Re: [PATCH 20/25] arm64:ilp32: add sys_ilp32.c and a separate table (in entry.S) to use it

[Yury Norov]

On Fri, May 06, 2016 at 08:16:48PM +0800, Zhangjian (Bamvor) wrote:
> Hi,
> 
> On 2016/4/6 6:08, Yury Norov wrote:
> >From: Andrew Pinski <apinski@xxxxxxxxxx>
> >
> >Add a separate syscall-table for ILP32, which dispatches either to native
> >LP64 system call implementation or to compat-syscalls, as appropriate.
> >
> >Signed-off-by: Andrew Pinski <Andrew.Pinski@xxxxxxxxxxxxxxxxxx>
> >Signed-off-by: Yury Norov <ynorov@xxxxxxxxxxxxxxxxxx>
> >---
> >  arch/arm64/include/asm/unistd.h | 11 ++++++-
> >  arch/arm64/kernel/Makefile      |  2 +-
> >  arch/arm64/kernel/entry.S       | 12 +++++++-
> >  arch/arm64/kernel/sys_ilp32.c   | 65 +++++++++++++++++++++++++++++++++++++++++
> >  4 files changed, 87 insertions(+), 3 deletions(-)
> >  create mode 100644 arch/arm64/kernel/sys_ilp32.c
> >
> >diff --git a/arch/arm64/include/asm/unistd.h b/arch/arm64/include/asm/unistd.h
> >index 2971dea..5ea18ef 100644
> >--- a/arch/arm64/include/asm/unistd.h
> >+++ b/arch/arm64/include/asm/unistd.h
> >@@ -13,9 +13,18 @@
> >   * You should have received a copy of the GNU General Public License
> >   * along with this program.  If not, see <http://www.gnu.org/licenses/>.
> >   */
> >+
> >+#ifdef CONFIG_COMPAT
> >+#define __ARCH_WANT_COMPAT_STAT64
> >+#endif
> >+
> >+#ifdef CONFIG_ARM64_ILP32
> >+#define __ARCH_WANT_COMPAT_SYS_PREADV64
> >+#define __ARCH_WANT_COMPAT_SYS_PWRITEV64
> >+#endif
> >+
> >  #ifdef CONFIG_AARCH32_EL0
> >  #define __ARCH_WANT_COMPAT_SYS_GETDENTS64
> >-#define __ARCH_WANT_COMPAT_STAT64
> >  #define __ARCH_WANT_SYS_GETHOSTNAME
> >  #define __ARCH_WANT_SYS_PAUSE
> >  #define __ARCH_WANT_SYS_GETPGRP
> >diff --git a/arch/arm64/kernel/Makefile b/arch/arm64/kernel/Makefile
> >index 9dfdf86..7aa65ea 100644
> >--- a/arch/arm64/kernel/Makefile
> >+++ b/arch/arm64/kernel/Makefile
> >@@ -28,7 +28,7 @@ $(obj)/%.stub.o: $(obj)/%.o FORCE
> >  arm64-obj-$(CONFIG_AARCH32_EL0)		+= sys32.o kuser32.o signal32.o 	\
> >  					   sys_compat.o entry32.o		\
> >  					   ../../arm/kernel/opcodes.o binfmt_elf32.o
> >-arm64-obj-$(CONFIG_ARM64_ILP32)		+= binfmt_ilp32.o
> >+arm64-obj-$(CONFIG_ARM64_ILP32)		+= binfmt_ilp32.o sys_ilp32.o
> >  arm64-obj-$(CONFIG_FUNCTION_TRACER)	+= ftrace.o entry-ftrace.o
> >  arm64-obj-$(CONFIG_MODULES)		+= arm64ksyms.o module.o
> >  arm64-obj-$(CONFIG_ARM64_MODULE_PLTS)	+= module-plts.o
> >diff --git a/arch/arm64/kernel/entry.S b/arch/arm64/kernel/entry.S
> >index cf4d1ae..1f7a145 100644
> >--- a/arch/arm64/kernel/entry.S
> >+++ b/arch/arm64/kernel/entry.S
> >@@ -715,9 +715,13 @@ ENDPROC(ret_from_fork)
> >   */
> >  	.align	6
> >  el0_svc:
> >-	adrp	stbl, sys_call_table		// load syscall table pointer
> >  	uxtw	scno, w8			// syscall number in w8
> >  	mov	sc_nr, #__NR_syscalls
> >+#ifdef CONFIG_ARM64_ILP32
> >+	ldr	x16, [tsk, #TI_FLAGS]
> >+	tbnz	x16, #TIF_32BIT_AARCH64, el0_ilp32_svc // We are using ILP32
> >+#endif
> >+	adrp	stbl, sys_call_table		// load syscall table pointer
> >  el0_svc_naked:					// compat entry point
> >  	stp	x0, scno, [sp, #S_ORIG_X0]	// save the original x0 and syscall number
> >  	enable_dbg_and_irq
> >@@ -737,6 +741,12 @@ ni_sys:
> >  	b	ret_fast_syscall
> >  ENDPROC(el0_svc)
> >
> >+#ifdef CONFIG_ARM64_ILP32
> >+el0_ilp32_svc:
> >+	adrp	stbl, sys_call_ilp32_table // load syscall table pointer
> >+	b el0_svc_naked
> >+#endif
> >+
> >  	/*
> >  	 * This is the really slow path.  We're going to be doing context
> >  	 * switches, and waiting for our parent to respond.
> >diff --git a/arch/arm64/kernel/sys_ilp32.c b/arch/arm64/kernel/sys_ilp32.c
> >new file mode 100644
> >index 0000000..0996d8e
> >--- /dev/null
> >+++ b/arch/arm64/kernel/sys_ilp32.c
> >@@ -0,0 +1,65 @@
> >+/*
> >+ * AArch64- ILP32 specific system calls implementation
> >+ *
> >+ * Copyright (C) 2016 Cavium Inc.
> >+ * Author: Andrew Pinski <apinski@xxxxxxxxxx>
> >+ *
> >+ * This program is free software; you can redistribute it and/or modify
> >+ * it under the terms of the GNU General Public License version 2 as
> >+ * published by the Free Software Foundation.
> >+ *
> >+ * This program is distributed in the hope that it will be useful,
> >+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
> >+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
> >+ * GNU General Public License for more details.
> >+ *
> >+ * You should have received a copy of the GNU General Public License
> >+ * along with this program.  If not, see <http://www.gnu.org/licenses/>.
> >+ */
> >+
> >+#include <linux/compiler.h>
> >+#include <linux/errno.h>
> >+#include <linux/fs.h>
> >+#include <linux/mm.h>
> >+#include <linux/msg.h>
> >+#include <linux/export.h>
> >+#include <linux/sched.h>
> >+#include <linux/slab.h>
> >+#include <linux/syscalls.h>
> >+#include <linux/compat.h>
> >+#include <asm-generic/syscalls.h>
> >+
> >+/* Using non-compat syscalls where necessary */
> >+#define compat_sys_fadvise64_64        sys_fadvise64_64
> >+#define compat_sys_fallocate           sys_fallocate
> >+#define compat_sys_ftruncate64         sys_ftruncate
> >+#define compat_sys_lookup_dcookie      sys_lookup_dcookie
> >+#define compat_sys_pread64             sys_pread64
> >+#define compat_sys_pwrite64            sys_pwrite64
> >+#define compat_sys_readahead           sys_readahead
> >+#define compat_sys_shmat               sys_shmat
> >+#define compat_sys_sync_file_range     sys_sync_file_range
> >+#define compat_sys_truncate64          sys_truncate
> >+#define sys_llseek                     sys_lseek
> >+#define sys_mmap2		       sys_mmap
> I am a little bit confused here. We wrap the mmap to mmap2 in glibc
> without shift the 4096 and We map mmap2 to mmap in kernel which
> means we shift with the real page size. It works unless the
> application want to mmap the offset bigger then 2G. In ILP32 app,
> if the offset is bigger than 2G(e.g. 0xfb000000), it is a negative
> number and extend to 64bit nagetive number in kernel
> (0xfffffff fb000000). I add the "COMPAT_SYSCALL_WRAP6(mmap, ...)" in
> kernel/compat_wrapper.c. But it is not works. I am not sure if it is
> already sign extended in userspace.
> 
> On the other hand, I read the code of mmap in arm and other
> architecture. Usually, they will shift 4096 in userspace and shift
> others in kernel if needed. Should we follow the similar ways or we
> could call mmap_pgoff in glibc and do the shift according the real
> page shift(getpages())?
> 
> Thanks
> 
> Bamvor
> 
> 

Hi,

AFAIR, here we don't shift offset, as it's 64-bit both in user-
and kernel-space, and just pass it from user to kernel thru glibc
with no changes.

It definitely works, as there are many mappings made by linker and
libc in 2G+ area, and there are no problems with them. This is a
typical ILP32 application map:

00400000-00401000 r-xp 00000000 08:00 130400     /root/mykill
00410000-00411000 rwxp 00000000 08:00 130400     /root/mykill
00527000-00549000 rwxp 00000000 00:00 0          [heap]
c6278000-c6298000 rwxp 00000000 00:00 0 
c6298000-c63d0000 r-xp 00000000 08:00 135293     /root/sys-root/libilp32/libc-2.22.so
c63d0000-c63e0000 ---p 00138000 08:00 135293     /root/sys-root/libilp32/libc-2.22.so
c63e0000-c63e2000 r-xp 00138000 08:00 135293     /root/sys-root/libilp32/libc-2.22.so
c63e2000-c63e3000 rwxp 0013a000 08:00 135293     /root/sys-root/libilp32/libc-2.22.so
c63e3000-c63e6000 rwxp 00000000 00:00 0 
c63e6000-c63fc000 r-xp 00000000 08:00 135313     /root/sys-root/libilp32/libpthread-2.22.so
c63fc000-c640b000 ---p 00016000 08:00 135313     /root/sys-root/libilp32/libpthread-2.22.so
c640b000-c640c000 r-xp 00015000 08:00 135313     /root/sys-root/libilp32/libpthread-2.22.so
c640c000-c640d000 rwxp 00016000 08:00 135313     /root/sys-root/libilp32/libpthread-2.22.so
c640d000-c640f000 rwxp 00000000 00:00 0 
c640f000-c642c000 r-xp 00000000 08:00 135288     /root/sys-root/libilp32/ld-2.22.so
c6437000-c6439000 rwxp 00000000 00:00 0 
c6439000-c643a000 r--p 00000000 00:00 0          [vvar]
c643a000-c643b000 r-xp 00000000 00:00 0          [vdso]
c643b000-c643c000 r-xp 0001c000 08:00 135288     /root/sys-root/libilp32/ld-2.22.so
c643c000-c643d000 rwxp 0001d000 08:00 135288     /root/sys-root/libilp32/ld-2.22.so
ffe2d000-ffe4e000 rw-p 00000000 00:00 0          [stack]


> _______________________________________________
> linux-arm-kernel mailing list
> linux-arm-kernel@xxxxxxxxxxxxxxxxxxx
> http://lists.infradead.org/mailman/listinfo/linux-arm-kernel