[dpdk-dev] [PATCH] eal: add missing long-options for short option arguments
On Thu, Mar 3, 2016 at 4:02 PM, Wiles, Keith wrote: >>On Thu, Feb 25, 2016 at 11:12 PM, Wiles, Keith >>wrote: On Thu, Feb 25, 2016 at 01:09:16PM -0600, Keith Wiles wrote: > A number of short options for EAL are missing long options > and this patch adds those missing options. > > The missing long options are for: > -c add --coremask > -d add --driver > -l add --corelist > -m add --memsize > -n add --mem-channels > -r add --mem-ranks > -v add --version > Add an alias for --lcores using --lcore-map [snip] >>No strong opinion on this. >> >>Just, why "memsize" with no - but "mem-channels" ? >>And why cut down to mem rather than memory ? > > I debated on mem-size, but I noticed in a couple places some used memsize. I > can change them to any thing someone wants. If you want memory-channels and > memory-ranks I am good with that too. Ok, since I can see no real argument against, let's go with explicit options : --memory-size --memory-channels --memory-ranks The best would be then to align all other existing long options (but keeping compat with existing ones). Deal ? -- David Marchand
[dpdk-dev] [PATCH] eal: add missing long-options for short option arguments
>On Thu, Feb 25, 2016 at 11:12 PM, Wiles, Keith >wrote: >>>On Thu, Feb 25, 2016 at 01:09:16PM -0600, Keith Wiles wrote: A number of short options for EAL are missing long options and this patch adds those missing options. The missing long options are for: -c add --coremask -d add --driver -l add --corelist -m add --memsize -n add --mem-channels -r add --mem-ranks -v add --version Add an alias for --lcores using --lcore-map Signed-off-by: Keith Wiles >>> >>>Why do we need long options for all the short options? >> >> I think we need the long options to match the short options just because it >> makes sense to me to have long options for all short options. Take the case >> of -v, just about everyone else has a ?version long-option, but we do not. >> >> The real reason is to allow for DPDK configuration via a configuration file >> and I wanted to use the same strings for the config file variables as the >> command line options. I figured I would add the long options now as they do >> not effect the configuration file patch. > >No strong opinion on this. > >Just, why "memsize" with no - but "mem-channels" ? >And why cut down to mem rather than memory ? I debated on mem-size, but I noticed in a couple places some used memsize. I can change them to any thing someone wants. If you want memory-channels and memory-ranks I am good with that too. > > >-- >David Marchand > Regards, Keith
[dpdk-dev] [PATCH] eal: add missing long-options for short option arguments
>>On Thu, Feb 25, 2016 at 01:09:16PM -0600, Keith Wiles wrote: >>> A number of short options for EAL are missing long options >>> and this patch adds those missing options. >>> >>> The missing long options are for: >>> -c add --coremask >>> -d add --driver >>> -l add --corelist >>> -m add --memsize >>> -n add --mem-channels >>> -r add --mem-ranks >>> -v add --version >>> Add an alias for --lcores using --lcore-map >>> >>> Signed-off-by: Keith Wiles >> >>Why do we need long options for all the short options? > >I think we need the long options to match the short options just because it >makes sense to me to have long options for all short options. Take the case of >-v, just about everyone else has a ?version long-option, but we do not. > >The real reason is to allow for DPDK configuration via a configuration file >and I wanted to use the same strings for the config file variables as the >command line options. I figured I would add the long options now as they do >not effect the configuration file patch. Ping. I really want to have long options for the short option to allow me to use those same options for the config file support I would like to use for DPDK. A config file support is much more reasonable for live or production systems IMHO. Plus it could be very nice for the examples to have a config file on how that example could be configured. I can create the config file support without the long option names for the short ones, but it would be a lot cleaner to have the same names for config file and command line. Thanks ++Keith >> >>/Bruce >> >> > > >Regards, >Keith > > > > > Regards, Keith
[dpdk-dev] [PATCH] eal: add missing long-options for short option arguments
>On Thu, Feb 25, 2016 at 01:09:16PM -0600, Keith Wiles wrote: >> A number of short options for EAL are missing long options >> and this patch adds those missing options. >> >> The missing long options are for: >> -c add --coremask >> -d add --driver >> -l add --corelist >> -m add --memsize >> -n add --mem-channels >> -r add --mem-ranks >> -v add --version >> Add an alias for --lcores using --lcore-map >> >> Signed-off-by: Keith Wiles > >Why do we need long options for all the short options? I think we need the long options to match the short options just because it makes sense to me to have long options for all short options. Take the case of -v, just about everyone else has a ?version long-option, but we do not. The real reason is to allow for DPDK configuration via a configuration file and I wanted to use the same strings for the config file variables as the command line options. I figured I would add the long options now as they do not effect the configuration file patch. > >/Bruce > > Regards, Keith
[dpdk-dev] [PATCH] eal: add missing long-options for short option arguments
On Thu, Feb 25, 2016 at 01:09:16PM -0600, Keith Wiles wrote: > A number of short options for EAL are missing long options > and this patch adds those missing options. > > The missing long options are for: > -c add --coremask > -d add --driver > -l add --corelist > -m add --memsize > -n add --mem-channels > -r add --mem-ranks > -v add --version > Add an alias for --lcores using --lcore-map > > Signed-off-by: Keith Wiles Why do we need long options for all the short options? /Bruce
[dpdk-dev] [PATCH] eal: add missing long-options for short option arguments
A number of short options for EAL are missing long options
and this patch adds those missing options.
The missing long options are for:
-c add --coremask
-d add --driver
-l add --corelist
-m add --memsize
-n add --mem-channels
-r add --mem-ranks
-v add --version
Add an alias for --lcores using --lcore-map
Signed-off-by: Keith Wiles
---
doc/guides/testpmd_app_ug/run_app.rst | 16 +++
lib/librte_eal/common/eal_common_options.c | 31 --
lib/librte_eal/common/eal_options.h| 16 +++
3 files changed, 45 insertions(+), 18 deletions(-)
diff --git a/doc/guides/testpmd_app_ug/run_app.rst
b/doc/guides/testpmd_app_ug/run_app.rst
index f605564..753a013 100644
--- a/doc/guides/testpmd_app_ug/run_app.rst
+++ b/doc/guides/testpmd_app_ug/run_app.rst
@@ -38,18 +38,18 @@ The following are the EAL command-line options that can be
used in conjunction w
or any other DPDK application.
See the DPDK Getting Started Guides for more information on these options.
-* ``-c COREMASK``
+* ``-c, --coremask COREMASK``
Set the hexadecimal bitmask of the cores to run on.
-* ``-l CORELIST``
+* ``-l, --corelist CORELIST``
List of cores to run on
The argument format is ``[-c2][,c3[-c4],...]``
where ``c1``, ``c2``, etc are core indexes between 0 and 128.
-* ``--lcores COREMAP``
+* ``--lcores COREMAP or --lcore-map COREMAP``
Map lcore set to physical cpu set
@@ -66,7 +66,7 @@ See the DPDK Getting Started Guides for more information on
these options.
Core ID that is used as master.
-* ``-n NUM``
+* ``-n, --mem-channels NUM``
Set the number of memory channels to use.
@@ -74,7 +74,7 @@ See the DPDK Getting Started Guides for more information on
these options.
Blacklist a PCI devise to prevent EAL from using it. Multiple -b options
are allowed.
-* ``-d LIB.so``
+* ``-d, --driver LIB.so|DIR``
Load an external driver. Multiple -d options are allowed.
@@ -82,15 +82,15 @@ See the DPDK Getting Started Guides for more information on
these options.
Add a PCI device in white list.
-* ``-m MB``
+* ``-m, --memsize MB``
Memory to allocate. See also ``--socket-mem``.
-* ``-r NUM``
+* ``-r, --mem-ranks NUM``
Set the number of memory ranks (auto-detected by default).
-* ``-v``
+* ``-v, --version``
Display the version information on startup.
diff --git a/lib/librte_eal/common/eal_common_options.c
b/lib/librte_eal/common/eal_common_options.c
index 29942ea..cf9801d 100644
--- a/lib/librte_eal/common/eal_common_options.c
+++ b/lib/librte_eal/common/eal_common_options.c
@@ -95,6 +95,14 @@ eal_long_options[] = {
{OPT_VFIO_INTR, 1, NULL, OPT_VFIO_INTR_NUM},
{OPT_VMWARE_TSC_MAP,0, NULL, OPT_VMWARE_TSC_MAP_NUM },
{OPT_XEN_DOM0, 0, NULL, OPT_XEN_DOM0_NUM },
+ {OPT_COREMASK, 1, NULL, OPT_COREMASK_NUM },
+ {OPT_DRIVER,1, NULL, OPT_DRIVER_NUM },
+ {OPT_CORELIST, 1, NULL, OPT_CORELIST_NUM },
+ {OPT_MEM_SIZE, 1, NULL, OPT_MEM_SIZE_NUM },
+ {OPT_MEM_CHANNELS, 1, NULL, OPT_MEM_CHANNELS_NUM },
+ {OPT_MEM_RANKS, 1, NULL, OPT_MEM_RANKS_NUM},
+ {OPT_VERSION, 0, NULL, OPT_VERSION_NUM },
+ {OPT_LCORE_MAP, 1, NULL, OPT_LCORE_MAP_NUM},
{0, 0, NULL, 0}
};
@@ -889,6 +897,7 @@ eal_parse_common_option(int opt, const char *optarg,
conf->log_level = log;
break;
}
+ case OPT_LCORE_MAP_NUM:
case OPT_LCORES_NUM:
if (eal_parse_lcores(optarg) < 0) {
RTE_LOG(ERR, EAL, "invalid parameter for --"
@@ -978,11 +987,13 @@ eal_common_usage(void)
{
printf("[options]\n\n"
"EAL common options:\n"
- " -c COREMASK Hexadecimal bitmask of cores to run on\n"
- " -l CORELIST List of cores to run on\n"
+ " -c, --"OPT_COREMASK" Hexadecimal bitmask of cores to run
on\n"
+ " -l, --"OPT_CORELIST" List of cores to run on\n"
" The argument format is
[-c2][,c3[-c4],...]\n"
" where c1, c2, etc are core indexes
between 0 and %d\n"
- " --"OPT_LCORES" COREMAPMap lcore set to physical cpu set\n"
+ "(ex: 1-3,7,9-10) skipping 4,5,6 and 8
cores.\n"
+ " --"OPT_LCORES" COREMAP\n"
+ " --"OPT_LCORE_MAP" COREMAP Map lcore set to physical cpu set\n"
" The argument format is\n"
"
'[<,lcores[@cpus]>...]'\n"
" lcores and cpus list are grouped by '('
and ')'\n"
@@ -991,9
