[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Xen-devel] [RFC 5/7] WIP:arm64:armds: Build XEN with ARM Compiler 6.6
- To: Stefano Stabellini <sstabellini@xxxxxxxxxx>
- From: Andrii Anisov <andrii.anisov@xxxxxxxxx>
- Date: Thu, 14 Nov 2019 13:14:35 +0200
- Cc: Julien Grall <julien@xxxxxxx>, Andrii Anisov <andrii_anisov@xxxxxxxx>, Wei Liu <wl@xxxxxxx>, Konrad Rzeszutek Wilk <konrad.wilk@xxxxxxxxxx>, George Dunlap <George.Dunlap@xxxxxxxxxxxxx>, Andrew Cooper <andrew.cooper3@xxxxxxxxxx>, Ian Jackson <ian.jackson@xxxxxxxxxxxxx>, Jan Beulich <jbeulich@xxxxxxxx>, xen-devel@xxxxxxxxxxxxxxxxxxxx, Volodymyr Babchuk <Volodymyr_Babchuk@xxxxxxxx>
- Delivery-date: Thu, 14 Nov 2019 11:14:54 +0000
- List-id: Xen developer discussion <xen-devel.lists.xenproject.org>
On 11.11.19 23:26, Stefano Stabellini wrote:
Why _srodata and __start_bug_frames cannot both be defined as
Load$$_rodata_bug_frames_0$$Base when actually in the case of:
+#define __per_cpu_data_end Load$$_bss_percpu$$Limit
+#define __bss_end Load$$_bss_percpu$$Limit
+#define _end Load$$_bss_percpu$$Limit
They are all defined as "Load$$_bss_percpu$$Limit"? And:
+#define __init_end Load$$_bss$$Base
+#define __bss_start Load$$_bss$$Base
They are both defined as "Load$$_bss$$Base"? What's special about
"Load$$_rodata_bug_frames_0$$Base"?
Because in xen/include/asm-arm/bug.h:
extern const struct bug_frame __start_bug_frames[]
and in xen/include/xen/kernel.h
extern const char _srodata[];
After preprocessing we, effectively, have symbol declared with conflicting
types:
extern const struct bug_frame Load$$_rodata_bug_frames_0$$Base[];
extern const char Load$$_rodata_bug_frames_0$$Base[];
Eventually other symbols do not have such conflicts.
- C style shift operators are missed among supported scatter file expressions,
so some needed values are hardcoded in scatter file.
- Rename correspondent ARM Linker defined symbols to those needed by
`symbols` tool
using steering file feature.
- ARM Compiler 6.6 tools are not able to rename sections, so we still need
GNU toolchain's objcopy for this.
Signed-off-by: Andrii Anisov <andrii_anisov@xxxxxxxx>
---
xen/Rules.mk | 6 +
xen/arch/arm/Makefile | 24 ++++
xen/arch/arm/xen.scat.S | 266 ++++++++++++++++++++++++++++++++++++++++++++
I would strongly suggest to rename this file with something not
potentially related to scat
I just followed examples from ARM documentation, e.g. [1]. I suppose they know
something about their product.
Yet, the suggestion is reasonable.
xen/arch/arm/xen.steer | 5 +
xen/include/asm-arm/armds.h | 91 +++++++++++++++
5 files changed, 392 insertions(+)
create mode 100644 xen/arch/arm/xen.scat.S
create mode 100644 xen/arch/arm/xen.steer
create mode 100644 xen/include/asm-arm/armds.h
diff --git a/xen/Rules.mk b/xen/Rules.mk
index 41a1c26..67bedce 100644
--- a/xen/Rules.mk
+++ b/xen/Rules.mk
@@ -60,6 +60,12 @@ CFLAGS += -nostdinc -fno-builtin -fno-common
CFLAGS += -Werror -Wredundant-decls -Wno-pointer-arith
$(call cc-option-add,CFLAGS,CC,-Wvla)
CFLAGS += -pipe -D__XEN__ -include $(BASEDIR)/include/xen/config.h
+
+ifeq ($(armds),y)
+CFLAGS += -nostdlibinc -nostdlib -Wno-unused-command-line-argument
+CFLAGS += -include $(BASEDIR)/include/asm/armds.h
+endif
+
CFLAGS-$(CONFIG_DEBUG_INFO) += -g
CFLAGS += '-D__OBJECT_FILE__="$@"'
diff --git a/xen/arch/arm/Makefile b/xen/arch/arm/Makefile
index 70f532e..a5a3479 100644
--- a/xen/arch/arm/Makefile
+++ b/xen/arch/arm/Makefile
@@ -83,11 +83,16 @@ else
all_symbols =
endif
+ifeq ($(armds),y)
+$(TARGET): $(TARGET)-syms
+ fromelf --bin $< --output $@
+else
$(TARGET): $(TARGET)-syms
$(OBJCOPY) -O binary -S $< $@
ifeq ($(CONFIG_ARM_64),y)
ln -sf $(notdir $@) ../../$(notdir $@).efi
endif
+endif
ifeq ($(CONFIG_LTO),y)
# Gather all LTO objects together
@@ -102,6 +107,19 @@ prelink.o: $(ALL_OBJS)
$(LD) $(LDFLAGS) -r -o $@ $^
endif
+ifeq ($(armds),y)
+$(TARGET)-syms: prelink.o xen.scat
+ armlink --scatter="xen.scat" --edit="xen.steer" --no_scanlib $(LDFLAGS)
prelink.o $(BASEDIR)/common/symbols-dummy.o -o $(@D)/.$(@F).0
+ $(NM) -pa --format=sysv $(@D)/.$(@F).0 \
+ | $(BASEDIR)/tools/symbols $(all_symbols) --sysv --sort
>$(@D)/.$(@F).0.S
+ $(MAKE) -f $(BASEDIR)/Rules.mk $(@D)/.$(@F).0.o
+ armlink --scatter="xen.scat" --edit="xen.steer" --no_scanlib $(LDFLAGS)
prelink.o $(@D)/.$(@F).0.o -o $(@D)/.$(@F).1
+ $(NM) -pa --format=sysv $(@D)/.$(@F).1 \
+ | $(BASEDIR)/tools/symbols $(all_symbols) --sysv --sort
>$(@D)/.$(@F).1.S
+ $(MAKE) -f $(BASEDIR)/Rules.mk $(@D)/.$(@F).1.o
+ armlink --scatter="xen.scat" --edit="xen.steer" --no_scanlib
--symdefs="$(@D)/$(@F).map" $(LDFLAGS) prelink.o $(build_id_linker) $(@D)/.$(@F).1.o -o $@
+ rm -f $(@D)/.$(@F).[0-9]*
+else
$(TARGET)-syms: prelink.o xen.lds
$(LD) $(LDFLAGS) -T xen.lds -N prelink.o \
$(BASEDIR)/common/symbols-dummy.o -o $(@D)/.$(@F).0
@@ -119,14 +137,20 @@ $(TARGET)-syms: prelink.o xen.lds
| $(BASEDIR)/tools/symbols --xensyms --sysv --sort \
>$(@D)/$(@F).map
rm -f $(@D)/.$(@F).[0-9]*
+endif
asm-offsets.s: $(TARGET_SUBARCH)/asm-offsets.c
$(CC) $(filter-out -flto,$(CFLAGS)) -S -o $@ $<
+ifeq ($(armds),y)
+xen.scat: xen.scat.S
+ $(CC) -P -E --target=aarch64-arm-none-eabi -o $@ $<
+else
xen.lds: xen.lds.S
$(CC) -P -E -Ui386 $(AFLAGS) -o $@ $<
sed -e 's/xen\.lds\.o:/xen\.lds:/g' <.xen.lds.d >.xen.lds.d.new
mv -f .xen.lds.d.new .xen.lds.d
+endif
dtb.o: $(CONFIG_DTB_FILE)
diff --git a/xen/arch/arm/xen.scat.S b/xen/arch/arm/xen.scat.S
new file mode 100644
index 0000000..3bb405f
--- /dev/null
+++ b/xen/arch/arm/xen.scat.S
@@ -0,0 +1,266 @@
+#if 0
#if 0 must be a mistake?
It is here because of RFC. Because I do not like the hardcodes below, can't
find the better way, but eager for suggestions.
Also, is this basically the ARMCC version of a linker script? Is
xen.lds.S still used with ARMCC, or only xen.scat.S is used?
It is the ARMCC version of a linker script. xen.lds.S is not used.
+/*
+ * armlink does not understand shifts in scat file expressions
+ * so hardcode needed values
+ */
+#include <xen/cache.h>
+#include <asm/page.h>
+#include <asm/percpu.h>
+#undef ENTRY
+#undef ALIGN
+#else
+ #define PAGE_SIZE 4096
+ #define POINTER_ALIGN 8
+ #define SMP_CACHE_BYTES 128
+ #define STACK_SIZE 32768
+ #define XEN_VIRT_START 0x00200000
+#endif
+
+LOAD XEN_VIRT_START
+{
+;_start
+;_stext
+ _text AlignExpr(+0, PAGE_SIZE) PADVALUE 0x9090
+ {
+ *(.text*)
+ *(.text.cold)
+ *(.text.unlikely)
+ *(.fixup)
+ *(.gnu.warning)
+ }
+;_etext
+
+;_srodata
+;__start_bug_frames
+ _rodata_bug_frames_0 AlignExpr(+0, PAGE_SIZE) FIXED ZEROPAD
+ {
+ *(.bug_frames.0)
+ }
+;__stop_bug_frames_0
+
+ _rodata_bug_frames_1 +0 FIXED ZEROPAD
+ {
+ *(.bug_frames.1)
+ }
+;__stop_bug_frames_1
+
+ _rodata_bug_frames_2 +0 FIXED ZEROPAD
+ {
+ *(.bug_frames.2)
+ }
+;__stop_bug_frames_2
+
+ _rodata_data +0
+ {
+ *(.rodata)
+ *(.rodata.*)
+ *(.data.rel.ro)
+ *(.data.rel.ro.*)
+ }
+
+#ifdef CONFIG_LOCK_PROFILE
+;__lock_profile_start
+ _rodata_lockprofile_data AlignExpr(+0, POINTER_ALIGN) FIXED ZEROPAD
+ {
+ *(.lockprofile.data)
+ }
+;__lock_profile_end
+#endif
Should be below
OK.
+;__param_start
+ _rodata_data_param AlignExpr(+0, POINTER_ALIGN) FIXED ZEROPAD
+ {
+ *(.data.param)
+ }
+;__param_end
+
+;__proc_info_start
+ _rodata_proc_info +0 FIXED ZEROPAD
+ {
+ *(.proc.info)
+ }
+;__proc_info_end
+
+#if defined(CONFIG_HAS_VPCI) && defined(CONFIG_LATE_HWDOM)
+;__start_vpci_array
+ _rodata_data_vpci AlignExpr(+0, POINTER_ALIGN) SORTTYPE Lexical FIXED ZEROPAD
+ {
+ *(SORT(.data.vpci.*))
+ }
+;__end_vpci_array
+#endif
+
+#if defined(BUILD_ID)
+;__note_gnu_build_id_start
+ _note_gnu_build_id +0 FIXED ZEROPAD
+ {
+ *(.note.gnu.build-id)
+ }
+;__note_gnu_build_id_end
+#endif
+
+;_erodata
+
+ _data AlignExpr(+0, PAGE_SIZE) FIXED ZEROPAD
+ {
+ *(.data.page_aligned.*)
+ *(.data.*)
+ }
+
+;__start_schedulers_array
+ _data_schedulers AlignExpr(+0, 8) FIXED ZEROPAD
+ {
+ *(.data.schedulers)
+ }
+;__end_schedulers_array
+
+ _data_rel +0 FIXED ZEROPAD
+ {
+ *(.data.rel)
+ *(.data.rel.*)
+;#CONSTRUCTORS ????
Honestly I am not sure what this is either
Ah, it was a hint for myself to find place for constructors.
+ }
+
+;__start___ex_table
+ _data_ex_table AlignExpr(+0, SMP_CACHE_BYTES) FIXED ZEROPAD
+ {
+ *(.ex_table)
+ }
+;__stop___ex_table
+
+;__start___pre_ex_table
+ _data_ex_table_pre +0 FIXED ZEROPAD
+ {
+ *(.ex_table.pre)
+ }
+;__stop___pre_ex_table
+
+ _data_read_mostly +0 FIXED ZEROPAD
+ {
+ *(.data.read_mostly)
+ }
+
+;_splatform
+ _arch_info AlignExpr(+0, 8) FIXED ZEROPAD
+ {
+ *(.arch.info)
+ }
+;_eplatform
+
+;_sdevice
+ _dev_info AlignExpr(+0, 8) FIXED ZEROPAD
+ {
+ *(.dev.info)
+ }
+;_edevice
+
+;_asdevice
+ _adev_info AlignExpr(+0, 8) FIXED ZEROPAD
+ {
+ *(.adev.info)
+ }
+;_aedevice
_steemediator/_eteemediator
I suppose I did this .scat before _steemediator/_eteemediator were introduced
to .lds.
OK.
+;__init_begin
+;_sinittext
+ _init_text AlignExpr(+0, PAGE_SIZE) FIXED ZEROPAD
+ {
+ *(.init.text)
+ }
+;_einittext
+
+ _init_rodata AlignExpr(+0, PAGE_SIZE) FIXED ZEROPAD
+ {
+ *(.init.rodata)
+ *(.init.rodata.rel)
+ *(.init.rodata.str*)
+ }
+
+;__setup_start
+ _init_setup AlignExpr(+0, POINTER_ALIGN) FIXED ZEROPAD
+ {
+ *(.init.setup)
+ }
+;__setup_end
+
+;__initcall_start
+ _initcallpresmp_init +0 FIXED ZEROPAD
+ {
+ *(.initcallpresmp.init)
+ }
+;__presmp_initcall_end
+
+ _initcall1_init +0 FIXED ZEROPAD
+ {
+ *(.initcall1.init)
+ }
+;__initcall_end
+
+;__alt_instructions
+ _altinstructions AlignExpr(+0, 4) FIXED ZEROPAD
+ {
+ *(.altinstructions)
+ }
+;__alt_instructions_end
+
+ _altinstr_replacement AlignExpr(+0, 4) FIXED ZEROPAD
+ {
+ *(.altinstr_replacement)
+ }
__lock_profile_start should be here
OK.
+
+ _init_data +0 FIXED ZEROPAD
+ {
+ *(.init.data)
+ *(.init.data.rel)
+ *(.init.data.rel.*)
+ }
+
+;__ctors_start
+ _ctors AlignExpr(+0, 8) FIXED ZEROPAD
+ {
+ *(.ctors)
+ *(.init_array)
+ }
+
+ _init_array_sorted AlignExpr(+0, 8) SORTTYPE Lexical FIXED ZEROPAD
+ {
+ *(.init_array.*)
+ }
+;__ctors_end
+
+#if defined(CONFIG_HAS_VPCI) && !defined(CONFIG_LATE_HWDOM)
+ _data_vpci AlignExpr(+0, POINTER_ALIGN) SORTTYPE Lexical FIXED ZEROPAD
+ {
+ *(.data.vpci.*)
+ }
+#endif
__start_vpci_array/__end_vpci_array?
OK.
+;__init_end_efi
+
+;__init_end
+;__bss_start
+ _bss AlignExpr(+0, STACK_SIZE) FIXED ZEROPAD
+ {
+ *(.bss.stack_aligned*)
+ *(.bss.page_aligned*, OVERALIGN PAGE_SIZE)
+ *(.bss*)
+ }
+
+;__per_cpu_start
this should be page aligned too?
+ _bss_percpu AlignExpr(+0, SMP_CACHE_BYTES) FIXED ZEROPAD
+ {
+ *(.bss.percpu)
+ *(.bss.percpu.read_mostly, OVERALIGN SMP_CACHE_BYTES)
+ }
+;__per_cpu_data_end
+;__bss_end
__bss_end should be page aligned?
Well... Maybe, but I'm not sure how to do that.
+;_end
+
+#ifdef CONFIG_DTB_FILE
+;_sdtb
+ _dtb FIXED ZEROPAD
+ {
+ *(.dtb)
+ }
+#endif
+
+}
diff --git a/xen/arch/arm/xen.steer b/xen/arch/arm/xen.steer
new file mode 100644
index 0000000..646e912
--- /dev/null
+++ b/xen/arch/arm/xen.steer
@@ -0,0 +1,5 @@
+RESOLVE _srodata AS Load$$_rodata_bug_frames_0$$Base
+RENAME Load$$_text$$Base AS _stext
+RENAME Load$$_text$$Limit AS _etext
+RENAME Load$$_init_text$$Base AS _sinittext
+RENAME Load$$_init_text$$Limit AS _einittext
I don't get why some if the "symbols" get renamed using RENAME here, and
some other we are using a #define in xen/include/asm-arm/armds.h. Can't
we rename them all here?
Well, the situation here is really complicated. I described above a reason why
_srodata is resolved here.
Other symbols are renamed here because the tool "xen/tools/symbols", used at
the latest linking stages, needs `_stext`, `_etext`, and the rest to be present in the
elf.
diff --git a/xen/include/asm-arm/armds.h b/xen/include/asm-arm/armds.h
new file mode 100644
index 0000000..5ee2e5d
--- /dev/null
+++ b/xen/include/asm-arm/armds.h
Missing guards. Also, probably you want to make sure this is only #ifdef
ARMCC.
OK.
Is this meant to be used when building C files, asm files, or both?
Well, I have to check.
I would avoid this header file if we can get away with just xen.steer.
We can't go with xen.steer only. One of the armlink issues is "ARM linker defined
symbols are not counted as referred if only mentioned in a steering file for rename or
resolve". Also, linker-defined symbols are only generated when the code references
them [2].
I tried resolving existing symbols (e.g. _start) to armlink defined symbols
with .steer only, and got errors that armlink can't find those linker-defined
symbols.
I tried a specific C file with references to all needed linker-defined symbols,
then resolving all .lds-style symbols to armlink defined symbols with the
steering file. But it did not work, I don't remember exactly the issue. Maybe C
file with externs only (without using them in the effective code) did not
result in an object file referring those linker-defined symbols.
@@ -0,0 +1,91 @@
+#define _start Load$$_text$$Base
+#define _stext Load$$_text$$Base
+
+#define _etext Load$$_text$$Limit
+
+//#define _srodata Load$$_rodata_bug_frames_0$$Base
+#define __start_bug_frames Load$$_rodata_bug_frames_0$$Base
+
+#define __stop_bug_frames_0 Load$$_rodata_bug_frames_0$$Limit
+#define __stop_bug_frames_1 Load$$_rodata_bug_frames_1$$Limit
+#define __stop_bug_frames_2 Load$$_rodata_bug_frames_2$$Limit
+
+#ifdef CONFIG_LOCK_PROFILE
+#define __lock_profile_start Load$$_rodata_lockprofile_data$$Base
+#define __lock_profile_end Load$$_rodata_lockprofile_data$$Limit
+#endif
+
+#define __param_start Load$$_rodata_data_param$$Base
+#define __param_end Load$$_rodata_data_param$$Limit
+
+#define __proc_info_start Load$$_rodata_proc_info$$Base
+#define __proc_info_end Load$$_rodata_proc_info$$Limit
+
+#define _erodata Load$$_rodata_proc_info$$Limit
+
+#if defined(CONFIG_HAS_VPCI) && defined(CONFIG_LATE_HWDOM)
+#define __start_vpci_array Load$$_rodata_data_vpci$$Base
+#define __end_vpci_array Load$$_rodata_data_vpci$$Limit
+
+#undef _erodata
+#define _erodata Load$$_rodata_data_vpci$$Limit
+#endif
+
+#if defined(BUILD_ID)
+#define __note_gnu_build_id_start Load$$_note_gnu_build_id$$Base
+#define __note_gnu_build_id_end Load$$_note_gnu_build_id$$Limit
+
+#undef _erodata
+#define _erodata Load$$_note_gnu_build_id$$Limit
+#endif
+
+#define __start_schedulers_array Load$$_data_schedulers$$Base
+#define __end_schedulers_array Load$$_data_schedulers$$Limit
+
+/* Does not exist for ARM
+#define __start___ex_table Load$$_data_ex_table$$Base
+#define __stop___ex_table Load$$_data_ex_table$$Limit
+*/
+
+#define __start___pre_ex_table Load$$_data_ex_table_pre$$Base
+#define __stop___pre_ex_table Load$$_data_ex_table_pre$$Limit
+
+#define _splatform Load$$_arch_info$$Base
+#define _eplatform Load$$_arch_info$$Limit
+
+#define _sdevice Load$$_dev_info$$Base
+#define _edevice Load$$_dev_info$$Limit
+
+#define _asdevice Load$$_adev_info$$Base
+#define _aedevice Load$$_adev_info$$Limit
+
+#define __init_begin Load$$_init_text$$Base
+#define _sinittext Load$$_init_text$$Base
+#define _einittext Load$$_init_text$$Limit
+
+#define __setup_start Load$$_init_setup$$Base
+#define __setup_end Load$$_init_setup$$Limit
+
+#define __initcall_start Load$$_initcallpresmp_init$$Base
+#define __presmp_initcall_end Load$$_initcallpresmp_init$$Limit
+#define __initcall_end Load$$_initcall1_init$$Limit
+
+#define __alt_instructions Load$$_altinstructions$$Base
+#define __alt_instructions_end Load$$_altinstructions$$Limit
+
+#define __ctors_start Load$$_ctors$$Base
+#define __ctors_end Load$$_init_array_sorted$$Limit
+#define __init_end_efi Load$$_init_array_sorted$$Limit
+
+#if defined(CONFIG_HAS_VPCI) && !defined(CONFIG_LATE_HWDOM)
+#undef __init_end_efi
+#define __init_end_efi Load$$_data_vpci$$Limit
+#endif
+
+#define __init_end Load$$_bss$$Base
+#define __bss_start Load$$_bss$$Base
+
+#define __per_cpu_start Load$$_bss_percpu$$Base
+#define __per_cpu_data_end Load$$_bss_percpu$$Limit
+#define __bss_end Load$$_bss_percpu$$Limit
+#define _end Load$$_bss_percpu$$Limit
[1]
http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc.dui0803j/rbm1505486312921.html
[2]
http://infocenter.arm.com/help/topic/com.arm.doc.dui0803j/nbd1509536435303.html
--
Sincerely,
Andrii Anisov.
_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel
|
