[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: gEDA-user: Iverilog support SWIFT models??
Hi all,
The PLI interface seems to be implemented in the module lmtv.o, that is
shipped in the smartmodel lib directory. An `nm lmtv.o' reveals the following
undefined symbols:
----
[409] | 0| 0|FUNC |GLOB |0 |UNDEF |.mul
[357] | 0| 0|NOTY |GLOB |0 |UNDEF |__ctype
[456] | 0| 0|NOTY |GLOB |0 |UNDEF |__iob
[350] | 0| 0|NOTY |GLOB |0 |UNDEF |_GLOBAL_OFFSET_TABLE_
[388] | 0| 0|FUNC |GLOB |0 |UNDEF |acc_close
[542] | 0| 0|FUNC |GLOB |0 |UNDEF |acc_configure
[489] | 0| 0|FUNC |GLOB |0 |UNDEF |acc_fetch_fullname
[540] | 0| 0|FUNC |GLOB |0 |UNDEF |acc_fetch_location
[400] | 0| 0|FUNC |GLOB |0 |UNDEF |acc_fetch_name
[553] | 0| 0|FUNC |GLOB |0 |UNDEF |acc_fetch_paramval
[402] | 0| 0|FUNC |GLOB |0 |UNDEF |acc_fetch_size
[333] | 0| 0|FUNC |GLOB |0 |UNDEF |acc_fetch_tfarg_int
[493] | 0| 0|FUNC |GLOB |0 |UNDEF |acc_fetch_tfarg_str
[564] | 0| 0|FUNC |GLOB |0 |UNDEF
|acc_fetch_timescale_info
[347] | 0| 0|FUNC |GLOB |0 |UNDEF |acc_fetch_type
[341] | 0| 0|FUNC |GLOB |0 |UNDEF |acc_fetch_type_str
[331] | 0| 0|FUNC |GLOB |0 |UNDEF |acc_fetch_value
[561] | 0| 0|FUNC |GLOB |0 |UNDEF |acc_handle_object
[421] | 0| 0|FUNC |GLOB |0 |UNDEF |acc_handle_tfarg
[512] | 0| 0|FUNC |GLOB |0 |UNDEF |acc_handle_tfinst
[375] | 0| 0|FUNC |GLOB |0 |UNDEF |acc_initialize
[315] | 0| 0|FUNC |GLOB |0 |UNDEF |acc_next_bit
[484] | 0| 0|FUNC |GLOB |0 |UNDEF |acc_next_port
[502] | 0| 0|FUNC |GLOB |0 |UNDEF |acc_object_in_typelist
[429] | 0| 0|FUNC |GLOB |0 |UNDEF |acc_object_of_type
[389] | 0| 0|FUNC |GLOB |0 |UNDEF |acc_set_value
[448] | 0| 0|FUNC |GLOB |0 |UNDEF |acc_vcl_add
[556] | 0| 0|FUNC |GLOB |0 |UNDEF |acc_vcl_delete
[360] | 0| 0|FUNC |GLOB |0 |UNDEF |acc_version
[369] | 0| 0|FUNC |GLOB |0 |UNDEF |exit
[428] | 0| 0|FUNC |GLOB |0 |UNDEF |fclose
[528] | 0| 0|FUNC |GLOB |0 |UNDEF |fflush
[515] | 0| 0|FUNC |GLOB |0 |UNDEF |fopen
[562] | 0| 0|FUNC |GLOB |0 |UNDEF |fprintf
[495] | 0| 0|FUNC |GLOB |0 |UNDEF |free
[478] | 0| 0|FUNC |GLOB |0 |UNDEF |io_printf
[384] | 0| 0|FUNC |GLOB |0 |UNDEF |malloc
[558] | 0| 0|FUNC |GLOB |0 |UNDEF |mc_scan_plusargs
[516] | 0| 0|FUNC |GLOB |0 |UNDEF |memset
[481] | 0| 0|FUNC |GLOB |0 |UNDEF |perror
[469] | 0| 0|FUNC |GLOB |0 |UNDEF |realloc
[321] | 0| 0|NOTY |GLOB |0 |UNDEF |slm_lsm_EP
[399] | 0| 0|FUNC |GLOB |0 |UNDEF |slm_lsm_provide
[386] | 0| 0|FUNC |GLOB |0 |UNDEF |sprintf
[508] | 0| 0|FUNC |GLOB |0 |UNDEF |strcmp
[521] | 0| 0|FUNC |GLOB |0 |UNDEF |strcpy
[543] | 0| 0|FUNC |GLOB |0 |UNDEF |strlen
[451] | 0| 0|FUNC |GLOB |0 |UNDEF |strncmp
[550] | 0| 0|FUNC |GLOB |0 |UNDEF |strstr
[466] | 0| 0|FUNC |GLOB |0 |UNDEF |tf_asynchon
[376] | 0| 0|FUNC |GLOB |0 |UNDEF |tf_divide_long
[346] | 0| 0|FUNC |GLOB |0 |UNDEF |tf_error
[518] | 0| 0|FUNC |GLOB |0 |UNDEF |tf_getlongtime
[343] | 0| 0|FUNC |GLOB |0 |UNDEF |tf_getp
[471] | 0| 0|FUNC |GLOB |0 |UNDEF |tf_gettime
[452] | 0| 0|FUNC |GLOB |0 |UNDEF |tf_getworkarea
[349] | 0| 0|FUNC |GLOB |0 |UNDEF |tf_igettimeprecision
[329] | 0| 0|FUNC |GLOB |0 |UNDEF |tf_isynchronize
[423] | 0| 0|FUNC |GLOB |0 |UNDEF |tf_multiply_long
[473] | 0| 0|FUNC |GLOB |0 |UNDEF |tf_nodeinfo
[430] | 0| 0|FUNC |GLOB |0 |UNDEF |tf_nump
[503] | 0| 0|FUNC |GLOB |0 |UNDEF |tf_putp
[442] | 0| 0|FUNC |GLOB |0 |UNDEF |tf_read_restart
[367] | 0| 0|FUNC |GLOB |0 |UNDEF |tf_setdelay
[551] | 0| 0|FUNC |GLOB |0 |UNDEF |tf_setworkarea
[379] | 0| 0|FUNC |GLOB |0 |UNDEF |tf_spname
[490] | 0| 0|FUNC |GLOB |0 |UNDEF |tf_synchronize
[549] | 0| 0|FUNC |GLOB |0 |UNDEF |tf_typep
[424] | 0| 0|FUNC |GLOB |0 |UNDEF |tf_write_save
[380] | 0| 0|FUNC |GLOB |0 |UNDEF |tolower
[342] | 0| 0|FUNC |GLOB |0 |UNDEF |vfprintf
[426] | 0| 0|FUNC |GLOB |0 |UNDEF |vsprintf
------
Most are either C library functions or PLI functions. So, if the Icarus
Verilog has all these tf_* and acc_* functions swift is in business. The only
two functions that I don't recognize are the slm_* ones. They may turn up in
some other shared library. You will have to try and link it to see what goes
wrong.
Mike
Paul Michael Hartke <phartke@Stanford.EDU> said:
> This issue has come up before and I've talked to stephen about it.
> Basically, Synopysys controls the specification for the SWIFT interface
> and they are not willing to let it out unless they know exactly what you
> are planning to do with it--open-source projects need not apply. I've
> tried several channels and have not gotten anywhere.
>
> There apparently is a PLI only interface into the SWIFT model that does
> not use the SWIFT interface. Its called LMTV (appendix B of Simulator
> Configuration Guide for Synopsys Models--simcfg.pdf supplied with any
> SWIFT product). I haven't had time to fully track this down but this
> would be a good place to start--if anyone gets any more leads, please give
> a hollar.
>
> Paul
>
> >
> > Does the Iverilog compiler support SWIFT models?? Anyone have any
> > experience with this??
> > Thanks
> > Lonnie
> >
>
--
--------------------------------------------------
Mike Jarabek
FPGA/ASIC Designer
http://www.istop.com/~mjarabek
--------------------------------------------------