Using LTP on Andriod - Current State, Latest Patch, and EMail address of folks involved with this effort.

LTP [email protected] 1 participant 2 messages Aug 3, 2012 – Aug 8, 2012 2 messages not in the archive

lore

Share & tools

lore

A message that is not in the archive (sent off-list or not yet received): [email protected]

[email protected]

author

2012-08-03 13:59 UTC #1 lore raw

Hi!

Quoted 7 lines

I'm looking into using LTP on Android and reviewing the discussions by /* Cyril Hrubi, Anupam Srivastava*/ discussion at:
   *http://comments.gmane.org/gmane.linux.ltp/16065
*
and

Quoted 8 lines

   *http://old.nabble.com/-LTP--Porting-of-LTP-on-Android---Patch-to33745295.html#a33812376*
Any pointers/suggestions on this work would be appreciated; like the best patch
available.
Looking through the Ltp-list archives now and the above mentioned discussions.

Well as far as I remeber the attached patch wasn't complete and it was more of a hack than anything else.

On the other hand I'm not against having LTP ported to Android. If the changes comes to us in reasonable small chunks that could be reviewed, I would be happy to help and possibly merge them into the upstream codebase.

Sorry that I couldn't help more here. I don't have any expirience with Android SDK and I do not intend to play with it in the near future.

Signature · 13 lines

Cyril Hrubis
[email protected]

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
Ltp-list mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/ltp-list

A message that is not in the archive (sent off-list or not yet received): [email protected]

[email protected]

author

2012-08-08 08:27 UTC #3 lore raw

Hi!

Quoted 11 lines

Yesterday we got the MIPS env starting to build the LTP test, currently we aren't
using the Android.mk approach, just minor tweaks to the stock LTP repo. I'm also
building Busybox using:
*http://code.google.com/p/busybox-android/downloads/detail?name=.config
*
as a guide. I've applied Sterisco's busybox patch but didn't apply his Bionic patch (yet). It's starting to build and like LPT it's starting to compile.

That seems to be way better than the Android.mk mess. And it would certainly be great to have Android HOWTO document bundled with ltp sources.

Quoted 8 lines

Currently my changes to the git-hub LTP master branch are just a bunch of hacks. Seemed like a clear concept on how to do it cleanly wasn't clear yet
and it made more sense to get something starting to work and then clean it up once a clear approach was precipitating out. I'm using the attached Setup
script to use the r8 NDK crosstools and will test with a jellyBean (jb) MIPS emulator.

The patch seems to be relatively small. Is this all that is necessary to build LTP on Android? Have build the open posix testsuite?

Looking at the patch itself. The ifdef 0 shouldn't be there, the !defined(_ANDROID) is better. The FILTER_OUT_DIRS should be either triggered by Android environment or the testcases should be fixed.

The GLIBC_PREREQ() in the accept4_01 should be moved to configure script.

The GLIBC_PREREQ() from tst_tmpdir() is about to be removed. Hm and why is that defined to 0? The uclibc implements mkdtemp() or not?

Quoted 3 lines

Supporting Android in the LTP and Busybox repositories seems like a good idea to me. I see little sense in folks re-doing this procedure from email descriptions.

Well I can't do much for Busybox but I could help to fix LTP to build on Android ;).

Quoted 257 lines

include/tlibio.h +1 -1

index 0ac52e4..14ad640 100644
@@ -142,7 +142,7 @@ int  lio_check_asyncio(char *io_type, int size, struct iosw *status);
 int  lio_wait4asyncio(int method, int fd, aiocb_t *aiocbp);
 int  lio_check_asyncio(char *io_type, int size, aiocb_t *aiocbp, int method);
 #endif /* sgi */
-#if defined(__linux__) && !defined(__UCLIBC__)
+#if defined(__linux__) && !defined(__UCLIBC__) && 0
 #include <aio.h>
 int  lio_wait4asyncio(int method, int fd, struct aiocb *aiocbp);
 int  lio_check_asyncio(char *io_type, int size, struct aiocb *aiocbp, int method);

lib/Makefile +2 -0

index 453cf95..a0081d6 100644
@@ -24,6 +24,8 @@ top_srcdir		?= ..
 
 include $(top_srcdir)/include/mk/env_pre.mk
 
+CFLAGS			+=-D__UCLIBC__
+
 ifeq ($(UCLINUX),1)
 CFLAGS			+= -D_USC_LIB_
 endif

lib/tlibio.c +4 -4

index eaaaaa0..7bb507f 100644
@@ -108,7 +108,7 @@
 #include <string.h>  /* bzero */
 #endif
 #if defined(__linux__) || defined(__sun) || defined(__hpux) || defined(_AIX)
-#if !defined(UCLINUX) && !defined(__UCLIBC__)
+#if !defined(UCLINUX) && !defined(__UCLIBC__) && !defined(_ANDROID)
 #include <aio.h>
 #endif
 #endif

@@ -561,7 +561,7 @@ long wrd;	/* to allow future features, use zero for now */
     aiocb_t *aiolist[1]; /* list of aio control blocks for lio_listio */
     off64_t poffset;	/* pwrite(2) offset */
 #endif
-#if defined(__linux__) && !defined(__UCLIBC__)
+#if defined(__linux__) && !defined(__UCLIBC__) && 0
 	struct aiocb aiocbp;	/* POSIX aio control block */
 	struct aiocb *aiolist[1]; /* list of aio control blocks for lio_listio */
 	off64_t poffset;	/* pwrite(2) offset */

@@ -595,7 +595,7 @@ long wrd;	/* to allow future features, use zero for now */
     iov.iov_base = buffer;
     iov.iov_len = size;
 #endif
-#if defined(sgi) || (defined(__linux__) && !defined(__UCLIBC__))
+#if defined(sgi) || (defined(__linux__) && !defined(__UCLIBC__) && 0)
 #if defined(sgi)
     memset(&aiocbp, 0x00, sizeof(aiocb_t));
 #else

@@ -841,7 +841,7 @@ long wrd;	/* to allow future features, use zero for now */
 	return ret;
 
 #endif
-#if defined(sgi) || (defined(__linux__) && !defined(__UCLIBC__))
+#if defined(sgi) || (defined(__linux__) && !defined(__UCLIBC__)) && 0
 
 	aiocbp.aio_lio_opcode = LIO_WRITE;
 	listio_cmd=LIO_WAIT;

lib/tst_tmpdir.c +3 -1

index d112e3e..3e50eaf 100644
@@ -132,7 +132,9 @@ void tst_tmpdir(void)
 	int  no_cleanup = 0;	/* !0 means TDIRECTORY env var was set */
 	char *env_tmpdir;	/* temporary storage for TMPDIR env var */
 	/* This is an AWFUL hack to figure out if mkdtemp() is available */
-#if defined(__GLIBC_PREREQ) && __GLIBC_PREREQ(2,2)
+
+// #if 0 && defined(__GLIBC_PREREQ) && __GLIBC_PREREQ(2,2)
+#if 0 
 #define HAVE_MKDTEMP
 #endif

testcases/Makefile +1 -1

index 61993fe..fa3986a 100644
@@ -28,7 +28,7 @@ include $(top_srcdir)/include/mk/env_pre.mk
 # 1. kdump shouldn't be compiled by default, because it's runtime based and
 #    WILL crash the build host (the tests need to be fixed to just build, not
 #    run).
-FILTER_OUT_DIRS		:= kdump
+FILTER_OUT_DIRS		:= kdump commands
 
 ifneq ($(WITH_OPEN_POSIX_TESTSUITE),yes)
 FILTER_OUT_DIRS		+= open_posix_testsuite

testcases/commands/Makefile +3 -1

index 9da27c3..ec85f8b 100644
@@ -24,7 +24,9 @@
 top_srcdir		?= ../..
 
 ifneq ($(WITH_EXPECT),yes)
-FILTER_OUT_SUBDIRS	:= tpm-tools
+FILTER_OUT_SUBDIRS	:= tpm-tools cpio
+else
+FILTER_OUT_SUBDIRS      := cpio
 endif
 
 include $(top_srcdir)/include/mk/env_pre.mk

testcases/kernel/syscalls/Makefile +2 -0

index cee99e9..facf741 100644
@@ -29,4 +29,6 @@ FILTER_OUT_DIRS	+= capget capset chmod chown clone fork getcontext llseek \
 		   nftw profil remap_file_pages
 endif
 
+FILTER_OUT_DIRS += add_key adjtimex bdflush
+
 include $(top_srcdir)/include/mk/generic_trunk_target.mk

testcases/kernel/syscalls/accept4/accept4_01.c +4 -2

index d3a39df..e270d39 100644
@@ -135,7 +135,8 @@ void setup()
 	TEST_PAUSE;
 	tst_tmpdir();
 }
-#if !(__GLIBC_PREREQ(2,10))
+// #if !(__GLIBC_PREREQ(2,10))
+#if 1
 static int
 accept4_01(int fd, struct sockaddr *sockaddr, socklen_t *addrlen, int flags)
 {

@@ -192,7 +193,8 @@ do_test(int lfd, struct sockaddr_in *conn_addr,
 		die("Connect Error");
 
 	addrlen = sizeof(struct sockaddr_in);
-	#if !(__GLIBC_PREREQ(2,10))
+//	#if !(__GLIBC_PREREQ(2,10))
+	#if 1
 	acceptfd = accept4_01(lfd, (struct sockaddr *)&claddr, &addrlen,
                            closeonexec_flag | nonblock_flag);
 	#else

testcases/kernel/syscalls/chmod/Makefile +2 -0

index bd617d8..f293025 100644
@@ -20,4 +20,6 @@ top_srcdir		?= ../../../..
 
 include $(top_srcdir)/include/mk/testcases.mk
 
+FILTER_OUT_MAKE_TARGETS := chmod06
+
 include $(top_srcdir)/include/mk/generic_leaf_target.mk

testcases/kernel/syscalls/ptrace/Makefile +1 -1

index 078792b..486ab56 100644
@@ -44,6 +44,6 @@ include $(top_srcdir)/include/mk/testcases.mk
 #  make: *** [testcases-all] Error 2
 #
 
-FILTER_OUT_MAKE_TARGETS	:= ptrace06 simple_tracer
+FILTER_OUT_MAKE_TARGETS	:= ptrace03 ptrace04 ptrace06 simple_tracer
 
 include $(top_srcdir)/include/mk/generic_leaf_target.mk

utils/benchmark/ebizzy-0.3/Makefile +1 -1

index 41aadd4..b23ded6 100644
@@ -27,7 +27,7 @@ top_srcdir		?= ../../..
 
 include $(top_srcdir)/include/mk/env_pre.mk
 
-LDLIBS			+= -lpthread
+# LDLIBS			+= -lpthread
 
 WCPPFLAGS		+= -Wshadow

utils/benchmark/ebizzy-0.3/ebizzy.c +3 -2

index 7c2c2c1..7b87f42 100644
@@ -218,9 +218,10 @@ read_options(int argc, char *argv[])
 			"\"never mmap\" option specified\n");
 		usage();
 	}
-
+#if 0
 	if (never_mmap)
 		mallopt(M_MMAP_MAX, 0);
+#endif
 
 	if (chunk_size < record_size) {
 		fprintf(stderr, "Chunk size %u smaller than record size %u\n",

@@ -533,4 +534,4 @@ main(int argc, char *argv[])
 	start_threads();
 
 	return 0;
-}
\ No newline at end of file
+}

export LTP_ROOT=pwd

export NDK_ROOT=/home/piet/src/dev-mips-jb/mozilla/android-ndk-r8

export MIPS_TOOL_CHAIN=${NDK_ROOT}/toolchains/mipsel-linux-android-4.4.3/prebuilt/linux-x86

# export MIPS_INCLUDES=${NDK_ROOT}/platforms/android-3/arch-mips/usr/include
export MIPS_INCLUDES=${NDK_ROOT}/platforms/android-14/arch-mips/usr/include

export PATH=${MIPS_TOOL_CHAIN}/bin:$PATH

export ARCH=mips

export CROSS_COMPILE=mipsel-linux-android-

export CC=${MIPS_TOOL_CHAIN}/bin/${CROSS_COMPILE}gcc

export LD=${MIPS_TOOL_CHAIN}/bin/${CROSS_COMPILE}ld

export AR=${MIPS_TOOL_CHAIN}/bin/${CROSS_COMPILE}ar

export STRIP=${MIPS_TOOL_CHAIN}/bin/${CROSS_COMPILE}strip

export RANLIB=${MIPS_TOOL_CHAIN}/bin/${CROSS_COMPILE}ranlib

export CXX=${MIPS_TOOL_CHAIN}/bin/${CROSS_COMPILE}c++

export AS=${MIPS_TOOL_CHAIN}/bin/${CROSS_COMPILE}as

export GXX=${MIPS_TOOL_CHAIN}/bin/${CROSS_COMPILE}g++

export GCC=${MIPS_TOOL_CHAIN}/bin/${CROSS_COMPILE}gcc

export NM=${MIPS_TOOL_CHAIN}/bin/${CROSS_COMPILE}nm

export CPPFLAGS=-I${MIPS_INCLUDES}/

export CFLAGS='-nostdlib -DDEV_BSIZE=512'

# export WCFLAGS='-pedantic -std=c99'
# export WCFLAGS='-pedantic'
export WCFLAGS=''

# export LDFLAGS=-Wl,???entry=main,???no-undefined,-rpath-link=${MIPS_TOOL_CHAIN}/lib/-L${MIPS_TOOL_CHAIN}/lib/
export LDFLAGS=-Wl,-rpath-link=${MIPS_TOOL_CHAIN}/lib/-L${MIPS_TOOL_CHAIN}/lib/

# export LIBS=-lc
export LIBS=

# ./configure --host=mipsel-linux-android  --target=emulator-5554 --build=x86_64-linux-gnu --prefix=${LTP_ROOT}/objdir
./configure --host=mipsel-linux            --target=emulator-5554 --build=x86_64-linux-gnu --prefix=${LTP_ROOT}/objdir

Signature · 13 lines

Cyril Hrubis
[email protected]

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
Ltp-list mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/ltp-list