Re: [PATCH v1] xen/riscv: make calculation of stack address PC-relative

[Oleksii <oleksii.kurochko@xxxxxxxxx>]

On Wed, 2023-03-15 at 08:59 +0100, Jan Beulich wrote:
> On 14.03.2023 21:16, Oleksii wrote:
> > I checked in Linux binary how 'la' instruction is transformed, and
> > it
> > looks like it is translated as I expect to auipc/addi pair:
> > ffffffe000001066: 00027517 auipc a0,0x27
> > ffffffe00000106a: f9a50513 addi a0,a0,-102 # ffffffe000028000
> > <early_pg_dir>
> > 
> > I checked compiler flags between Xen and Linux. The difference is
> > in-
> > fno-PIE (Linux also adds -mabi and -march to AFLAGS):
> > 
> > 1. Linux build command of head.S: riscv64-linux-gnu-gcc -Wp,-
> > MD,arch/riscv/kernel/.head.o.d -nostdinc -isystem /usr/lib/gcc-
> > cross/riscv64-linux-gnu/9/include -I./arch/riscv/include -
> > I./arch/riscv/include/generated -I./include -
> > I./arch/riscv/include/uapi
> > -I./arch/riscv/include/generated/uapi -I./include/uapi -
> > I./include/generated/uapi -include ./include/linux/kconfig.h -
> > D__KERNEL__ -D__ASSEMBLY__ -fno-PIE -mabi=lp64 -march=rv64imafdc -c
> > -o
> > arch/riscv/kernel/head.o arch/riscv/kernel/head.S
> > 
> > 2. Xen build command of head.S:riscv64-linux-gnu-gcc -MMD -MP -MF
> > arch/riscv/riscv64/.head.o.d -D__ASSEMBLY__ -Wa,--noexecstack -
> > DBUILD_ID -fno-strict-aliasing -Wall -Wstrict-prototypes -
> > Wdeclaration-
> > after-statement -Wno-unused-but-set-variable -Wno-unused-local-
> > typedefs
> > -O1 -fno-omit-frame-pointer -nostdinc -fno-builtin -fno-common -
> > Werror
> > -Wredundant-decls -Wno-pointer-arith -Wvla -pipe -D__XEN__ -include
> > ./include/xen/config.h -Wa,--strip-local-absolute -g -mabi=lp64 -
> > I./include -I./arch/riscv/include -march=rv64gc -mstrict-align -
> > mcmodel=medany - -c arch/riscv/riscv64/head.S -o
> > arch/riscv/riscv64/head.o
> 
> Looking into why you see different code generated than I: Nothing in
> here directs gcc to pass -fpic to gas; in upstream gcc (consistent
> from gcc7 through gcc12, which are the versions I've checked; the
> actual range may be wider) there is
> 
> #define ASM_SPEC "\
> %(subtarget_asm_debugging_spec) \
> %{" FPIE_OR_FPIC_SPEC ":-fpic} \
> ...
> 
> Can you check whether your gcc passes -fpic to gas even when there's
> no -fPIC / -fPIE (or alike) on the gcc command line?
I am not sure that I know how to check specifically if -fpic flag
passes to gas.
Could you please tell me?

>  Or whether your
> gas (unlike upstream's) defaults to PIC mode? (For .S files ASM_SPEC
> is all that counts. For .c files gcc is redundantly passing -fpic
> along with also emitting ".option pic" or, in the opposite case, it
> is omitting -fpic along with emitting ".option nopic".)
it looks like it should be by default -fpic because if look at gcc spec
file for the RISC-V architecture:

[user@49295ae49cbe build]$ riscv64-linux-gnu-gcc -dumpspecs | grep -i
pic
--traditional-format %(subtarget_asm_debugging_spec) %{fno-pie|fno-
PIE|fno-pic|fno-PIC:;:-fpic} %{march=*} %{mabi=*} %{mno-relax} %{mbig-
endian} %{mlittle-endian} %(subtarget_asm_spec)%{misa-spec=*}

which means that -fpic is enabled if none of the following options are
present on the command line:
    -fno-pie
    -fno-PIE
    -fno-pic
    -fno-PIC

> 
> You gcc may have been configured with --enable-default-pie, while I
> know mine hasn't been (simply because that's the default).
You are right my gcc is configured with --enable-default-pie:

[user@49295ae49cbe build]$ riscv64-linux-gnu-gcc -v 
Using built-in specs.
COLLECT_GCC=riscv64-linux-gnu-gcc
COLLECT_LTO_WRAPPER=/usr/lib/gcc/riscv64-linux-gnu/12.2.0/lto-wrapper
Target: riscv64-linux-gnu
Configured with: /build/riscv64-linux-gnu-gcc/src/gcc-12.2.0/configure
--prefix=/usr --program-prefix=riscv64-linux-gnu- --with-local-
prefix=/usr/riscv64-linux-gnu --with-sysroot=/usr/riscv64-linux-gnu --
with-build-sysroot=/usr/riscv64-linux-gnu --libdir=/usr/lib --
libexecdir=/usr/lib --target=riscv64-linux-gnu --host=x86_64-pc-linux-
gnu --build=x86_64-pc-linux-gnu --with-system-zlib --with-isl --with-
linker-hash-style=gnu --disable-nls --disable-libunwind-exceptions --
disable-libstdcxx-pch --disable-libssp --disable-multilib --disable-
werror --enable-languages=c,c++ --enable-shared --enable-threads=posix
--enable-__cxa_atexit --enable-clocale=gnu --enable-gnu-unique-object -
-enable-linker-build-id --enable-lto --enable-plugin --enable-install-
libiberty --enable-gnu-indirect-function --enable-default-pie --enable-
checking=release
Thread model: posix
Supported LTO compression algorithms: zlib zstd
gcc version 12.2.0 (GCC)

So should we pass to CFLAGS and AFLAGS at least for RISC-V -fno-PIE?

~ Oleksii