4 > + rsbs r4,r4,#0 > +__rt_sdiv_label4 > + mov r1,r4 > + pop {r4-r6,pc} > + > +// > +// typedef struct { > +// uint64_t quotient; > +// uint64_t modulus; > +// } udiv64_result_t; > +// > +// void __rt_udiv64_internal(udiv64_result_t *result, uint64_t divisor, uint64_t dividend) > +// > + > +__rt_udiv64_internal > + orrs r1,r2,r3 > + beq __brkdiv0 > + push {r4-r8,lr} > + mov r7,r3 > + mov r6,r2 > + mov r4,r0 > + ldrd r0,r5,[sp,#0x18] > + cmp r7,r5 > + bcc __rt_udiv64_internal_label2 > + bhi __rt_udiv64_internal_label1 > + cmp r6,r0 > + bls __rt_udiv64_internal_label2 > +__rt_udiv64_internal_label1 > + movs r3,#0 > + strd r3,r3,[r4] > + b __rt_udiv64_internal_label8 > +__rt_udiv64_internal_label2 > + clz r2,r7 > + cmp r2,#0x20 > + bne __rt_udiv64_internal_label3 > + clz r3,r6 > + add r2,r2,r3 > +__rt_udiv64_internal_label3 > + clz r1,r5 ; > + cmp r1,#0x20 > + bne __rt_udiv64_internal_label4 > + clz r3,r0 > + add r1,r1,r3 > +__rt_udiv64_internal_label4 > + subs r1,r2,r1 > + rsb r3,r1,#0x20 > + lsr r3,r6,r3 > + lsl r2,r7,r1 > + orrs r2,r2,r3 > + sub r3,r1,#0x20 > + lsl r3,r6,r3 > + orrs r2,r2,r3 > + lsl r7,r6,r1 > + sub r3,r1,#0x20 > + movs r6,#1 > + lsls r6,r6,r3 > + movs r3,#1 > + mov lr,#0 > + lsl r1,r3,r1 > + mov r8,lr > +__rt_udiv64_internal_label5 > + cmp r5,r2 > + bcc __rt_udiv64_internal_label7 > + bhi __rt_udiv64_internal_label6 > + cmp r0,r7 > + bcc __rt_udiv64_internal_label7 > +__rt_udiv64_internal_label6 > + subs r0,r0,r7 > + sbcs r5,r5,r2 > + orr lr,lr,r1 > + orr r8,r8,r6 > +__rt_udiv64_internal_label7 > + lsls r3,r2,#0x1F > + orr r7,r3,r7,lsr #1 > + lsls r3,r6,#0x1F > + orr r1,r3,r1,lsr #1 > + lsrs r6,r6,#1 > + lsrs r2,r2,#1 > + orrs r3,r1,r6 > + bne __rt_udiv64_internal_label5 > + strd lr,r8,[r4] > +__rt_udiv64_internal_label8 > + str r5,[r4,#0xC] > + str r0,[r4,#8] > + pop {r4-r8,pc} > + > +// > +// {int64_t, int64_t} __rt_sdiv64(int64_t divisor, int64_t dividend) > +// > + > +__rt_sdiv64 > + push {r4-r6,lr} > + sub sp,sp,#0x18 > + and r6,r1,#0x80000000 > + movs r4,r6 > + mov r5,r0 > + beq __rt_sdiv64_label1 > + movs r0,#0 > + rsbs r2,r2,#0 > + sbc r3,r0,r3 > +__rt_sdiv64_label1 > + movs r4,r6 > + beq __rt_sdiv64_label2 > + movs r0,#0 > + rsbs r5,r5,#0 > + sbc r1,r0,r1 > +__rt_sdiv64_label2 > + str r2,[sp] > + str r3,[sp,#4] > + mov r3,r1 > + mov r2,r5 > + add r0,sp,#8 > + bl __rt_udiv64_internal > + movs r3,r6 > + beq __rt_sdiv64_label3 > + ldrd r3,r2,[sp,#0x10] > + movs r1,#0 > + rsbs r3,r3,#0 > + sbcs r1,r1,r2 > + b __rt_sdiv64_label4 > +__rt_sdiv64_label3 > + ldrd r3,r1,[sp,#0x10] > +__rt_sdiv64_label4 > + mov r2,r3 > + ldr r0,[sp,#8] > + mov r3,r1 > + ldr r1,[sp,#0xC] > + add sp,sp,#0x18 > + pop {r4-r6,pc} > + > +// > +// {uint64_t, uint64_t} __rt_udiv64(uint64_t divisor, uint64_t dividend) > +// > + > +__rt_udiv64 > + push {r4,r5,lr} > + sub sp,sp,#0x1C > + mov r4,r2 > + mov r2,r0 > + mov r5,r3 > + add r0,sp,#8 > + mov r3,r1 > + str r4,[sp] > + str r5,[sp,#4] > + bl __rt_udiv64_internal > + ldrd r2,r3,[sp,#0x10] > + ldrd r0,r1,[sp,#8] > + add sp,sp,#0x1C > + pop {r4,r5,pc} > + > + END > diff --git a/ArmPkg/Library/CompilerIntrinsicsLib/Arm/rtsrsh.asm b/ArmPkg/Library/CompilerIntrinsicsLib/Arm/rtsrsh.asm > new file mode 100644 > index 000000000000..2332dda823f2 > --- /dev/null > +++ b/ArmPkg/Library/CompilerIntrinsicsLib/Arm/rtsrsh.asm > @@ -0,0 +1,45 @@ > +///** @file > +// > +// This code provides replacement for MSVC CRT __rt_srsh > +// > +// Copyright (c) Timo Kreuzer. All rights reserved.<BR> > +// > +// This program and the accompanying materials > +// are licensed and made available under the terms and conditions of the BSD License > +// which accompanies this distribution. The full text of the license may be found at > +// http://opensource.org/licenses/bsd-license.php > +// > +// THE PROGRAM IS DISTRIBUTED UNDER THE BSD LICENSE ON AN "AS IS" BASIS, > +// WITHOUT WARRANTIES OR REPRESENTATIONS OF ANY KIND, EITHER EXPRESS OR IMPLIED. > +// > +//**/ > + > + EXPORT __rt_srsh > + > + AREA Math, CODE, READONLY > + > +// > +// int64_t __rt_srsh(int64_t value, uint32_t shift); > +// > + > +__rt_srsh > + rsbs r3, r2, #32 > + bmi __rt_srsh_label1 > + lsr r0, r0, r2 > + lsl r3, r1, r3 > + orr r0, r0, r3 > + asr r1, r1, r2 > + bx lr > +__rt_srsh_label1 > + cmp r2, 64 > + bhs __rt_srsh_label2 > + sub r3, r2, #32 > + asr r0, r1, r3 > + asr r1, r1, #32 > + bx lr > +__rt_srsh_label2 > + asr r1, r1, #32 > + mov r0, r1 > + bx lr > + > + END > diff --git a/ArmPkg/Library/CompilerIntrinsicsLib/CompilerIntrinsicsLib.inf > b/ArmPkg/Library/CompilerIntrinsicsLib/CompilerIntrinsicsLib.inf > index 44333141a70a..0dacc5e5117d 100644 > --- a/ArmPkg/Library/CompilerIntrinsicsLib/CompilerIntrinsicsLib.inf > +++ b/ArmPkg/Library/CompilerIntrinsicsLib/CompilerIntrinsicsLib.inf > @@ -23,8 +23,12 @@ [Defines] > LIBRARY_CLASS = CompilerIntrinsicsLib > > [Sources] > - memcpy.c > - memset.c > + memcpy.c | RVCT > + memcpy.c | GCC > + memcpy_ms.c | MSFT > + memset.c | RVCT > + memset.c | GCC > + memset_ms.c | MSFT > > [Sources.ARM] > Arm/mullu.asm | RVCT > @@ -94,6 +98,8 @@ [Sources.ARM] > Arm/llsr.S | GCC > Arm/llsl.S | GCC > > + Arm/rtdiv.asm | MSFT > + Arm/rtsrsh.asm | MSFT > > [Packages] > MdePkg/MdePkg.dec > @@ -101,3 +107,6 @@ [Packages] > > [LibraryClasses] > > +[BuildOptions] > + MSFT:*_*_ARM_CC_FLAGS = /GL- > + MSFT:*_*_AARCH64_CC_FLAGS = /GL- > diff --git a/ArmPkg/Library/CompilerIntrinsicsLib/memcpy_ms.c b/ArmPkg/Library/CompilerIntrinsicsLib/memcpy_ms.c > new file mode 100644 > index 000000000000..90bbbb930d31 > --- /dev/null > +++ b/ArmPkg/Library/CompilerIntrinsicsLib/memcpy_ms.c > @@ -0,0 +1,34 @@ > +//------------------------------------------------------------------------------ > +// > +// Copyright (c) 2017, Pete Batard. All rights reserved.<BR> > +// > +// This program and the accompanying materials are licensed and made > +// available under the terms and conditions of the BSD License which > +// accompanies this distribution. The full text of the license may be > +// found at http://opensource.org/licenses/bsd-license.php > +// > +// THE PROGRAM IS DISTRIBUTED UNDER THE BSD LICENSE ON AN "AS IS" BASIS, > +// WITHOUT WARRANTIES OR REPRESENTATIONS OF ANY KIND, EITHER EXPRESS OR > +// IMPLIED. > +// > +//------------------------------------------------------------------------------ > + > +#if defined(_M_ARM64) > +typedef unsigned __int64 size_t; > +#else > +typedef unsigned __int32 size_t; > +#endif > + > +void* memcpy(void *, const void *, size_t); > +#pragma intrinsic(memcpy) > +#pragma function(memcpy) > +void* memcpy(void *dest, const void *src, size_t n) > +{ > + unsigned char *d = dest; > + unsigned char const *s = src; > + > + while (n--) > + *d++ = *s++; > + > + return dest; > +} > diff --git a/ArmPkg/Library/CompilerIntrinsicsLib/memset_ms.c b/ArmPkg/Library/CompilerIntrinsicsLib/memset_ms.c > new file mode 100644 > index 000000000000..64205e5d1012 > --- /dev/null > +++ b/ArmPkg/Library/CompilerIntrinsicsLib/memset_ms.c > @@ -0,0 +1,33 @@ > +//------------------------------------------------------------------------------ > +// > +// Copyright (c) 2017, Pete Batard. All rights reserved.<BR> > +// > +// This program and the accompanying materials are licensed and made > +// available under the terms and conditions of the BSD License which > +// accompanies this distribution. The full text of the license may be > +// found at http://opensource.org/licenses/bsd-license.php > +// > +// THE PROGRAM IS DISTRIBUTED UNDER THE BSD LICENSE ON AN "AS IS" BASIS, > +// WITHOUT WARRANTIES OR REPRESENTATIONS OF ANY KIND, EITHER EXPRESS OR > +// IMPLIED. > +// > +//------------------------------------------------------------------------------ > + > +#if defined(_M_ARM64) > +typedef unsigned __int64 size_t; > +#else > +typedef unsigned __int32 size_t; > +#endif > + > +void* memset(void *, int, size_t); > +#pragma intrinsic(memset) > +#pragma function(memset) > +void *memset(void *s, int c, size_t n) > +{ > + unsigned char *d = s; > + > + while (n--) > + *d++ = (unsigned char)c; > + > + return s; > +} > -- > 2.9.3.windows.2 ^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [PATCH v3 4/6] ArmPkg/Library/CompilerIntrinsicsLib: Enable VS2017/ARM builds 2017-12-08 14:06 ` [PATCH v3 4/6] ArmPkg/Library/CompilerIntrinsicsLib: " Pete Batard 2017-12-10 13:30 ` Gao, Liming @ 2017-12-12 19:59 ` Ard Biesheuvel 2017-12-13 13:56 ` Pete Batard 1 sibling, 1 reply; 12+ messages in thread From: Ard Biesheuvel @ 2017-12-12 19:59 UTC (permalink / raw) To: Pete Batard; +Cc: edk2-devel@lists.01.org, Gao, Liming Hi Pete, On 8 December 2017 at 14:06, Pete Batard <pete@akeo.ie> wrote: > Introduce CRT assembly replacements for __rt_sdiv, __rt_udiv, > __rt_udiv64, __rt_sdiv64, __rt_srsh, memcpy and memset. > > Contributed-under: TianoCore Contribution Agreement 1.1 > Signed-off-by: Pete Batard <pete@akeo.ie> > --- > ArmPkg/Library/CompilerIntrinsicsLib/Arm/rtdiv.asm | 255 ++++++++++++++++++++ > ArmPkg/Library/CompilerIntrinsicsLib/Arm/rtsrsh.asm | 45 ++++ > ArmPkg/Library/CompilerIntrinsicsLib/CompilerIntrinsicsLib.inf | 13 +- > ArmPkg/Library/CompilerIntrinsicsLib/memcpy_ms.c | 34 +++ > ArmPkg/Library/CompilerIntrinsicsLib/memset_ms.c | 33 +++ > 5 files changed, 378 insertions(+), 2 deletions(-) > > diff --git a/ArmPkg/Library/CompilerIntrinsicsLib/Arm/rtdiv.asm b/ArmPkg/Library/CompilerIntrinsicsLib/Arm/rtdiv.asm > new file mode 100644 > index 000000000000..096dc6317318 > --- /dev/null > +++ b/ArmPkg/Library/CompilerIntrinsicsLib/Arm/rtdiv.asm > @@ -0,0 +1,255 @@ > +///** @file > +// > +// This code provides replacement for MSVC CRT division functions > +// > +// Copyright (c) 2017, Pete Batard. All rights reserved.<BR> > +// Based on generated assembly of ReactOS' sdk/lib/crt/math/arm/__rt_###div.c, > +// Copyright (c) Timo Kreuzer. All rights reserved.<BR> > +// > +// This program and the accompanying materials > +// are licensed and made available under the terms and conditions of the BSD License > +// which accompanies this distribution. The full text of the license may be found at > +// http://opensource.org/licenses/bsd-license.php > +// > +// THE PROGRAM IS DISTRIBUTED UNDER THE BSD LICENSE ON AN "AS IS" BASIS, > +// WITHOUT WARRANTIES OR REPRESENTATIONS OF ANY KIND, EITHER EXPRESS OR IMPLIED. > +// > +//**/ > + > + EXPORT _fltused > + EXPORT __brkdiv0 > + Why are these being exported? Are these part of the CRT ABI? > + EXPORT __rt_sdiv > + EXPORT __rt_udiv > + EXPORT __rt_udiv64 > + EXPORT __rt_sdiv64 > + > + AREA Math, CODE, READONLY > + > +_fltused > + dcd 0x9875 > + > +__brkdiv0 > + udf #249 > + This will cause a crash when invoked, in a way that gives very little to go on to figure out what happens. Perhaps branch to a C function here that calls ASSERT() and CpuDeadLoop ? > +// > +// uint64_t __rt_udiv(uint32_t divisor, uint32_t dividend) > +// There are many different division implementations already in the same directory. Do we really need a new one? > + > +__rt_udiv > + cmp r0, #0 > + beq __brkdiv0 > + push {r3-r5,lr} > + mov r5,r0 > + mov r4,r1 > + cmp r5,r4 > + it hi > + movhi r0,#0 > + bhi __rt_udiv_label3 > + clz r2,r5 > + clz r3,r4 > + subs r3,r2,r3 > + movs r1,#1 > + lsl r2,r5,r3 > + lsl r3,r1,r3 > + movs r0,#0 > +__rt_udiv_label1 > + cmp r4,r2 > + bcc __rt_udiv_label2 > + orrs r0,r0,r3 > + subs r4,r4,r2 > +__rt_udiv_label2 > + lsrs r2,r2,#1 > + lsrs r3,r3,#1 > + bne __rt_udiv_label1 > +__rt_udiv_label3 > + mov r1,r4 > + pop {r3-r5,pc} > + > +// > +// uint64_t __rt_sdiv(int32_t divisor, int32_t dividend) > +// > + Does signed division return an unsigned result? And do we really need yet another implementation? > +__rt_sdiv > + cmp r0, #0 > + beq __brkdiv0 > + push {r4-r6,lr} > + mov r4,r1 > + ands r6,r0,#0x80000000 > + it ne > + rsbne r4,r4,#0 > + mov r5,r0 > + rsbs r5,r5,#0 > + cmp r5,r4 > + it hi > + movhi r0,#0 > + bhi __rt_sdiv_label3 > + clz r2,r5 > + clz r3,r4 > + subs r3,r2,r3 > + movs r1,#1 > + lsl r2,r5,r3 > + lsl r3,r1,r3 > + movs r0,#0 > +__rt_sdiv_label1 > + cmp r4,r2 > + bcc __rt_sdiv_label2 > + orrs r0,r0,r3 > + subs r4,r4,r2 > +__rt_sdiv_label2 > + lsrs r2,r2,#1 > + lsrs r3,r3,#1 > + bne __rt_sdiv_label1 > +__rt_sdiv_label3 > + cbz r6,__rt_sdiv_label4 > + rsbs r4,r4,#0 > +__rt_sdiv_label4 > + mov r1,r4 > + pop {r4-r6,pc} > + > +// > +// typedef struct { > +// uint64_t quotient; > +// uint64_t modulus; > +// } udiv64_result_t; > +// > +// void __rt_udiv64_internal(udiv64_result_t *result, uint64_t divisor, uint64_t dividend) > +// > + > +__rt_udiv64_internal > + orrs r1,r2,r3 > + beq __brkdiv0 > + push {r4-r8,lr} > + mov r7,r3 > + mov r6,r2 > + mov r4,r0 > + ldrd r0,r5,[sp,#0x18] > + cmp r7,r5 > + bcc __rt_udiv64_internal_label2 > + bhi __rt_udiv64_internal_label1 > + cmp r6,r0 > + bls __rt_udiv64_internal_label2 > +__rt_udiv64_internal_label1 > + movs r3,#0 > + strd r3,r3,[r4] > + b __rt_udiv64_internal_label8 > +__rt_udiv64_internal_label2 > + clz r2,r7 > + cmp r2,#0x20 > + bne __rt_udiv64_internal_label3 > + clz r3,r6 > + add r2,r2,r3 > +__rt_udiv64_internal_label3 > + clz r1,r5 ; > + cmp r1,#0x20 > + bne __rt_udiv64_internal_label4 > + clz r3,r0 > + add r1,r1,r3 > +__rt_udiv64_internal_label4 > + subs r1,r2,r1 > + rsb r3,r1,#0x20 > + lsr r3,r6,r3 > + lsl r2,r7,r1 > + orrs r2,r2,r3 > + sub r3,r1,#0x20 > + lsl r3,r6,r3 > + orrs r2,r2,r3 > + lsl r7,r6,r1 > + sub r3,r1,#0x20 > + movs r6,#1 > + lsls r6,r6,r3 > + movs r3,#1 > + mov lr,#0 > + lsl r1,r3,r1 > + mov r8,lr > +__rt_udiv64_internal_label5 > + cmp r5,r2 > + bcc __rt_udiv64_internal_label7 > + bhi __rt_udiv64_internal_label6 > + cmp r0,r7 > + bcc __rt_udiv64_internal_label7 > +__rt_udiv64_internal_label6 > + subs r0,r0,r7 > + sbcs r5,r5,r2 > + orr lr,lr,r1 > + orr r8,r8,r6 > +__rt_udiv64_internal_label7 > + lsls r3,r2,#0x1F > + orr r7,r3,r7,lsr #1 > + lsls r3,r6,#0x1F > + orr r1,r3,r1,lsr #1 > + lsrs r6,r6,#1 > + lsrs r2,r2,#1 > + orrs r3,r1,r6 > + bne __rt_udiv64_internal_label5 > + strd lr,r8,[r4] > +__rt_udiv64_internal_label8 > + str r5,[r4,#0xC] > + str r0,[r4,#8] > + pop {r4-r8,pc} > + > +// > +// {int64_t, int64_t} __rt_sdiv64(int64_t divisor, int64_t dividend) > +// > + > +__rt_sdiv64 > + push {r4-r6,lr} > + sub sp,sp,#0x18 > + and r6,r1,#0x80000000 > + movs r4,r6 > + mov r5,r0 > + beq __rt_sdiv64_label1 > + movs r0,#0 > + rsbs r2,r2,#0 > + sbc r3,r0,r3 > +__rt_sdiv64_label1 > + movs r4,r6 > + beq __rt_sdiv64_label2 > + movs r0,#0 > + rsbs r5,r5,#0 > + sbc r1,r0,r1 > +__rt_sdiv64_label2 > + str r2,[sp] > + str r3,[sp,#4] > + mov r3,r1 > + mov r2,r5 > + add r0,sp,#8 > + bl __rt_udiv64_internal > + movs r3,r6 > + beq __rt_sdiv64_label3 > + ldrd r3,r2,[sp,#0x10] > + movs r1,#0 > + rsbs r3,r3,#0 > + sbcs r1,r1,r2 > + b __rt_sdiv64_label4 > +__rt_sdiv64_label3 > + ldrd r3,r1,[sp,#0x10] > +__rt_sdiv64_label4 > + mov r2,r3 > + ldr r0,[sp,#8] > + mov r3,r1 > + ldr r1,[sp,#0xC] > + add sp,sp,#0x18 > + pop {r4-r6,pc} > + > +// > +// {uint64_t, uint64_t} __rt_udiv64(uint64_t divisor, uint64_t dividend) > +// > + > +__rt_udiv64 > + push {r4,r5,lr} > + sub sp,sp,#0x1C > + mov r4,r2 > + mov r2,r0 > + mov r5,r3 > + add r0,sp,#8 > + mov r3,r1 > + str r4,[sp] > + str r5,[sp,#4] > + bl __rt_udiv64_internal > + ldrd r2,r3,[sp,#0x10] > + ldrd r0,r1,[sp,#8] > + add sp,sp,#0x1C > + pop {r4,r5,pc} > + > + END > diff --git a/ArmPkg/Library/CompilerIntrinsicsLib/Arm/rtsrsh.asm b/ArmPkg/Library/CompilerIntrinsicsLib/Arm/rtsrsh.asm > new file mode 100644 > index 000000000000..2332dda823f2 > --- /dev/null > +++ b/ArmPkg/Library/CompilerIntrinsicsLib/Arm/rtsrsh.asm > @@ -0,0 +1,45 @@ > +///** @file > +// > +// This code provides replacement for MSVC CRT __rt_srsh > +// > +// Copyright (c) Timo Kreuzer. All rights reserved.<BR> > +// > +// This program and the accompanying materials > +// are licensed and made available under the terms and conditions of the BSD License > +// which accompanies this distribution. The full text of the license may be found at > +// http://opensource.org/licenses/bsd-license.php > +// > +// THE PROGRAM IS DISTRIBUTED UNDER THE BSD LICENSE ON AN "AS IS" BASIS, > +// WITHOUT WARRANTIES OR REPRESENTATIONS OF ANY KIND, EITHER EXPRESS OR IMPLIED. > +// > +//**/ > + > + EXPORT __rt_srsh > + > + AREA Math, CODE, READONLY > + > +// > +// int64_t __rt_srsh(int64_t value, uint32_t shift); > +// > + Same questions here, basically. > +__rt_srsh > + rsbs r3, r2, #32 > + bmi __rt_srsh_label1 > + lsr r0, r0, r2 > + lsl r3, r1, r3 > + orr r0, r0, r3 > + asr r1, r1, r2 > + bx lr > +__rt_srsh_label1 > + cmp r2, 64 > + bhs __rt_srsh_label2 > + sub r3, r2, #32 > + asr r0, r1, r3 > + asr r1, r1, #32 > + bx lr > +__rt_srsh_label2 > + asr r1, r1, #32 > + mov r0, r1 > + bx lr > + > + END > diff --git a/ArmPkg/Library/CompilerIntrinsicsLib/CompilerIntrinsicsLib.inf b/ArmPkg/Library/CompilerIntrinsicsLib/CompilerIntrinsicsLib.inf > index 44333141a70a..0dacc5e5117d 100644 > --- a/ArmPkg/Library/CompilerIntrinsicsLib/CompilerIntrinsicsLib.inf > +++ b/ArmPkg/Library/CompilerIntrinsicsLib/CompilerIntrinsicsLib.inf > @@ -23,8 +23,12 @@ [Defines] > LIBRARY_CLASS = CompilerIntrinsicsLib > > [Sources] > - memcpy.c > - memset.c > + memcpy.c | RVCT > + memcpy.c | GCC > + memcpy_ms.c | MSFT > + memset.c | RVCT > + memset.c | GCC > + memset_ms.c | MSFT > > [Sources.ARM] > Arm/mullu.asm | RVCT > @@ -94,6 +98,8 @@ [Sources.ARM] > Arm/llsr.S | GCC > Arm/llsl.S | GCC > > + Arm/rtdiv.asm | MSFT > + Arm/rtsrsh.asm | MSFT > > [Packages] > MdePkg/MdePkg.dec > @@ -101,3 +107,6 @@ [Packages] > > [LibraryClasses] > > +[BuildOptions] > + MSFT:*_*_ARM_CC_FLAGS = /GL- > + MSFT:*_*_AARCH64_CC_FLAGS = /GL- > diff --git a/ArmPkg/Library/CompilerIntrinsicsLib/memcpy_ms.c b/ArmPkg/Library/CompilerIntrinsicsLib/memcpy_ms.c > new file mode 100644 > index 000000000000..90bbbb930d31 > --- /dev/null > +++ b/ArmPkg/Library/CompilerIntrinsicsLib/memcpy_ms.c > @@ -0,0 +1,34 @@ > +//------------------------------------------------------------------------------ > +// > +// Copyright (c) 2017, Pete Batard. All rights reserved.<BR> > +// > +// This program and the accompanying materials are licensed and made > +// available under the terms and conditions of the BSD License which > +// accompanies this distribution. The full text of the license may be > +// found at http://opensource.org/licenses/bsd-license.php > +// > +// THE PROGRAM IS DISTRIBUTED UNDER THE BSD LICENSE ON AN "AS IS" BASIS, > +// WITHOUT WARRANTIES OR REPRESENTATIONS OF ANY KIND, EITHER EXPRESS OR > +// IMPLIED. > +// > +//------------------------------------------------------------------------------ > + > +#if defined(_M_ARM64) > +typedef unsigned __int64 size_t; > +#else > +typedef unsigned __int32 size_t; > +#endif > + > +void* memcpy(void *, const void *, size_t); > +#pragma intrinsic(memcpy) > +#pragma function(memcpy) > +void* memcpy(void *dest, const void *src, size_t n) > +{ > + unsigned char *d = dest; > + unsigned char const *s = src; > + > + while (n--) > + *d++ = *s++; > + > + return dest; > +} > diff --git a/ArmPkg/Library/CompilerIntrinsicsLib/memset_ms.c b/ArmPkg/Library/CompilerIntrinsicsLib/memset_ms.c > new file mode 100644 > index 000000000000..64205e5d1012 > --- /dev/null > +++ b/ArmPkg/Library/CompilerIntrinsicsLib/memset_ms.c > @@ -0,0 +1,33 @@ > +//------------------------------------------------------------------------------ > +// > +// Copyright (c) 2017, Pete Batard. All rights reserved.<BR> > +// > +// This program and the accompanying materials are licensed and made > +// available under the terms and conditions of the BSD License which > +// accompanies this distribution. The full text of the license may be > +// found at http://opensource.org/licenses/bsd-license.php > +// > +// THE PROGRAM IS DISTRIBUTED UNDER THE BSD LICENSE ON AN "AS IS" BASIS, > +// WITHOUT WARRANTIES OR REPRESENTATIONS OF ANY KIND, EITHER EXPRESS OR > +// IMPLIED. > +// > +//------------------------------------------------------------------------------ > + > +#if defined(_M_ARM64) > +typedef unsigned __int64 size_t; > +#else > +typedef unsigned __int32 size_t; > +#endif > + > +void* memset(void *, int, size_t); > +#pragma intrinsic(memset) > +#pragma function(memset) > +void *memset(void *s, int c, size_t n) > +{ > + unsigned char *d = s; > + > + while (n--) > + *d++ = (unsigned char)c; > + > + return s; > +} These look fine to me. ^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [PATCH v3 4/6] ArmPkg/Library/CompilerIntrinsicsLib: Enable VS2017/ARM builds 2017-12-12 19:59 ` Ard Biesheuvel @ 2017-12-13 13:56 ` Pete Batard 2017-12-14 14:05 ` Ard Biesheuvel 0 siblings, 1 reply; 12+ messages in thread From: Pete Batard @ 2017-12-13 13:56 UTC (permalink / raw) To: edk2-devel@lists.01.org; +Cc: Ard Biesheuvel, Gao, Liming Hi Ard. Thanks for reviewing this. On 2017.12.12 19:59, Ard Biesheuvel wrote: >> + EXPORT _fltused >> + EXPORT __brkdiv0 >> + > > Why are these being exported? Are these part of the CRT ABI? Good point. I exported them because I was encountering undefined symbols for those when I started working on adding ARM support. But I now suspect they may have come from an export statement from VS headers (which we no longer include), as re-test shows that the exports seem to be superfluous. I will test further to confirm that this is the case, and remove the unneeded exports. >> +__brkdiv0 >> + udf #249 >> + > > This will cause a crash when invoked, in a way that gives very little > to go on to figure out what happens. Yes. I mostly copied what ReactOS was doing here, since it should be similar to what Microsoft does internally. > Perhaps branch to a C function > here that calls ASSERT() and CpuDeadLoop ? I don't mind giving it a try, if we keep the separate .asm files for MSVC. On the other hand, if I am to try to reuse existing assembly, as you suggest below, then I don't think I want to introduce anything that will differ from how divisions by zero are currently handled there. As far as I can tell (for RVCT, which is the only assembly that might be suitable for reuse with MSFT) the current way of handling a zero divisor is to set result and remainder to 0 and return (as per uldiv.asm). For gcc, we also have the following comment in div.S: "@ What to do about division by zero? For now, just return." So, while I agree that calling ASSERT and CpuDeadLoop is probably the better approach, I'd rather have the introduction of this behaviour be the subject of a separate ARM harmonization patch, that applies to all the toolchains, rather than something that's specific to Visual Studio usage. > There are many different division implementations already in the same > directory. Do we really need a new one? I'll look into that. My goal with this series was to make sure that the changes being introduced would not break existing toolchains, which is why I saw it preferable to keep the MSVC intrinsics separate. Now, gcc assembly is not something that can be reused easily with the MS toolchain, so the only possibility are the RVCT .asm files. There's a fair bit of work involved into trying to figure out if and how we can reuse that code, but I'll see what I can do. It'll probably be a few weeks before I get back to this list on that however. Regards, /Pete ^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [PATCH v3 4/6] ArmPkg/Library/CompilerIntrinsicsLib: Enable VS2017/ARM builds 2017-12-13 13:56 ` Pete Batard @ 2017-12-14 14:05 ` Ard Biesheuvel 0 siblings, 0 replies; 12+ messages in thread From: Ard Biesheuvel @ 2017-12-14 14:05 UTC (permalink / raw) To: Pete Batard; +Cc: edk2-devel@lists.01.org, Gao, Liming On 13 December 2017 at 14:56, Pete Batard <pete@akeo.ie> wrote: > Hi Ard. Thanks for reviewing this. > > On 2017.12.12 19:59, Ard Biesheuvel wrote: >>> >>> + EXPORT _fltused >>> + EXPORT __brkdiv0 >>> + >> >> >> Why are these being exported? Are these part of the CRT ABI? > > > Good point. I exported them because I was encountering undefined symbols for > those when I started working on adding ARM support. But I now suspect they > may have come from an export statement from VS headers (which we no longer > include), as re-test shows that the exports seem to be superfluous. > > I will test further to confirm that this is the case, and remove the > unneeded exports. > >>> +__brkdiv0 >>> + udf #249 >>> + >> >> >> This will cause a crash when invoked, in a way that gives very little >> to go on to figure out what happens. > > > Yes. I mostly copied what ReactOS was doing here, since it should be similar > to what Microsoft does internally. > >> Perhaps branch to a C function >> here that calls ASSERT() and CpuDeadLoop ? > > > I don't mind giving it a try, if we keep the separate .asm files for MSVC. > > On the other hand, if I am to try to reuse existing assembly, as you suggest > below, then I don't think I want to introduce anything that will differ from > how divisions by zero are currently handled there. As far as I can tell (for > RVCT, which is the only assembly that might be suitable for reuse with MSFT) > the current way of handling a zero divisor is to set result and remainder to > 0 and return (as per uldiv.asm). For gcc, we also have the following comment > in div.S: > > "@ What to do about division by zero? For now, just return." > > So, while I agree that calling ASSERT and CpuDeadLoop is probably the better > approach, I'd rather have the introduction of this behaviour be the subject > of a separate ARM harmonization patch, that applies to all the toolchains, > rather than something that's specific to Visual Studio usage. > I agree. Reusing code is more important than having one flavour of division that will spot div-by-0. But we really don't want the UDF, because it will crash unconditionally without a hint of what happened. >> There are many different division implementations already in the same >> directory. Do we really need a new one? > > > I'll look into that. My goal with this series was to make sure that the > changes being introduced would not break existing toolchains, which is why I > saw it preferable to keep the MSVC intrinsics separate. > > Now, gcc assembly is not something that can be reused easily with the MS > toolchain, so the only possibility are the RVCT .asm files. > > There's a fair bit of work involved into trying to figure out if and how we > can reuse that code, but I'll see what I can do. It'll probably be a few > weeks before I get back to this list on that however. > I'm prepared to be pragmatic here. I would prefer to have all platforms run the same low level division code, but I understand that it is tightly coupled to the compiler, given that they are intrinsics. So yes, please try to reuse the RVCT code if feasible without a ton of rework, but it seems like we will need to live with some MSFT specific hooks anyway (if I understand the prototypes correctly) ^ permalink raw reply [flat|nested] 12+ messages in thread
* [PATCH v3 5/6] MdePkg/Include: Add VA list support for VS2017/ARM 2017-12-08 14:06 [PATCH v3 0/6] Add ARM support for VS2017 Pete Batard ` (3 preceding siblings ...) 2017-12-08 14:06 ` [PATCH v3 4/6] ArmPkg/Library/CompilerIntrinsicsLib: " Pete Batard @ 2017-12-08 14:06 ` Pete Batard 2017-12-08 14:06 ` [PATCH v3 6/6] BaseTools/Conf: Add VS2017/ARM support Pete Batard 2017-12-08 14:13 ` [PATCH v3 0/6] Add ARM support for VS2017 Gao, Liming 6 siblings, 0 replies; 12+ messages in thread From: Pete Batard @ 2017-12-08 14:06 UTC (permalink / raw) To: edk2-devel; +Cc: liming.gao VA_START, VA_END and VA_COPY are the same as the generic macros. VA_ARG was reverse engineered from MS ARM assembly output. Contributed-under: TianoCore Contribution Agreement 1.1 Signed-off-by: Pete Batard <pete@akeo.ie> --- MdePkg/Include/Base.h | 13 +++++++++++++ 1 file changed, 13 insertions(+) diff --git a/MdePkg/Include/Base.h b/MdePkg/Include/Base.h index 02140a5ac2ee..f3467ecdc22f 100644 --- a/MdePkg/Include/Base.h +++ b/MdePkg/Include/Base.h @@ -631,6 +631,19 @@ struct _LIST_ENTRY { #define VA_COPY(Dest, Start) __va_copy (Dest, Start) +#elif defined(_M_ARM) +// +// MSFT ARM variable argument list support. +// Same as the generic macros below, except for VA_ARG that needs extra adjustment. +// + +typedef char* VA_LIST; + +#define VA_START(Marker, Parameter) (Marker = (VA_LIST) ((UINTN) & (Parameter) + _INT_SIZE_OF(Parameter))) +#define VA_ARG(Marker, TYPE) (*(TYPE *) ((Marker += _INT_SIZE_OF(TYPE) + ((-(INTN)Marker) & (sizeof(TYPE) - 1))) - _INT_SIZE_OF (TYPE))) +#define VA_END(Marker) (Marker = (VA_LIST) 0) +#define VA_COPY(Dest, Start) ((void)((Dest) = (Start))) + #elif defined(__GNUC__) #if defined(MDE_CPU_X64) && !defined(NO_MSABI_VA_FUNCS) -- 2.9.3.windows.2 ^ permalink raw reply related [flat|nested] 12+ messages in thread
* [PATCH v3 6/6] BaseTools/Conf: Add VS2017/ARM support 2017-12-08 14:06 [PATCH v3 0/6] Add ARM support for VS2017 Pete Batard ` (4 preceding siblings ...) 2017-12-08 14:06 ` [PATCH v3 5/6] MdePkg/Include: Add VA list support for VS2017/ARM Pete Batard @ 2017-12-08 14:06 ` Pete Batard 2017-12-08 14:13 ` [PATCH v3 0/6] Add ARM support for VS2017 Gao, Liming 6 siblings, 0 replies; 12+ messages in thread From: Pete Batard @ 2017-12-08 14:06 UTC (permalink / raw) To: edk2-devel; +Cc: liming.gao We duplicate the Assembly-Code-File section from build_rule.template because --convert-hex cannot be used with the MSFT ARM assembler. Contributed-under: TianoCore Contribution Agreement 1.1 Signed-off-by: Pete Batard <pete@akeo.ie> --- BaseTools/Conf/build_rule.template | 31 +++++++++++++++++++- BaseTools/Conf/tools_def.template | 31 ++++++++++++++++++++ 2 files changed, 61 insertions(+), 1 deletion(-) diff --git a/BaseTools/Conf/build_rule.template b/BaseTools/Conf/build_rule.template index 3e6aa8ff0f34..6ea14fb7aa02 100755 --- a/BaseTools/Conf/build_rule.template +++ b/BaseTools/Conf/build_rule.template @@ -182,7 +182,6 @@ <Command> - [Assembly-Code-File.COMMON.COMMON] <InputFile.MSFT, InputFile.INTEL, InputFile.RVCT> ?.asm, ?.Asm, ?.ASM @@ -207,6 +206,36 @@ # For RVCTCYGWIN ASM_FLAGS must be first to work around pathing issues "$(ASM)" $(ASM_FLAGS) -o ${dst} $(INC) ${d_path}(+)${s_base}.iii +[Assembly-Code-File.COMMON.ARM] + # Remove --convert-hex for ARM as it breaks MSFT assemblers + <InputFile.MSFT, InputFile.INTEL, InputFile.RVCT> + ?.asm, ?.Asm, ?.ASM + + <InputFile.GCC, InputFile.GCCLD> + ?.S, ?.s + + <ExtraDependency> + $(MAKE_FILE) + + <OutputFile> + $(OUTPUT_DIR)(+)${s_dir}(+)${s_base}.obj + + <Command.INTEL> + "$(PP)" $(PP_FLAGS) $(INC) ${src} > ${d_path}(+)${s_base}.i + Trim --source-code --convert-hex --trim-long -o ${d_path}(+)${s_base}.iii ${d_path}(+)${s_base}.i + "$(ASM)" /Fo${dst} $(ASM_FLAGS) /I${s_path} $(INC) ${d_path}(+)${s_base}.iii + + <Command.MSFT> + "$(PP)" $(PP_FLAGS) $(INC) ${src} > ${d_path}(+)${s_base}.i + Trim --source-code --trim-long -o ${d_path}(+)${s_base}.iii ${d_path}(+)${s_base}.i + "$(ASM)" /Fo${dst} $(ASM_FLAGS) /I${s_path} $(INC) ${d_path}(+)${s_base}.iii + + <Command.GCC, Command.GCCLD, Command.RVCT> + "$(PP)" $(PP_FLAGS) $(INC) ${src} > ${d_path}(+)${s_base}.i + Trim --trim-long --source-code -o ${d_path}(+)${s_base}.iii ${d_path}(+)${s_base}.i + # For RVCTCYGWIN ASM_FLAGS must be first to work around pathing issues + "$(ASM)" $(ASM_FLAGS) -o ${dst} $(INC) ${d_path}(+)${s_base}.iii + [Nasm-Assembly-Code-File.COMMON.COMMON] <InputFile> ?.nasm diff --git a/BaseTools/Conf/tools_def.template b/BaseTools/Conf/tools_def.template index 91b135c2e569..0bf30f05756f 100755 --- a/BaseTools/Conf/tools_def.template +++ b/BaseTools/Conf/tools_def.template @@ -79,6 +79,7 @@ DEFINE VS2017_HOST = x86 DEFINE VS2017_BIN_HOST = DEF(VS2017_BIN)\HostDEF(VS2017_HOST)\DEF(VS2017_HOST) DEFINE VS2017_BIN_IA32 = DEF(VS2017_BIN)\HostDEF(VS2017_HOST)\x86 DEFINE VS2017_BIN_X64 = DEF(VS2017_BIN)\HostDEF(VS2017_HOST)\x64 +DEFINE VS2017_BIN_ARM = DEF(VS2017_BIN)\HostDEF(VS2017_HOST)\arm DEFINE WINSDK_BIN = ENV(WINSDK_PREFIX) DEFINE WINSDKx86_BIN = ENV(WINSDKx86_PREFIX) @@ -335,6 +336,9 @@ DEFINE DTC_BIN = ENV(DTC_PREFIX)dtc # Required to build platforms or ACPI tables: # Intel(r) ACPI Compiler (iasl.exe) from # https://acpica.org/downloads +# Note: +# Building of XIP firmware images for ARM is not currently supported (only applications). +# /FILEALIGN:4096 and other changes are needed for ARM firmware builds. # DDK3790 -win32- Requires: # Microsoft Windows Server 2003 Driver Development Kit (Microsoft WINDDK) version 3790.1830 # Optional: @@ -4169,6 +4173,33 @@ NOOPT_VS2017_X64_NASM_FLAGS = -O0 -f win64 -g RELEASE_VS2017_X64_DLINK_FLAGS = /NOLOGO /NODEFAULTLIB /IGNORE:4001 /IGNORE:4254 /OPT:REF /OPT:ICF=10 /MAP /ALIGN:32 /SECTION:.xdata,D /SECTION:.pdata,D /Machine:X64 /LTCG /DLL /ENTRY:$(IMAGE_ENTRY_POINT) /SUBSYSTEM:EFI_BOOT_SERVICE_DRIVER /SAFESEH:NO /BASE:0 /DRIVER /MERGE:.rdata=.data NOOPT_VS2017_X64_DLINK_FLAGS = /NOLOGO /NODEFAULTLIB /IGNORE:4001 /OPT:REF /OPT:ICF=10 /MAP /ALIGN:32 /SECTION:.xdata,D /SECTION:.pdata,D /Machine:X64 /LTCG /DLL /ENTRY:$(IMAGE_ENTRY_POINT) /SUBSYSTEM:EFI_BOOT_SERVICE_DRIVER /SAFESEH:NO /BASE:0 /DRIVER /DEBUG +################# +# ARM definitions +################# +*_VS2017_ARM_CC_PATH = DEF(VS2017_BIN_ARM)\cl.exe +*_VS2017_ARM_VFRPP_PATH = DEF(VS2017_BIN_ARM)\cl.exe +*_VS2017_ARM_SLINK_PATH = DEF(VS2017_BIN_ARM)\lib.exe +*_VS2017_ARM_DLINK_PATH = DEF(VS2017_BIN_ARM)\link.exe +*_VS2017_ARM_APP_PATH = DEF(VS2017_BIN_ARM)\cl.exe +*_VS2017_ARM_PP_PATH = DEF(VS2017_BIN_ARM)\cl.exe +*_VS2017_ARM_ASM_PATH = DEF(VS2017_BIN_ARM)\armasm.exe +*_VS2017_ARM_ASLCC_PATH = DEF(VS2017_BIN_ARM)\cl.exe +*_VS2017_ARM_ASLPP_PATH = DEF(VS2017_BIN_ARM)\cl.exe +*_VS2017_ARM_ASLDLINK_PATH = DEF(VS2017_BIN_ARM)\link.exe + + *_VS2017_ARM_MAKE_FLAGS = /nologo + DEBUG_VS2017_ARM_CC_FLAGS = /nologo /c /WX /GS- /W4 /Gs32768 /D UNICODE /O1b2 /GL /FIAutoGen.h /EHs-c- /GR- /GF /Gy /Zi /Gm /Gw /Oi- +RELEASE_VS2017_ARM_CC_FLAGS = /nologo /c /WX /GS- /W4 /Gs32768 /D UNICODE /O1b2 /GL /FIAutoGen.h /EHs-c- /GR- /GF /Gw /Oi- +NOOPT_VS2017_ARM_CC_FLAGS = /nologo /c /WX /GS- /W4 /Gs32768 /D UNICODE /FIAutoGen.h /EHs-c- /GR- /GF /Gy /Zi /Gm /Od /Oi- + + DEBUG_VS2017_ARM_ASM_FLAGS = /nologo /g +RELEASE_VS2017_ARM_ASM_FLAGS = /nologo +NOOPT_VS2017_ARM_ASM_FLAGS = /nologo + + DEBUG_VS2017_ARM_DLINK_FLAGS = /NOLOGO /NODEFAULTLIB /IGNORE:4001 /OPT:REF /OPT:ICF=10 /MAP /SECTION:.xdata,D /SECTION:.pdata,D /MACHINE:ARM /LTCG /DLL /ENTRY:$(IMAGE_ENTRY_POINT) /SUBSYSTEM:EFI_BOOT_SERVICE_DRIVER /SAFESEH:NO /BASE:0 /DRIVER /DEBUG +RELEASE_VS2017_ARM_DLINK_FLAGS = /NOLOGO /NODEFAULTLIB /IGNORE:4001 /IGNORE:4254 /OPT:REF /OPT:ICF=10 /MAP /SECTION:.xdata,D /SECTION:.pdata,D /MACHINE:ARM /LTCG /DLL /ENTRY:$(IMAGE_ENTRY_POINT) /SUBSYSTEM:EFI_BOOT_SERVICE_DRIVER /SAFESEH:NO /BASE:0 /DRIVER /MERGE:.rdata=.data +NOOPT_VS2017_ARM_DLINK_FLAGS = /NOLOGO /NODEFAULTLIB /IGNORE:4001 /OPT:REF /OPT:ICF=10 /MAP /SECTION:.xdata,D /SECTION:.pdata,D /MACHINE:ARM /LTCG /DLL /ENTRY:$(IMAGE_ENTRY_POINT) /SUBSYSTEM:EFI_BOOT_SERVICE_DRIVER /SAFESEH:NO /BASE:0 /DRIVER /DEBUG + ################## # EBC definitions ################## -- 2.9.3.windows.2 ^ permalink raw reply related [flat|nested] 12+ messages in thread
* Re: [PATCH v3 0/6] Add ARM support for VS2017 2017-12-08 14:06 [PATCH v3 0/6] Add ARM support for VS2017 Pete Batard ` (5 preceding siblings ...) 2017-12-08 14:06 ` [PATCH v3 6/6] BaseTools/Conf: Add VS2017/ARM support Pete Batard @ 2017-12-08 14:13 ` Gao, Liming 6 siblings, 0 replies; 12+ messages in thread From: Gao, Liming @ 2017-12-08 14:13 UTC (permalink / raw) To: Pete Batard, edk2-devel@lists.01.org Pete: Thanks for your update. I have no other comment on BaseTools and MdePkg part. Reviewed-by: Liming Gao <liming.gao@intel.com> > -----Original Message----- > From: Pete Batard [mailto:pete@akeo.ie] > Sent: Friday, December 8, 2017 10:06 PM > To: edk2-devel@lists.01.org > Cc: Gao, Liming <liming.gao@intel.com> > Subject: [PATCH v3 0/6] Add ARM support for VS2017 > > This v3 is the same as v2, except for the following comment added in > tools_def for VS2017: > # Building of XIP firmware images for ARM is not currently supported (only applications). > # /FILEALIGN:4096 and other changes are needed for ARM firmware builds. > > The following series adds ARM compilation support for the VS2017 toolchain. > * PATCH 1 targets the disabling of VS Level 4 warnings. The disabled warnings > for ARM are now aligned with IA32 and X64. > * PATCH 2 adds a NULL handler for the base stack check, since this is a GCC > functionality. > * PATCH 3 updates MdePkg/Library/BaseLib so that the RVCT assembly sources > are also used for MSFT. > * PATCH 4 adds the required compiler intrinsics replacements for division, > shift and memset/memcpy. > * PATCH 5 adds variable argument handlers for print output. Note that this > is done without relying on any external headers, with the VA_ARG macro > having been reverse engineered from MSFT ARM assembly output. > * PATCH 6 enables the selection of ARM in the conf templates. > > With these patches, VS2017 toolchain users should be able to compile > regular UEFI ARM applications using EDK2. Note that, unlike ARM64 support, > ARM support does not require a specific update of Visual Studio 2017, as > the ARM toolchain has been available from the very first release. > > Additional notes: > > We tested compiling and running the full UEFI Shell with this series, as > well as a small set of applications and drivers, and found no issues. > With an additional patch [1], it is also possible to use this proposal to > compile a complete QEMU ARM firmware. As the patch shows, the changes that > need to be applied to the EDK2 sources to achieve this are actually very > minimal. > > However, the generated firmware does not currently boot, possibly because > of the following warnings being generated by the MS compiler: > - ArmCpuDxe.dll : warning LNK4072: section count 118 exceeds max (96); image may not run > - UiApp.dll : warning LNK4072: section count 113 exceeds max (96); image may not run > > As far as I could see, the section count max is hardcoded so a workaround > would be needed to address those. > > Also, because the VS2017 ARM compiler forces a section alignment of 4096 > bytes (which in turn forces use to use /FILEALIGN:4096 as a linker option > for the firmware generation), the generated firmware exceeds 2MB and we > had to double its size to 4MB. > > At this stage, since the goal of this series is to allow users to compile > regular ARM UEFI applications using the VS2017 toolchain, I have non plans > to spend more time on the QEMU firmware issues, especially as I suspect > that reducing the firmware size back to 2 MB may not be achievable without > Microsoft altering their compiler. I am however hopeful that ARM > specialists can take this matter over eventually... > > Regards, > > /Pete > > [1] https://github.com/pbatard/edk2/commit/c4ce41094a46f4f3dc7ccc64a90604813f037b13 > > Pete Batard (6): > MdePkg: Disable some Level 4 warnings for VS2017/ARM > MdePkg/Library/BaseStackCheckLib: Add Null handler for VS2017/ARM > MdePkg/Library/BaseLib: Enable VS2017/ARM builds > ArmPkg/Library/CompilerIntrinsicsLib: Enable VS2017/ARM builds > MdePkg/Include: Add VA list support for VS2017/ARM > BaseTools/Conf: Add VS2017/ARM support > > ArmPkg/Library/CompilerIntrinsicsLib/Arm/rtdiv.asm | 255 ++++++++++++++++++++ > ArmPkg/Library/CompilerIntrinsicsLib/Arm/rtsrsh.asm | 45 ++++ > ArmPkg/Library/CompilerIntrinsicsLib/CompilerIntrinsicsLib.inf | 13 +- > ArmPkg/Library/CompilerIntrinsicsLib/memcpy_ms.c | 34 +++ > ArmPkg/Library/CompilerIntrinsicsLib/memset_ms.c | 33 +++ > BaseTools/Conf/build_rule.template | 31 ++- > BaseTools/Conf/tools_def.template | 31 +++ > MdePkg/Include/Arm/ProcessorBind.h | 96 ++++++-- > MdePkg/Include/Base.h | 13 + > MdePkg/Library/BaseLib/Arm/CpuBreakpoint.asm | 5 +- > MdePkg/Library/BaseLib/BaseLib.inf | 16 +- > MdePkg/Library/BaseStackCheckLib/BaseStackCheckLib.inf | 5 +- > MdePkg/Library/BaseStackCheckLib/BaseStackCheckNull.c | 18 ++ > 13 files changed, 565 insertions(+), 30 deletions(-) > create mode 100644 ArmPkg/Library/CompilerIntrinsicsLib/Arm/rtdiv.asm > create mode 100644 ArmPkg/Library/CompilerIntrinsicsLib/Arm/rtsrsh.asm > create mode 100644 ArmPkg/Library/CompilerIntrinsicsLib/memcpy_ms.c > create mode 100644 ArmPkg/Library/CompilerIntrinsicsLib/memset_ms.c > create mode 100644 MdePkg/Library/BaseStackCheckLib/BaseStackCheckNull.c > > -- > 2.9.3.windows.2 ^ permalink raw reply [flat|nested] 12+ messages in thread
end of thread, other threads:[~2017-12-14 14:00 UTC | newest] Thread overview: 12+ messages (download: mbox.gz follow: Atom feed -- links below jump to the message on this page -- 2017-12-08 14:06 [PATCH v3 0/6] Add ARM support for VS2017 Pete Batard 2017-12-08 14:06 ` [PATCH v3 1/6] MdePkg: Disable some Level 4 warnings for VS2017/ARM Pete Batard 2017-12-08 14:06 ` [PATCH v3 2/6] MdePkg/Library/BaseStackCheckLib: Add Null handler " Pete Batard 2017-12-08 14:06 ` [PATCH v3 3/6] MdePkg/Library/BaseLib: Enable VS2017/ARM builds Pete Batard 2017-12-08 14:06 ` [PATCH v3 4/6] ArmPkg/Library/CompilerIntrinsicsLib: " Pete Batard 2017-12-10 13:30 ` Gao, Liming 2017-12-12 19:59 ` Ard Biesheuvel 2017-12-13 13:56 ` Pete Batard 2017-12-14 14:05 ` Ard Biesheuvel 2017-12-08 14:06 ` [PATCH v3 5/6] MdePkg/Include: Add VA list support for VS2017/ARM Pete Batard 2017-12-08 14:06 ` [PATCH v3 6/6] BaseTools/Conf: Add VS2017/ARM support Pete Batard 2017-12-08 14:13 ` [PATCH v3 0/6] Add ARM support for VS2017 Gao, Liming
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox