Skip to contentSkip to conversation

Convert .flat files to 32-bit elf format

PATCH In review Avi Kivity 2 participants 4 messages Dec 5, 2010 KVM

lore
Share & tools
lore

Avi Kivity

author
#0 lore raw

PATCH 1/1 Convert .flat files to 32-bit elf format

Recent qemus no longer load 64-bit elf kernels. Fix by converting to 32-bit elf.

---
 config-x86-common.mak |   39 +++++++++++++++++++++------------------
 1 files changed, 21 insertions(+), 18 deletions(-)
config-x86-common.mak +21 -18
index b541c1c..c5508b3 100644
@@ -20,9 +20,12 @@ CFLAGS += -m$(bits)
 libgcc := $(shell $(CC) -m$(bits) --print-libgcc-file-name)
 
 FLATLIBS = lib/libcflat.a $(libgcc)
-%.flat: %.o $(FLATLIBS) flat.lds
+%.elf: %.o $(FLATLIBS) flat.lds
 	$(CC) $(CFLAGS) -nostdlib -o $@ -Wl,-T,flat.lds $(filter %.o, $^) $(FLATLIBS)
 
+%.flat: %.elf
+	objcopy -O elf32-i386 $^ $@
+
 tests-common = $(TEST_DIR)/vmexit.flat $(TEST_DIR)/tsc.flat \
                $(TEST_DIR)/smptest.flat  $(TEST_DIR)/port80.flat \
                $(TEST_DIR)/realmode.flat $(TEST_DIR)/msr.flat \
@@ -35,47 +38,47 @@ test_cases: $(tests-common) $(tests)
 
 $(TEST_DIR)/%.o: CFLAGS += -std=gnu99 -ffreestanding -I lib -I lib/x86
 
-$(TEST_DIR)/access.flat: $(cstart.o) $(TEST_DIR)/access.o
+$(TEST_DIR)/access.elf: $(cstart.o) $(TEST_DIR)/access.o
 
-$(TEST_DIR)/hypercall.flat: $(cstart.o) $(TEST_DIR)/hypercall.o
+$(TEST_DIR)/hypercall.elf: $(cstart.o) $(TEST_DIR)/hypercall.o
 
-$(TEST_DIR)/sieve.flat: $(cstart.o) $(TEST_DIR)/sieve.o \
+$(TEST_DIR)/sieve.elf: $(cstart.o) $(TEST_DIR)/sieve.o \
 		$(TEST_DIR)/vm.o
 
-$(TEST_DIR)/vmexit.flat: $(cstart.o) $(TEST_DIR)/vmexit.o
+$(TEST_DIR)/vmexit.elf: $(cstart.o) $(TEST_DIR)/vmexit.o
 
-$(TEST_DIR)/smptest.flat: $(cstart.o) $(TEST_DIR)/smptest.o
+$(TEST_DIR)/smptest.elf: $(cstart.o) $(TEST_DIR)/smptest.o
 
-$(TEST_DIR)/emulator.flat: $(cstart.o) $(TEST_DIR)/emulator.o \
+$(TEST_DIR)/emulator.elf: $(cstart.o) $(TEST_DIR)/emulator.o \
 			   $(TEST_DIR)/vm.o $(TEST_DIR)/idt.o
 
-$(TEST_DIR)/port80.flat: $(cstart.o) $(TEST_DIR)/port80.o
+$(TEST_DIR)/port80.elf: $(cstart.o) $(TEST_DIR)/port80.o
 
-$(TEST_DIR)/tsc.flat: $(cstart.o) $(TEST_DIR)/tsc.o
+$(TEST_DIR)/tsc.elf: $(cstart.o) $(TEST_DIR)/tsc.o
 
-$(TEST_DIR)/apic.flat: $(cstart.o) $(TEST_DIR)/apic.o $(TEST_DIR)/vm.o
+$(TEST_DIR)/apic.elf: $(cstart.o) $(TEST_DIR)/apic.o $(TEST_DIR)/vm.o
 
-$(TEST_DIR)/realmode.flat: $(TEST_DIR)/realmode.o
+$(TEST_DIR)/realmode.elf: $(TEST_DIR)/realmode.o
 	$(CC) -m32 -nostdlib -o $@ -Wl,-T,$(TEST_DIR)/realmode.lds $^
 
 $(TEST_DIR)/realmode.o: bits = 32
 
-$(TEST_DIR)/msr.flat: $(cstart.o) $(TEST_DIR)/msr.o
+$(TEST_DIR)/msr.elf: $(cstart.o) $(TEST_DIR)/msr.o
 
-$(TEST_DIR)/idt_test.flat: $(cstart.o) $(TEST_DIR)/idt.o $(TEST_DIR)/idt_test.o
+$(TEST_DIR)/idt_test.elf: $(cstart.o) $(TEST_DIR)/idt.o $(TEST_DIR)/idt_test.o
 
-$(TEST_DIR)/xsave.flat: $(cstart.o) $(TEST_DIR)/idt.o $(TEST_DIR)/xsave.o
+$(TEST_DIR)/xsave.elf: $(cstart.o) $(TEST_DIR)/idt.o $(TEST_DIR)/xsave.o
 
-$(TEST_DIR)/rmap_chain.flat: $(cstart.o) $(TEST_DIR)/rmap_chain.o \
+$(TEST_DIR)/rmap_chain.elf: $(cstart.o) $(TEST_DIR)/rmap_chain.o \
 			     $(TEST_DIR)/vm.o
 
-$(TEST_DIR)/svm.flat: $(cstart.o) $(TEST_DIR)/vm.o
+$(TEST_DIR)/svm.elf: $(cstart.o) $(TEST_DIR)/vm.o
 
-$(TEST_DIR)/kvmclock_test.flat: $(cstart.o) $(TEST_DIR)/kvmclock.o \
+$(TEST_DIR)/kvmclock_test.elf: $(cstart.o) $(TEST_DIR)/kvmclock.o \
                                 $(TEST_DIR)/kvmclock_test.o
 
 arch_clean:
-	$(RM) $(TEST_DIR)/*.o $(TEST_DIR)/*.flat \
+	$(RM) $(TEST_DIR)/*.o $(TEST_DIR)/*.flat $(TEST_DIR)/*.elf \
 	$(TEST_DIR)/.*.d $(TEST_DIR)/lib/.*.d $(TEST_DIR)/lib/*.o
 
 -include $(TEST_DIR)/.*.d lib/.*.d lib/x86/.*.d
Signature · 1 line
1.7.1
#1 lore raw
Quoted 13 lines · Avi Kivity

Recent qemus no longer load 64-bit elf kernels. Fix by converting to 32-bit elf.

---
config-x86-common.mak | 39 +++++++++++++++++++++------------------

1 files changed, 21 insertions(+), 18 deletions(-)
config-x86-common.mak +0 -0
index b541c1c..c5508b3 100644

config-x86-common.mak CFLAGS += -m$(bits) · line 27

Quoted 9 lines
config-x86-common.mak +3 -1
@@ -20,9 +20,12 @@ CFLAGS += -m$(bits)
 libgcc := $(shell $(CC) -m$(bits) --print-libgcc-file-name)
 
 FLATLIBS = lib/libcflat.a $(libgcc)
-%.flat: %.o $(FLATLIBS) flat.lds
+%.elf: %.o $(FLATLIBS) flat.lds
 	$(CC) $(CFLAGS) -nostdlib -o $@ -Wl,-T,flat.lds $(filter %.o, $^) $(FLATLIBS)
 
+%.flat: %.elf
+	objcopy -O elf32-i386 $^ $@

Not sure it's that great to call 32bit elf binaries ".flat". I'd rather expect a flat file when reading that file extension ;). Mind to just call it ".elf32"?

Alternatively, you could also use real flat files and shove a multiboot header into the binary.

Alex

author
#2 lore raw
Quoted 13 lines · Alexander Graf
Quoted 13 lines

diff --git a/config-x86-common.mak b/config-x86-common.mak index b541c1c..c5508b3 100644
--- a/config-x86-common.mak
+++ b/config-x86-common.mak
@@ -20,9 +20,12 @@ CFLAGS += -m$(bits)
libgcc := $(shell $(CC) -m$(bits) --print-libgcc-file-name)

FLATLIBS = lib/libcflat.a $(libgcc)
-%.flat: %.o $(FLATLIBS) flat.lds
+%.elf: %.o $(FLATLIBS) flat.lds
$(CC) $(CFLAGS) -nostdlib -o $@ -Wl,-T,flat.lds $(filter %.o, $^) $(FLATLIBS)

Quoted 4 lines
Quoted 2 lines

+%.flat: %.elf
+ objcopy -O elf32-i386 $^ $@

Not sure it's that great to call 32bit elf binaries ".flat". I'd rather expect a flat file when reading that file extension ;). Mind to just call it ".elf32"?

It's a reminder from the good old days that these were flat files loaded by kvmctl.

Quoted 1 line

Alternatively, you could also use real flat files and shove a multiboot header into the binary.

I'll rename them in a later patch. .elf32 isn't a good name either, it implies 32-bit execution, whereas it's just the file format due to multiboot constraints. Maybe .test (and .test32 and .test64 when we build both).

Signature · 1 line
error compiling committee.c: too many arguments to function
#3 lore raw
Quoted 14 lines · Avi Kivity
Quoted 13 lines · Alexander Graf
Quoted 13 lines

diff --git a/config-x86-common.mak b/config-x86-common.mak index b541c1c..c5508b3 100644
--- a/config-x86-common.mak
+++ b/config-x86-common.mak
@@ -20,9 +20,12 @@ CFLAGS += -m$(bits)
libgcc := $(shell $(CC) -m$(bits) --print-libgcc-file-name)

FLATLIBS = lib/libcflat.a $(libgcc)
-%.flat: %.o $(FLATLIBS) flat.lds
+%.elf: %.o $(FLATLIBS) flat.lds
$(CC) $(CFLAGS) -nostdlib -o $@ -Wl,-T,flat.lds $(filter %.o, $^) $(FLATLIBS)

Quoted 6 lines
Quoted 4 lines
Quoted 2 lines

+%.flat: %.elf
+ objcopy -O elf32-i386 $^ $@

Not sure it's that great to call 32bit elf binaries ".flat". I'd rather expect a flat file when reading that file extension ;). Mind to just call it ".elf32"?

It's a reminder from the good old days that these were flat files loaded by kvmctl.

Nothing's as permanent as an improvised solution :).

Quoted 3 lines
Quoted 1 line

Alternatively, you could also use real flat files and shove a multiboot header into the binary.

I'll rename them in a later patch. .elf32 isn't a good name either, it implies 32-bit execution, whereas it's just the file format due to multiboot constraints. Maybe .test (and .test32 and .test64 when we build both).

Yeah, anything really. The .flat is definitely just wrong ;). .test32 and .test64 always sounds good to me. No reason to be clever and also have .test.

Alex