Re: Waf branch - Overview [2/6]

[Chris Johns]

On 6/02/2015 6:03 pm, Amar Takhar wrote:


Here is the timing for the build at the moment:

verm@peach# waf config --bsp sparc/sis --prefix /mnt/devel/rtems/prefix/rtems
--path-tools /mnt/devel/rtems/tools/bin/
'config' finished successfully (0.026s)

verm@peach# waf configure
'configure' finished successfully (4.034s)

verm@peach# waf -j 14
'build' finished successfully (16.204s)

   note: the above is 1,337 objects.

verm@peach# waf -j 14 --enable-tests
'build' finished successfully (13.414s)

   note: the above is 881 targets.

 From a clone we have: 0.026 + 4.034 + 16.204 + 13.414 = 33.678s.



I think we need to create a table in the wiki for the performance we get 
for each type of machine, an so at the top of the table is 


MacPro (Maverick)

sparc/sis:
 config: 0.016s
 configure : 3.836s
 build : 12.793s
 tests : 10.073s
 Total : 26.718s

Chris
___
devel mailing list
devel@rtems.org
http://lists.rtems.org/mailman/listinfo/devel

Re: [PATCH] bsp/sparc: Move BSP_ISR_handler to a separate file

[Gedare Bloom]

On Fri, Feb 6, 2015 at 10:48 AM, Daniel Cederman  wrote:
>> I'd like to entertain a new name for this "BSP_ISR_handler" function
>> now that it becomes an exported symbol. We should put it into a proper
>> namespace.
>
> Any suggestion on what to change it to?
>
bsp_isr_handler() is fine with me.

> /Daniel C
>
>
> On 2015-02-04 16:12, Gedare Bloom wrote:
>>
>> I'd like to entertain a new name for this "BSP_ISR_handler" function
>> now that it becomes an exported symbol. We should put it into a proper
>> namespace.
>>
>> Also, I'm not sure on the legality of putting a copyright for Cobham
>> Gaisler that predates Cobham's acquisition of Aeroflex. But I am not a
>> lawyer.
>>
>> -Gedare
>>
>> On Wed, Feb 4, 2015 at 4:52 AM, Daniel Cederman 
>> wrote:
>>>
>>> This allows it to be wrapped by another function at link-time
>>> and can be used to trace interrupts. If not placed in a separate
>>> file, the function pointer address used in BSP_shared_interrupt_init
>>> will be resolved at compile-time, and the function will not be wrappable.
>>> ---
>>>   c/src/lib/libbsp/sparc/Makefile.am |  1 +
>>>   c/src/lib/libbsp/sparc/erc32/Makefile.am   |  1 +
>>>   c/src/lib/libbsp/sparc/erc32/include/bsp.h |  3 ++
>>>   c/src/lib/libbsp/sparc/leon2/Makefile.am   |  1 +
>>>   c/src/lib/libbsp/sparc/leon2/include/bsp.h |  3 ++
>>>   c/src/lib/libbsp/sparc/leon3/Makefile.am   |  1 +
>>>   c/src/lib/libbsp/sparc/leon3/include/bsp.h |  3 ++
>>>   .../lib/libbsp/sparc/shared/irq/bsp_isr_handler.c  | 35
>>> ++
>>>   c/src/lib/libbsp/sparc/shared/irq/irq-shared.c | 32
>>> +++-
>>>   9 files changed, 58 insertions(+), 22 deletions(-)
>>>   create mode 100644 c/src/lib/libbsp/sparc/shared/irq/bsp_isr_handler.c
>>>
>>> diff --git a/c/src/lib/libbsp/sparc/Makefile.am
>>> b/c/src/lib/libbsp/sparc/Makefile.am
>>> index 3d433fb..615aea7 100644
>>> --- a/c/src/lib/libbsp/sparc/Makefile.am
>>> +++ b/c/src/lib/libbsp/sparc/Makefile.am
>>> @@ -11,6 +11,7 @@ EXTRA_DIST += shared/start/start.S
>>>
>>>   # Interrupt
>>>   EXTRA_DIST += shared/irq/irq-shared.c
>>> +EXTRA_DIST += shared/irq/bsp_isr_handler.c
>>>
>>>   # AMBA Plug&Play bus
>>>   EXTRA_DIST += shared/include/ambapp.h
>>> diff --git a/c/src/lib/libbsp/sparc/erc32/Makefile.am
>>> b/c/src/lib/libbsp/sparc/erc32/Makefile.am
>>> index c99d2fb..8dcdd32 100644
>>> --- a/c/src/lib/libbsp/sparc/erc32/Makefile.am
>>> +++ b/c/src/lib/libbsp/sparc/erc32/Makefile.am
>>> @@ -71,6 +71,7 @@ include_bsp_HEADERS += \
>>>   include/bsp/irq.h
>>>   libbsp_a_SOURCES += \
>>>   ../../sparc/shared/irq/irq-shared.c \
>>> +../../sparc/shared/irq/bsp_isr_handler.c \
>>>   ../../shared/src/irq-default-handler.c \
>>>   ../../shared/src/irq-generic.c \
>>>   ../../shared/src/irq-info.c \
>>> diff --git a/c/src/lib/libbsp/sparc/erc32/include/bsp.h
>>> b/c/src/lib/libbsp/sparc/erc32/include/bsp.h
>>> index 3393910..2e531ec 100644
>>> --- a/c/src/lib/libbsp/sparc/erc32/include/bsp.h
>>> +++ b/c/src/lib/libbsp/sparc/erc32/include/bsp.h
>>> @@ -106,6 +106,9 @@ typedef void (*bsp_shared_isr)(void *arg);
>>>   /* Initializes the Shared System Interrupt service */
>>>   extern void BSP_shared_interrupt_init(void);
>>>
>>> +/* Called directly from IRQ trap handler TRAP[0x10..0x1F] = IRQ[0..15]
>>> */
>>> +void BSP_ISR_handler(rtems_vector_number vector);
>>> +
>>>   /* Registers a shared IRQ handler, and enable it at IRQ controller.
>>> Multiple
>>>* interrupt handlers may use the same IRQ number, all ISRs will be
>>> called
>>>* when an interrupt on that line is fired.
>>> diff --git a/c/src/lib/libbsp/sparc/leon2/Makefile.am
>>> b/c/src/lib/libbsp/sparc/leon2/Makefile.am
>>> index d1e3817..efcb286 100644
>>> --- a/c/src/lib/libbsp/sparc/leon2/Makefile.am
>>> +++ b/c/src/lib/libbsp/sparc/leon2/Makefile.am
>>> @@ -80,6 +80,7 @@ include_bsp_HEADERS += \
>>>   include/bsp/irq.h
>>>   libbsp_a_SOURCES += \
>>>   ../../sparc/shared/irq/irq-shared.c \
>>> +../../sparc/shared/irq/bsp_isr_handler.c \
>>>   ../../shared/src/irq-default-handler.c \
>>>   ../../shared/src/irq-generic.c \
>>>   ../../shared/src/irq-info.c \
>>> diff --git a/c/src/lib/libbsp/sparc/leon2/include/bsp.h
>>> b/c/src/lib/libbsp/sparc/leon2/include/bsp.h
>>> index 41a1e43..807f8d9 100644
>>> --- a/c/src/lib/libbsp/sparc/leon2/include/bsp.h
>>> +++ b/c/src/lib/libbsp/sparc/leon2/include/bsp.h
>>> @@ -130,6 +130,9 @@ typedef void (*bsp_shared_isr)(void *arg);
>>>   /* Initializes the Shared System Interrupt service */
>>>   extern void BSP_shared_interrupt_init(void);
>>>
>>> +/* Called directly from IRQ trap handler TRAP[0x10..0x1F] = IRQ[0..15]
>>> */
>>> +void BSP_ISR_handler(rtems_vector_number vector);
>>> +
>>>   /* Registers a shared IRQ handler, and enable it at IRQ controller.
>>> Multiple
>>>* interrupt handlers may use the same IRQ number, all ISRs will be
>>> called
>>>* when

Re: Waf branch - Overview [2/6]

[Amar Takhar]

On 2015-02-07 19:22 +1100, Chris Johns wrote:
> MacPro (Maverick)
 

I have added a page here for this:

  https://devel.rtems.org/wiki/waf/Timing

Note the performance on Debian.  There is something serious going on with 
FreeBSD and python that also hits OS X.  Not sure what it is yet.


Amar.
___
devel mailing list
devel@rtems.org
http://lists.rtems.org/mailman/listinfo/devel

Re: Waf branch - Overview [2/6]

[Chris Johns]

On 6/02/2015 6:03 pm, Amar Takhar wrote:


verm@peach# waf -j 14 --enable-tests
'build' finished successfully (13.414s)

   note: the above is 881 targets.



I have updated the rtems-test command to install when the 
rtems-tools.git package is installed (which the RSB automatically does).


I ran the tests from the top of the waf build with:

 $ rtems-test --rtems-bsp=sis --filter=test_* build/sparc/sis/testsuites

The important part of the results are:

Passed:   289
Failed: 0
Timeouts:  13
Invalid:5
-
Total:307

Timeouts:
 test_sptests_sp22
 test_sptests_sp31
 test_sptests_sp16
 test_sptests_sp14
 test_sptests_sp12
 test_sptests_sp13
 test_sptests_sp11
 test_sptests_sp07
 test_sptests_sp25
 test_tmtests_tmoverhd
 test_tmtests_tmck
 test_sptests_sp43
 test_sptests_sp42
Invalid:
 test_extend.c.35.o
 test_support.c.3.o
 test_sptests_sp36
 test_cat.c.35.o
 test_write.c.35.o

I am not sure if this is the total number of tests. I thought it was 
around 500+ tests including samples.


With the autotools build system the executables all have a .exe 
extension and I cannot see how to get a suitable list of executables.


Chris
___
devel mailing list
devel@rtems.org
http://lists.rtems.org/mailman/listinfo/devel