[dpdk-dev] [PATCH] spinlock:move constructor function out of header
2016-07-14 22:45, Wiles, Keith: > > On Jul 14, 2016, at 2:59 PM, Thomas Monjalon > > wrote: > > > > Thanks Keith for continuing work. > > > > 2016-07-14 14:31, Keith Wiles: > >> lib/librte_eal/bsdapp/eal/Makefile | 1 + > >> lib/librte_eal/common/arch/arm/rte_spinlock.c | 46 > >> ++ > >> lib/librte_eal/common/arch/ppc_64/rte_spinlock.c | 46 > >> ++ > >> lib/librte_eal/common/arch/tile/rte_spinlock.c | 46 > >> ++ > >> lib/librte_eal/common/arch/x86/rte_spinlock.c | 46 > >> ++ > >> .../common/include/arch/x86/rte_spinlock.h | 14 ++- > >> lib/librte_eal/linuxapp/eal/Makefile | 1 + > > > > I am not sure we should add a .c file for each arch, given it is called only > > from arch/x86/rte_spinlock.h. > > I did not like having the .c for everyone, but the previous comment seemed > to suggest it. I am willing to change it any better method, just let me > know what you think. I would like just one. I will make sure it is not needed. In this case, we can keep the original patch from Damjan and just do some trivial changes. I can make them quickly before RC3. > On a side note I have combined the bsdapp and linuxapp into a single > directory before. It is doable and it eliminates a number of duplicate > files or code. Yes patches to remove duplicated code are welcome. But please do not introduce more #ifdefs. I think it is better to keep separate directories bsdapp/ and linuxapp/ while increasing the shared code in common/ as much as possible. Some functions are really different and are better handled separately. > Plus a also added support for OS X for DPDK, but I do not have access > to any NICs with that version yet other then virtual ones. > I could submit it and may be someone will write the kext to make it work. :-) Maybe that OS X would deserve a third separate directory. I guess you wanted it only for dev testing? Why not just use a Linux or FreeBSD VM?
[dpdk-dev] [PATCH] spinlock:move constructor function out of header
> On Jul 14, 2016, at 2:59 PM, Thomas Monjalon
> wrote:
>
> Thanks Keith for continuing work.
>
> 2016-07-14 14:31, Keith Wiles:
>> Moving the rte_rtm_init() constructor routine out of the
>> header file and into a new rte_spinlock.c for all archs/platforms.
>> Having constructor routines in a header file is not a good
>> place to keep these types of functions.
>>
>> The problem is with linking 3rd party libraries when
>> an application is not linked directly to dpdk libraries, which
>> in this case causes a missing symbol in the linking phase.
>>
>> Fixes: ba7468997ea6 ("spinlock: add HTM lock elision for x86")
>>
>> Originally submitted by Damjan Marion
>
> You should keep the original Signed-off and authorship (From: field)
> with "git am".
> It is also easier to track when using -v2 --in-reply-to='?.
I can try and add that into the patch for v2.
>
>> Signed-off-by: Keith Wiles
>> ---
>> lib/librte_eal/bsdapp/eal/Makefile | 1 +
>> lib/librte_eal/common/arch/arm/rte_spinlock.c | 46
>> ++
>> lib/librte_eal/common/arch/ppc_64/rte_spinlock.c | 46
>> ++
>> lib/librte_eal/common/arch/tile/rte_spinlock.c | 46
>> ++
>> lib/librte_eal/common/arch/x86/rte_spinlock.c | 46
>> ++
>> .../common/include/arch/x86/rte_spinlock.h | 14 ++-
>> lib/librte_eal/linuxapp/eal/Makefile | 1 +
>
> I am not sure we should add a .c file for each arch, given it is called only
> from arch/x86/rte_spinlock.h.
I did not like having the .c for everyone, but the previous comment seemed to
suggest it. I am willing to change it any better method, just let me know what
you think. I would like just one.
On a side note I have combined the bsdapp and linuxapp into a single directory
before. It is doable and it eliminates a number of duplicate files or code.
Plus a also added support for OS X for DPDK, but I do not have access to any
NICs with that version yet other then virtual ones. I could submit it and may
be someone will write the kext to make it work. :-)
[dpdk-dev] [PATCH] spinlock:move constructor function out of header
Thanks Keith for continuing work.
2016-07-14 14:31, Keith Wiles:
> Moving the rte_rtm_init() constructor routine out of the
> header file and into a new rte_spinlock.c for all archs/platforms.
> Having constructor routines in a header file is not a good
> place to keep these types of functions.
>
> The problem is with linking 3rd party libraries when
> an application is not linked directly to dpdk libraries, which
> in this case causes a missing symbol in the linking phase.
>
> Fixes: ba7468997ea6 ("spinlock: add HTM lock elision for x86")
>
> Originally submitted by Damjan Marion
You should keep the original Signed-off and authorship (From: field)
with "git am".
It is also easier to track when using -v2 --in-reply-to=''.
> Signed-off-by: Keith Wiles
> ---
> lib/librte_eal/bsdapp/eal/Makefile | 1 +
> lib/librte_eal/common/arch/arm/rte_spinlock.c | 46
> ++
> lib/librte_eal/common/arch/ppc_64/rte_spinlock.c | 46
> ++
> lib/librte_eal/common/arch/tile/rte_spinlock.c | 46
> ++
> lib/librte_eal/common/arch/x86/rte_spinlock.c | 46
> ++
> .../common/include/arch/x86/rte_spinlock.h | 14 ++-
> lib/librte_eal/linuxapp/eal/Makefile | 1 +
I am not sure we should add a .c file for each arch, given it is called only
from arch/x86/rte_spinlock.h.
[dpdk-dev] [PATCH] spinlock:move constructor function out of header
Moving the rte_rtm_init() constructor routine out of the
header file and into a new rte_spinlock.c for all archs/platforms.
Having constructor routines in a header file is not a good
place to keep these types of functions.
The problem is with linking 3rd party libraries when
an application is not linked directly to dpdk libraries, which
in this case causes a missing symbol in the linking phase.
Fixes: ba7468997ea6 ("spinlock: add HTM lock elision for x86")
Originally submitted by Damjan Marion
Signed-off-by: Keith Wiles
---
lib/librte_eal/bsdapp/eal/Makefile | 1 +
lib/librte_eal/common/arch/arm/rte_spinlock.c | 46 ++
lib/librte_eal/common/arch/ppc_64/rte_spinlock.c | 46 ++
lib/librte_eal/common/arch/tile/rte_spinlock.c | 46 ++
lib/librte_eal/common/arch/x86/rte_spinlock.c | 46 ++
.../common/include/arch/x86/rte_spinlock.h | 14 ++-
lib/librte_eal/linuxapp/eal/Makefile | 1 +
7 files changed, 190 insertions(+), 10 deletions(-)
create mode 100644 lib/librte_eal/common/arch/arm/rte_spinlock.c
create mode 100644 lib/librte_eal/common/arch/ppc_64/rte_spinlock.c
create mode 100644 lib/librte_eal/common/arch/tile/rte_spinlock.c
create mode 100644 lib/librte_eal/common/arch/x86/rte_spinlock.c
diff --git a/lib/librte_eal/bsdapp/eal/Makefile
b/lib/librte_eal/bsdapp/eal/Makefile
index 698fa0a..cd93f5b 100644
--- a/lib/librte_eal/bsdapp/eal/Makefile
+++ b/lib/librte_eal/bsdapp/eal/Makefile
@@ -89,6 +89,7 @@ SRCS-$(CONFIG_RTE_EXEC_ENV_BSDAPP) += rte_keepalive.c
# from arch dir
SRCS-$(CONFIG_RTE_EXEC_ENV_BSDAPP) += rte_cpuflags.c
+SRCS-$(CONFIG_RTE_EXEC_ENV_BSDAPP) += rte_spinlock.c
CFLAGS_eal_common_cpuflags.o := $(CPUFLAGS_LIST)
diff --git a/lib/librte_eal/common/arch/arm/rte_spinlock.c
b/lib/librte_eal/common/arch/arm/rte_spinlock.c
new file mode 100644
index 000..bc924e5
--- /dev/null
+++ b/lib/librte_eal/common/arch/arm/rte_spinlock.c
@@ -0,0 +1,46 @@
+/*-
+ * BSD LICENSE
+ *
+ * Copyright(c) 2016 Intel Corporation. All rights reserved.
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * * Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * * Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in
+ * the documentation and/or other materials provided with the
+ * distribution.
+ * * Neither the name of Intel Corporation nor the names of its
+ * contributors may be used to endorse or promote products derived
+ * from this software without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include
+
+#include "rte_cpuflags.h"
+
+/**< cache the flag to avoid the overhead of the function call */
+int rte_rtm_supported;
+
+static void __attribute__((constructor))
+rte_rtm_init(void)
+{
+ rte_rtm_supported = rte_cpu_get_flag_enabled(RTE_CPUFLAG_RTM);
+}
+
diff --git a/lib/librte_eal/common/arch/ppc_64/rte_spinlock.c
b/lib/librte_eal/common/arch/ppc_64/rte_spinlock.c
new file mode 100644
index 000..bc924e5
--- /dev/null
+++ b/lib/librte_eal/common/arch/ppc_64/rte_spinlock.c
@@ -0,0 +1,46 @@
+/*-
+ * BSD LICENSE
+ *
+ * Copyright(c) 2016 Intel Corporation. All rights reserved.
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ * * Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * * Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in
+ * the documentation and/or other materials provided with the
+ * distribution.
+ * * Neither the name of Intel Co
