Re: [UrJTAG-dev] data/MANUFACTURERS: Closer to JEP106
On Sat, Aug 03, 2019 at 09:17:48PM +0200, Karl wrote: > Geert Stappers: > > More JEP106 information into MANUFACTURERS file. > > It is based upon mailinglist postings in week 31 of 2019. > ... > > Here is the full generated file, though with a 0 in the directory > column. > 001 0 AMD > 010 0 AMI > 011 0 Fairchild > 100 0 Fujitsu > 101 0 GTE > 110 0 Harris > 111 0 Hitachi > 0001000 0 Inmos > 0001001 0 Intel > 0001010 0 I.T.T. > 0001011 0 Intersil > 0001100 0 Monolithic Memories > 0001101 0 Mostek > 0001110 0 Freescale (Motorola) > 000 0 National > 001 0 NEC > 0010001 0 RCA And who converts that into something to be processed by `git am`? Regards Geert Stappers -- It is *not my* project, it is *our* project ___ UrJTAG-development mailing list UrJTAG-development@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/urjtag-development
Re: [UrJTAG-dev] data/MANUFACTURERS updates
On Thursday, August 8, 2019, Acknowledge wrote: > > > > Then manufaturer (the full name) is only used in presentation, like > > in print chain. > > > > There is no way that we can know that this chip is from that year when > > that company owned the idcode. > > > > Why not just take current owner of the id as shown in (latest) jep106 > > and be done with it ? > > > We agree on that > * anybody should provide a patch > * it is OK if somebody does I am happy someone has worked on it. It is really annoying to NOT have the feature as an end user. Keep pushing! -- Benjamin Henrion (zoobab) Email: zoobab at gmail.com Mobile: +32-484-566109 Web: http://www.zoobab.com FFII.org Brussels "In July 2005, after several failed attempts to legalise software patents in Europe, the patent establishment changed its strategy. Instead of explicitly seeking to sanction the patentability of software, they are now seeking to create a central European patent court, which would establish and enforce patentability rules in their favor, without any possibility of correction by competing courts or democratically elected legislators." ___ UrJTAG-development mailing list UrJTAG-development@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/urjtag-development
Re: [UrJTAG-dev] data/MANUFACTURERS updates
> > Then manufaturer (the full name) is only used in presentation, like > in print chain. > > There is no way that we can know that this chip is from that year when > that company owned the idcode. > > Why not just take current owner of the id as shown in (latest) jep106 > and be done with it ? We agree on that * anybody should provide a patch * it is OK if somebody does ___ UrJTAG-development mailing list UrJTAG-development@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/urjtag-development
Re: [UrJTAG-dev] data/MANUFACTURERS, second column
> On Fri, Aug 02, 2019 at 10:14:28AM +0200, k...@aspodata.se wrote: ... > > Ok, how do you want to handle this then ? > With data/0/PARTS as intermediate step. That seems to be a good idea. > > To manually update the directory column is a pain that I want to avoid. > > Pain is a human thing. Got me there. > Yes, updating the directory column will require human attention. > Luckly there is no need for a bulk update that might wear down humans. ... /// Regarding the mapping idcode -> manufacturer JEP106AZ lists 75 comany name changes, e.g. Siemens -> Infineon Motorola -> Freescale Philips Semi (Signetics) -> NXP Toshiba -> Toshiba Memory Corporation and we have already seen that urjtag disagrees with jep106az in my mail from aug 1 (the first one in [1]). [1] https://sourceforge.net/p/urjtag/mailman/urjtag-development/?viewmonth=201908=1 Then manufaturer (the full name) is only used in presentation, like in print chain. There is no way that we can know that this chip is from that year when that company owned the idcode. Why not just take current owner of the id as shown in (latest) jep106 and be done with it ? /// And regarding idcode -> part If we are lucky, there is just one uniqe chip part (as in register setup, set of instructions, signal names, bsr use) that matches bits 27-12 of the idcode. Currently I have four devices matching the same idcode: $ grep -l 01100110 * STM32F1_Low_Med_density_value_LQFP100.bsd STM32F1_Low_Med_density_value_LQFP48.bsd STM32F1_Low_Med_density_value_LQFP64.bsd STM32F1_Low_Med_density_value_TFBGA64.bsd that has different pin map and boundary register setup. This cannot be expressed in the current data/// setup. So why bother ? /// It would be more efficient to have something like data/// and then the user has to tell the program which of theese files matches your chip (in this position of the jtag chain). Why bsdl files: . always use data from as close to the source as possiblea . they are what the manufacturer publishes . current datafile setup is just a cache of the above . if you miss a file, you can get it from mfg. or from e.g. http://www.bsdl.info/index.htm Regards, /Karl Hammar ___ UrJTAG-development mailing list UrJTAG-development@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/urjtag-development
Re: [UrJTAG-dev] data/MANUFACTURERS, second column
On Fri, Aug 02, 2019 at 10:14:28AM +0200, k...@aspodata.se wrote: > Geert Stappers: > > On Thu, Aug 01, 2019 at 12:07:23PM +0200, k...@aspodata.se wrote: > > > So making a data/MANUFACTURERS file like: > > > $ head MANUFACTURERS > > > 001 001 AMD > > > 010 010 AMI > > > 011 011 Fairchild > > > 100 100 Fujitsu > > > 101 101 GTE > > > 110 110 Harris > > > 111 111 Hitachi > > > 0001001 0001001 Intel > > > > > > and making theese links (in data directory): > > > 110 -> lexra > > > 111 -> hitachi > > > 0001001 -> intel/ > > > 0001110 -> freescale/ > > > 0010101 -> philips > > > > IMNSHO is that only moving the problem to another place. > > The symlinks add clutter. > > > > It is wrong approach of the challenge of updating data/MANUFACTERS > > Ok, how do you want to handle this then ? With data/0/PARTS as intermediate step. > To manually update the directory column is a pain that I want to avoid. Pain is a human thing. Yes, updating the directory column will require human attention. Luckly there is no need for a bulk update that might wear down humans. > BTW, I sent you a git format-patch. Regardless if that patch is > accepted or not, is that the way you want possible changes presented ? Yes. Or another form that contains code change plus commit message that can be processed automaticly. Groeten Geert Stappers -- Leven en laten leven ___ UrJTAG-development mailing list UrJTAG-development@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/urjtag-development
Re: [UrJTAG-dev] data/MANUFACTURERS
On Mon, Jul 29, 2019 at 11:22:18PM +0200, k...@aspodata.se wrote: >... > But what is the best way to update the manufacturers list ? Small steps ___ UrJTAG-development mailing list UrJTAG-development@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/urjtag-development
Re: [UrJTAG-dev] data/MANUFACTURERS, second column
Geert Stappers: > On Thu, Aug 01, 2019 at 12:07:23PM +0200, k...@aspodata.se wrote: > > Karl: > > > On Wed, Jul 31, 2019 at 10:44:45PM +0200, k...@aspodata.se wrote: > > > > The urjtag file has three columns: > > > > $ grep Hitachi data/MANUFACTURERS > > > > 111 hitachi Hitachi > > > > > and mine only has two: > > > > > $ ./jep106 | grep Hitachi > > > > > 111 Hitachi > > > e.g. data/analog/PARTS. > > > One way to handle this is to use the id instead of id_name for the > > > parts directories: e.g. instead of data/analog/PARTS, one coulde use > > > data/1100101/PARTS. But that makes it harder to browse the file > > > system. > > > > So making a data/MANUFACTURERS file like: > > $ head MANUFACTURERS > > 001 001 AMD > > 010 010 AMI > > 011 011 Fairchild > > 100 100 Fujitsu > > 101 101 GTE > > 110 110 Harris > > 111 111 Hitachi > > 0001000 0001000 Inmos > > 0001001 0001001 Intel > > 0001010 0001010 I.T.T. > > > > and making theese links (in data directory): > > 110 -> lexra > > 111 -> hitachi > > 0001001 -> intel/ > > 0001110 -> freescale/ > > 0010101 -> philips > > 0010111 -> ti > > 0011000 -> toshiba > > 001 -> atmel > > 011 -> lattice > > 0100100 -> ibm > > 0110101 -> dec > > 1001001 -> xilinx > > 1100101 -> analog > > IMNSHO is that only moving the problem to another place. > The symlinks add clutter. > > It is wrong approach of the challenge of updating data/MANUFACTERS Ok, how do you want to handle this then ? To manually update the directory column is a pain that I want to avoid. BTW, I sent you a git format-patch. Regardless if that patch is accepted or not, is that the way you want possible changes presented ? Regards, /Karl Hammar ___ UrJTAG-development mailing list UrJTAG-development@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/urjtag-development
Re: [UrJTAG-dev] data/MANUFACTURERS, second column
On Thu, Aug 01, 2019 at 12:07:23PM +0200, k...@aspodata.se wrote: > Karl: > > On Wed, Jul 31, 2019 at 10:44:45PM +0200, k...@aspodata.se wrote: > > > The urjtag file has three columns: > > > $ grep Hitachi data/MANUFACTURERS > > > 111 hitachi Hitachi > > > > and mine only has two: > > > > $ ./jep106 | grep Hitachi > > > > 111 Hitachi > > e.g. data/analog/PARTS. > > One way to handle this is to use the id instead of id_name for the > > parts directories: e.g. instead of data/analog/PARTS, one coulde use > > data/1100101/PARTS. But that makes it harder to browse the file > > system. > > So making a data/MANUFACTURERS file like: > $ head MANUFACTURERS > 001 001 AMD > 010 010 AMI > 011 011 Fairchild > 100 100 Fujitsu > 101 101 GTE > 110 110 Harris > 111 111 Hitachi > 0001000 0001000 Inmos > 0001001 0001001 Intel > 0001010 0001010 I.T.T. > > and making theese links (in data directory): > 110 -> lexra > 111 -> hitachi > 0001001 -> intel/ > 0001110 -> freescale/ > 0010101 -> philips > 0010111 -> ti > 0011000 -> toshiba > 001 -> atmel > 011 -> lattice > 0100100 -> ibm > 0110101 -> dec > 1001001 -> xilinx > 1100101 -> analog IMNSHO is that only moving the problem to another place. The symlinks add clutter. It is wrong approach of the challenge of updating data/MANUFACTERS Groeten Geert Stappers -- Leven en laten leven ___ UrJTAG-development mailing list UrJTAG-development@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/urjtag-development
Re: [UrJTAG-dev] data/MANUFACTURERS
On Thu, Aug 01, 2019 at 11:51:11PM +0200, k...@aspodata.se wrote: > Karl: > > /// > > There are some mismatches between the new set (below) and the one in > > data/MANUFACTURERS (above): > > > > 110 lexra Lexr > > 110 Harris > > > > Lexra went out of business 2003 and Harris is now running under a new > > name. > > > > 00010101011 lattice Lattice Semiconductors > > 00010101011 Vantis > > > > Lattice bought Vantis from AMD in 1999. > > > > 00011100101 analog Analog Devices, Inc. > > 00011100101 Gadzoox Networks > > > > Gadzoox went bancrupt in 2002, and Broadcom bought its assets. > > > > 0010101 broadcomBroadcom# or "Sibyte, Incorporated" > > ? > > 0010101 Sibyte, Incorporated > > > > Broadcom bought Sibyte in 2000. > > > > 00110101011 marvell Marvell > > 00110101011 Purple Ray > > > > Purple Ray was aquired by Integrated Silicon Solution 2004. > > Geert Stappers, in another email: > > Updating data/MANUFACTURERS is somewhat higher on my full list ... > ... > > Ok, fine with me. > > Did you like my proposal above or do you want to do it in another way ? As I don't understand the proposal of Karl it is hard to tell if my changes are along the same way. Unified patches such as generated by `git format-patch` are a clear way to express a change. Groeten Geert Stappers -- Leven en laten leven ___ UrJTAG-development mailing list UrJTAG-development@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/urjtag-development
Re: [UrJTAG-dev] data/MANUFACTURERS
Karl: > Karl: ... > > One way to handle this is to use the id instead of id_name for the > > parts directories: e.g. instead of data/analog/PARTS, one coulde use > > data/1100101/PARTS. But that makes it harder to browse the file > > system. > > So making a data/MANUFACTURERS file like: > $ head MANUFACTURERS > 001 001 AMD > 010 010 AMI > 011 011 Fairchild > 100 100 Fujitsu > 101 101 GTE > 110 110 Harris > 111 111 Hitachi > 0001000 0001000 Inmos > 0001001 0001001 Intel > 0001010 0001010 I.T.T. > > and making theese links (in data directory): > 110 -> lexra > 111 -> hitachi > 0001001 -> intel/ > 0001110 -> freescale/ > 0010101 -> philips > 0010111 -> ti > 0011000 -> toshiba > 001 -> atmel > 011 -> lattice > 0100100 -> ibm > 0110101 -> dec > 1001001 -> xilinx > 1100101 -> analog > 1101110 -> altera > 00010101011 -> lattice > 0001011 -> broadcom > 00011100101 -> analog > 0010101 -> broadcom > 0010111 -> brecis/ > 0001001 -> marvell > > I get: ... > jtag> detect > IR length: 9 > Chain length: 2 > Device Id: 001110111010010001110111 (0x3BA00477) > Manufacturer: ARM Ltd. (0x477) > Unknown part! (10111010) (/usr/local/share/urjtag/01000111011/PARTS) > Device Id: 011001100101 (0x06420041) > Manufacturer: STMicroelectronics (0x041) > Unknown part! (01100110) (/usr/local/share/urjtag/010/PARTS) > > So that works. > > /// > There are some mismatches between the new set (below) and the one in > data/MANUFACTURERS (above): > > 110 lexra Lexr > 110 Harris > > Lexra went out of business 2003 and Harris is now running under a new > name. > > 00010101011 lattice Lattice Semiconductors > 00010101011 Vantis > > Lattice bought Vantis from AMD in 1999. > > 00011100101 analog Analog Devices, Inc. > 00011100101 Gadzoox Networks > > Gadzoox went bancrupt in 2002, and Broadcom bought its assets. > > 0010101 broadcomBroadcom# or "Sibyte, Incorporated" ? > 0010101 Sibyte, Incorporated > > Broadcom bought Sibyte in 2000. > > 00110101011 marvell Marvell > 00110101011 Purple Ray > > Purple Ray was aquired by Integrated Silicon Solution 2004. Geert Stappers: > Updating data/MANUFACTURERS is somewhat higher on my full list ... ... Ok, fine with me. Did you like my proposal above or do you want to do it in another way ? BTW, as the code is written in usr/tap/detect.c, if you have entered bsdl some path and detect finds a bsdl, the data/MANUFACTURERS isn't consulted. Regards, /Karl Hammar ___ UrJTAG-development mailing list UrJTAG-development@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/urjtag-development
Re: [UrJTAG-dev] data/MANUFACTURERS
On Thu, Aug 01, 2019 at 09:59:32PM +0200, k...@aspodata.se wrote: > > Ok, fixed it with attached patch: > Patch lacks commit message. And make it at least two seperate commits. One with "jep106", the other with "the fix". Regards Geert Stappers P.S. Updating data/MANUFACTURERS is somewhat higher on my full list ... ___ UrJTAG-development mailing list UrJTAG-development@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/urjtag-development
Re: [UrJTAG-dev] data/MANUFACTURERS
Karl:
...
> Perhaps it is best to just drop the data directory and start collecting
> bsdl files instead.
>
> With bsdl files, I get:
>
> jtag> detect
> IR length: 9
> Chain length: 2
> Device Id: 001110111010010001110111 (0x3BA00477)
> Filename: ./CortexM3.bsd
> Device Id: 011001100101 (0x06420041)
> Filename: ./STM32F1_Low_Med_density_value_LQFP48.bsd
>
> Could be nice to have the manufacturers line there also.
Ok, fixed it with attached patch:
jtag> detect
IR length: 9
Chain length: 2
Device Id: 001110111010010001110111 (0x3BA00477)
Filename: ./CortexM3.bsd
Device Id: 011001100101 (0x06420041)
Filename: ./STM32F1_Low_Med_density_value_LQFP48.bsd
jtag> print chain
No. Manufacturer Part Stepping Instruction
Register
---
0 ARM Ltd. CortexM3 0x03 BYPASS
BYPASS
* 1 STMicroelectronicsSTM32F1_Low_Med_dens 0x00 BYPASS
BYPASS
Now manf. and stepping is available.
The files goes to:
include/urjtag/jep106.h
src/global/jep106.c
src/global/jep106.inc
Regards,
/Karl Hammar
diff --git a/urjtag/src/global/Makefile.am b/urjtag/src/global/Makefile.am
index 3fa82f0b..3bcf5ece 100644
--- a/urjtag/src/global/Makefile.am
+++ b/urjtag/src/global/Makefile.am
@@ -26,6 +26,7 @@ include $(top_srcdir)/Makefile.rules
noinst_LTLIBRARIES = libglobal.la
libglobal_la_SOURCES = \
+ jep106.c \
parse.c \
log-error.c \
data_dir.c \
diff --git a/urjtag/src/tap/detect.c b/urjtag/src/tap/detect.c
index 5c139507..ddb84508 100644
--- a/urjtag/src/tap/detect.c
+++ b/urjtag/src/tap/detect.c
@@ -45,6 +45,7 @@
#include
#include
#include
+#include
static int
find_record (char *filename, urj_tap_register_t *key,
@@ -315,7 +316,31 @@ urj_tap_detect_parts (urj_chain_t *chain, const char *db_path, int maxirlen)
#ifdef ENABLE_BSDL
if (urj_bsdl_scan_files (chain, urj_tap_register_get_string (did),
- URJ_BSDL_MODE_DETECT) <= 0)
+ URJ_BSDL_MODE_DETECT) > 0)
+ {
+ uint64_t idcode = urj_tap_register_get_value (did);
+ unsigned int Mfg = (unsigned int) EXTRACT_MFG(idcode);
+ unsigned int Bank = (unsigned int) EXTRACT_JEP106_BANK(idcode);
+ unsigned int Id = (unsigned int) EXTRACT_JEP106_ID(idcode);
+ unsigned int Part = (unsigned int) EXTRACT_PART(idcode);
+ unsigned int Ver = (unsigned int) EXTRACT_VER(idcode);
+
+ if (part->manufacturer[0] == '\0') {
+ const char *str = jep106_manufacturer(Bank, Id);
+ if (str == NULL) {
+ snprintf(part->manufacturer, URJ_PART_MANUFACTURER_MAXLEN + 1, "0x%02x", Mfg);
+ } else {
+ snprintf(part->manufacturer, URJ_PART_MANUFACTURER_MAXLEN + 1, "%s", str);
+ }
+ }
+ if (part->part[0] == '\0') {
+ snprintf(part->part, URJ_PART_PART_MAXLEN + 1, "0x%02x", Part);
+ }
+ if (part->stepping[0] == '\0') {
+ snprintf(part->stepping, URJ_PART_STEPPING_MAXLEN + 1, "0x%02x", Ver);
+ }
+ }
+ else
#endif
{
char *id_name = NULL, *id_fullname = NULL;
/* Autogenerated with update_jep106.pl*/
[0][0x01 - 1] = "AMD",
[0][0x02 - 1] = "AMI",
[0][0x03 - 1] = "Fairchild",
[0][0x04 - 1] = "Fujitsu",
[0][0x05 - 1] = "GTE",
[0][0x06 - 1] = "Harris",
[0][0x07 - 1] = "Hitachi",
[0][0x08 - 1] = "Inmos",
[0][0x09 - 1] = "Intel",
[0][0x0a - 1] = "I.T.T.",
[0][0x0b - 1] = "Intersil",
[0][0x0c - 1] = "Monolithic Memories",
[0][0x0d - 1] = "Mostek",
[0][0x0e - 1] = "Freescale (Motorola)",
[0][0x0f - 1] = "National",
[0][0x10 - 1] = "NEC",
[0][0x11 - 1] = "RCA",
[0][0x12 - 1] = "Raytheon",
[0][0x13 - 1] = "Conexant (Rockwell)",
[0][0x14 - 1] = "Seeq",
[0][0x15 - 1] = "NXP (Philips)",
[0][0x16 - 1] = "Synertek",
[0][0x17 - 1] = "Texas Instruments",
[0][0x18 - 1] = "Toshiba",
[0][0x19 - 1] = "Xicor",
[0][0x1a - 1] = "Zilog",
[0][0x1b - 1] = "Eurotechnique",
[0][0x1c - 1] = "Mitsubishi",
[0][0x1d - 1] = "Lucent (AT)",
[0][0x1e - 1] = "Exel",
[0][0x1f - 1] = "Atmel",
[0][0x20 - 1] = "STMicroelectronics",
[0][0x21 - 1] = "Lattice Semi.",
[0][0x22 - 1] = "NCR",
[0][0x23 - 1] = "Wafer Scale Integration",
[0][0x24 - 1] = "IBM",
[0][0x25 - 1] = "Tristar",
[0][0x26 - 1] = "Visic",
[0][0x27 - 1] = "Intl. CMOS Technology",
[0][0x28 - 1] = "SSSI",
[0][0x29 - 1] = "MicrochipTechnology",
[0][0x2a - 1] = "Ricoh Ltd.",
[0][0x2b - 1] = "VLSI",
[0][0x2c - 1] = "Micron Technology",
[0][0x2d - 1] = "SK Hynix",
[0][0x2e - 1] = "OKI Semiconductor",
[0][0x2f - 1] = "ACTEL",
[0][0x30 - 1] = "Sharp",
[0][0x31 - 1] = "Catalyst",
[0][0x32 - 1] = "Panasonic",
[0][0x33 - 1] = "IDT",
[0][0x34 - 1] = "Cypress",
[0][0x35 - 1] = "DEC",
[0][0x36 - 1] = "LSI Logic",
[0][0x37 - 1] = "Zarlink
Re: [UrJTAG-dev] data/MANUFACTURERS
Karl: > G G: > > On Wed, Jul 31, 2019 at 10:44:45PM +0200, k...@aspodata.se wrote: > ... > > > The urjtag file has three columns: > > > $ grep Hitachi data/MANUFACTURERS > > > 111 hitachi Hitachi > > > and mine only has two: > > > $ ./jep106 | grep Hitachi > > > 111 Hitachi ... > e.g. data/analog/PARTS. > One way to handle this is to use the id instead of id_name for the > parts directories: e.g. instead of data/analog/PARTS, one coulde use > data/1100101/PARTS. But that makes it harder to browse the file > system. So making a data/MANUFACTURERS file like: $ head MANUFACTURERS 001 001 AMD 010 010 AMI 011 011 Fairchild 100 100 Fujitsu 101 101 GTE 110 110 Harris 111 111 Hitachi 0001000 0001000 Inmos 0001001 0001001 Intel 0001010 0001010 I.T.T. and making theese links (in data directory): 110 -> lexra 111 -> hitachi 0001001 -> intel/ 0001110 -> freescale/ 0010101 -> philips 0010111 -> ti 0011000 -> toshiba 001 -> atmel 011 -> lattice 0100100 -> ibm 0110101 -> dec 1001001 -> xilinx 1100101 -> analog 1101110 -> altera 00010101011 -> lattice 0001011 -> broadcom 00011100101 -> analog 0010101 -> broadcom 0010111 -> brecis/ 0001001 -> marvell I get: jtag> detect IR length: 9 Chain length: 2 Device Id: 001110111010010001110111 (0x3BA00477) Manufacturer: ARM Ltd. (0x477) error: Unable to open file '/usr/local/share/urjtag/01000111011/PARTS' Unknown part! (10111010) (/usr/local/share/urjtag/01000111011/PARTS) Device Id: 011001100101 (0x06420041) Manufacturer: STMicroelectronics (0x041) error: Unable to open file '/usr/local/share/urjtag/010/PARTS' Unknown part! (01100110) (/usr/local/share/urjtag/010/PARTS) Creating directories and empty PARTS files, I get rid of the errors: jtag> detect IR length: 9 Chain length: 2 Device Id: 001110111010010001110111 (0x3BA00477) Manufacturer: ARM Ltd. (0x477) Unknown part! (10111010) (/usr/local/share/urjtag/01000111011/PARTS) Device Id: 011001100101 (0x06420041) Manufacturer: STMicroelectronics (0x041) Unknown part! (01100110) (/usr/local/share/urjtag/010/PARTS) So that works. /// There are some mismatches between the new set (below) and the one in data/MANUFACTURERS (above): 110 lexra Lexr 110 Harris Lexra went out of business 2003 and Harris is now running under a new name. 00010101011 lattice Lattice Semiconductors 00010101011 Vantis Lattice bought Vantis from AMD in 1999. 00011100101 analog Analog Devices, Inc. 00011100101 Gadzoox Networks Gadzoox went bancrupt in 2002, and Broadcom bought its assets. 0010101 broadcomBroadcom# or "Sibyte, Incorporated" ? 0010101 Sibyte, Incorporated Broadcom bought Sibyte in 2000. 00110101011 marvell Marvell 00110101011 Purple Ray Purple Ray was aquired by Integrated Silicon Solution 2004. /// Perhaps it is best to just drop the data directory and start collecting bsdl files instead. With bsdl files, I get: jtag> detect IR length: 9 Chain length: 2 Device Id: 001110111010010001110111 (0x3BA00477) Filename: ./CortexM3.bsd Device Id: 011001100101 (0x06420041) Filename: ./STM32F1_Low_Med_density_value_LQFP48.bsd Could be nice to have the manufacturers line there also. Regards, /Karl Hammar ___ UrJTAG-development mailing list UrJTAG-development@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/urjtag-development
Re: [UrJTAG-dev] data/MANUFACTURERS
G G:
> On Wed, Jul 31, 2019 at 10:44:45PM +0200, k...@aspodata.se wrote:
...
> > The urjtag file has three columns:
> >
> > $ grep Hitachi data/MANUFACTURERS
> > 111 hitachi Hitachi
> >
> > and mine only has two:
> >
> > $ ./jep106 | grep Hitachi
> > 111 Hitachi
...
> > Any idea how to fix this ?
>
> | 10010001011 Paragon Technology (Shenzhen) Ltd.
> | 10010001101 Shenzhen Yong Sheng Technology
>
> 10010001011 Mx48B Paragon_Technology__Shenzhen__Ltd_
> 10010001101 Mx48D Shenzhen_Yong_Sheng_Technology
I don't think that would do.
urj_tap_detect_parts()/find_record() in src/tap/detect.c are the
only ones reading the MANUFACTURERS file. find_record() finds the
line matching the id, so the id (key) must be the first column:
line 333:
if (!find_record (data_path, key, _name, _fullname))
...
urj_log (URJ_LOG_LEVEL_NORMAL, " %12s: %s (0x%03"PRIX64")\n",
_("Manufacturer"), id_fullname,
(urj_tap_register_get_value (key) << 1) | 1);
The third column is used in the logging output, but the second col.
is used to find the PARTS files:
line 388:
strncat_const (data_path, id_name);
strncat_const (data_path, "/PARTS");
e.g. data/analog/PARTS.
One way to handle this is to use the id instead of id_name for the
parts directories: e.g. instead of data/analog/PARTS, one coulde use
data/1100101/PARTS. But that makes it harder to browse the file
system.
Regards,
/Karl Hammar
___
UrJTAG-development mailing list
UrJTAG-development@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/urjtag-development
Re: [UrJTAG-dev] data/MANUFACTURERS
On Wed, Jul 31, 2019 at 10:44:45PM +0200, k...@aspodata.se wrote: > Karl: > > But what is the best way to update the manufacturers list ? > > > > Jedec uses bank, and id within bank. How do I convert that to the > > format used in data/MANUFACTURERS ? > ... > > I come up the the attached code. :-) > The urjtag file has three columns: > > $ grep Hitachi data/MANUFACTURERS > 111 hitachi Hitachi > > and mine only has two: > > $ ./jep106 | grep Hitachi > 111 Hitachi > > I don't know of any good way to produce the middle column > from my last column. Just picking the first token isn't good: > > $ ./jep106 | grep Shenzhen > 1000111 Shenzhen Jinge Information Co. Ltd. > 101 Shenzhen City Gcai Electronics > 1100100 Shenzhen Elicks Technology > 10001001001 Shenzhen Zhongteng Electronic Corp. Ltd. > 10001001100 Shenzhen Pango Microsystems Co. Ltd > 10001100111 Shenzhen Mic Electronics Technology > 10010001011 Paragon Technology (Shenzhen) Ltd. > 10010001101 Shenzhen Yong Sheng Technology > 10010001110 SNOAMOO (Shenzhen Kai Zhuo Yue) > 1001001 Shenzhen XinRuiYan Electronics > 10010010100 Shenzhen Chixingzhe Tech Co. Ltd. > > Any idea how to fix this ? | 10010001011 Paragon Technology (Shenzhen) Ltd. | 10010001101 Shenzhen Yong Sheng Technology 10010001011 Mx48B Paragon_Technology__Shenzhen__Ltd_ 10010001101 Mx48D Shenzhen_Yong_Sheng_Technology > Regards, > /Karl Hammar G G ___ UrJTAG-development mailing list UrJTAG-development@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/urjtag-development
Re: [UrJTAG-dev] data/MANUFACTURERS
Karl: > But what is the best way to update the manufacturers list ? > > Jedec uses bank, and id within bank. How do I convert that to the > format used in data/MANUFACTURERS ? ... I come up the the attached code. The urjtag file has three columns: $ grep Hitachi data/MANUFACTURERS 111 hitachi Hitachi and mine only has two: $ ./jep106 | grep Hitachi 111 Hitachi I don't know of any good way to produce the middle column from my last column. Just picking the first token isn't good: $ ./jep106 | grep Shenzhen 1000111 Shenzhen Jinge Information Co. Ltd. 101 Shenzhen City Gcai Electronics 1100100 Shenzhen Elicks Technology 10001001001 Shenzhen Zhongteng Electronic Corp. Ltd. 10001001100 Shenzhen Pango Microsystems Co. Ltd 10001100111 Shenzhen Mic Electronics Technology 10010001011 Paragon Technology (Shenzhen) Ltd. 10010001101 Shenzhen Yong Sheng Technology 10010001110 SNOAMOO (Shenzhen Kai Zhuo Yue) 1001001 Shenzhen XinRuiYan Electronics 10010010100 Shenzhen Chixingzhe Tech Co. Ltd. Any idea how to fix this ? Regards, /Karl Hammar /* Autogenerated with update_jep106.pl*/ [0][0x01 - 1] = "AMD", [0][0x02 - 1] = "AMI", [0][0x03 - 1] = "Fairchild", [0][0x04 - 1] = "Fujitsu", [0][0x05 - 1] = "GTE", [0][0x06 - 1] = "Harris", [0][0x07 - 1] = "Hitachi", [0][0x08 - 1] = "Inmos", [0][0x09 - 1] = "Intel", [0][0x0a - 1] = "I.T.T.", [0][0x0b - 1] = "Intersil", [0][0x0c - 1] = "Monolithic Memories", [0][0x0d - 1] = "Mostek", [0][0x0e - 1] = "Freescale (Motorola)", [0][0x0f - 1] = "National", [0][0x10 - 1] = "NEC", [0][0x11 - 1] = "RCA", [0][0x12 - 1] = "Raytheon", [0][0x13 - 1] = "Conexant (Rockwell)", [0][0x14 - 1] = "Seeq", [0][0x15 - 1] = "NXP (Philips)", [0][0x16 - 1] = "Synertek", [0][0x17 - 1] = "Texas Instruments", [0][0x18 - 1] = "Toshiba", [0][0x19 - 1] = "Xicor", [0][0x1a - 1] = "Zilog", [0][0x1b - 1] = "Eurotechnique", [0][0x1c - 1] = "Mitsubishi", [0][0x1d - 1] = "Lucent (AT)", [0][0x1e - 1] = "Exel", [0][0x1f - 1] = "Atmel", [0][0x20 - 1] = "STMicroelectronics", [0][0x21 - 1] = "Lattice Semi.", [0][0x22 - 1] = "NCR", [0][0x23 - 1] = "Wafer Scale Integration", [0][0x24 - 1] = "IBM", [0][0x25 - 1] = "Tristar", [0][0x26 - 1] = "Visic", [0][0x27 - 1] = "Intl. CMOS Technology", [0][0x28 - 1] = "SSSI", [0][0x29 - 1] = "MicrochipTechnology", [0][0x2a - 1] = "Ricoh Ltd.", [0][0x2b - 1] = "VLSI", [0][0x2c - 1] = "Micron Technology", [0][0x2d - 1] = "SK Hynix", [0][0x2e - 1] = "OKI Semiconductor", [0][0x2f - 1] = "ACTEL", [0][0x30 - 1] = "Sharp", [0][0x31 - 1] = "Catalyst", [0][0x32 - 1] = "Panasonic", [0][0x33 - 1] = "IDT", [0][0x34 - 1] = "Cypress", [0][0x35 - 1] = "DEC", [0][0x36 - 1] = "LSI Logic", [0][0x37 - 1] = "Zarlink (Plessey)", [0][0x38 - 1] = "UTMC", [0][0x39 - 1] = "Thinking Machine", [0][0x3a - 1] = "Thomson CSF", [0][0x3b - 1] = "Integrated CMOS (Vertex)", [0][0x3c - 1] = "Honeywell", [0][0x3d - 1] = "Tektronix", [0][0x3e - 1] = "Oracle Corporation", [0][0x3f - 1] = "Silicon Storage Technology", [0][0x40 - 1] = "ProMos/Mosel Vitelic", [0][0x41 - 1] = "Infineon (Siemens)", [0][0x42 - 1] = "Macronix", [0][0x43 - 1] = "Xerox", [0][0x44 - 1] = "Plus Logic", [0][0x45 - 1] = "SanDisk Corporation", [0][0x46 - 1] = "Elan Circuit Tech.", [0][0x47 - 1] = "European Silicon Str.", [0][0x48 - 1] = "Apple Computer", [0][0x49 - 1] = "Xilinx", [0][0x4a - 1] = "Compaq", [0][0x4b - 1] = "Protocol Engines", [0][0x4c - 1] = "SCI", [0][0x4d - 1] = "Seiko Instruments", [0][0x4e - 1] = "Samsung", [0][0x4f - 1] = "I3 Design System", [0][0x50 - 1] = "Klic", [0][0x51 - 1] = "Crosspoint Solutions", [0][0x52 - 1] = "Alliance Semiconductor", [0][0x53 - 1] = "Tandem", [0][0x54 - 1] = "Hewlett-Packard", [0][0x55 - 1] = "Integrated Silicon Solutions", [0][0x56 - 1] = "Brooktree", [0][0x57 - 1] = "New Media", [0][0x58 - 1] = "MHS Electronic", [0][0x59 - 1] = "Performance Semi.", [0][0x5a - 1] = "Winbond Electronic", [0][0x5b - 1] = "Kawasaki Steel", [0][0x5c - 1] = "Bright Micro", [0][0x5d - 1] = "TECMAR", [0][0x5e - 1] = "Exar", [0][0x5f - 1] = "PCMCIA", [0][0x60 - 1] = "LG Semi (Goldstar)", [0][0x61 - 1] = "Northern Telecom", [0][0x62 - 1] = "Sanyo", [0][0x63 - 1] = "Array Microsystems", [0][0x64 - 1] = "Crystal Semiconductor", [0][0x65 - 1] = "Analog Devices", [0][0x66 - 1] = "PMC-Sierra", [0][0x67 - 1] = "Asparix", [0][0x68 - 1] = "Convex Computer", [0][0x69 - 1] = "Quality Semiconductor", [0][0x6a - 1] = "Nimbus Technology", [0][0x6b - 1] = "Transwitch", [0][0x6c - 1] = "Micronas (ITT Intermetall)", [0][0x6d - 1] = "Cannon", [0][0x6e - 1] = "Altera", [0][0x6f - 1] = "NEXCOM", [0][0x70 - 1] = "Qualcomm", [0][0x71 - 1] = "Sony", [0][0x72 - 1] = "Cray Research", [0][0x73 - 1] = "AMS(Austria Micro)", [0][0x74 - 1] = "Vitesse", [0][0x75 - 1] = "Aster Electronics", [0][0x76 - 1] = "Bay Networks (Synoptic)", [0][0x77 - 1] = "Zentrum/ZMD", [0][0x78 - 1] = "TRW", [0][0x79 - 1] = "Thesys", [0][0x7a - 1] = "Solbourne Computer", [0][0x7b -
[UrJTAG-dev] data/MANUFACTURERS
As seen below, my driver seems to work. But what is the best way to update the manufacturers list ? Jedec uses bank, and id within bank. How do I convert that to the format used in data/MANUFACTURERS ? /// You can download the list from: http://www.jedec.org/standards-documents/results/jep106 in pdf format. Openocd uses https://sourceforge.net/p/openocd/code/ci/master/tree/src/helper/update_jep106.pl to extract the ids into c code and their latest list is available at: https://sourceforge.net/p/openocd/code/ci/master/tree/src/helper/jep106.inc /// $ openocd ... Info : JTAG tap: stm32f1x.cpu tap/device found: 0x3ba00477 (mfg: 0x23b (ARM Ltd.), part: 0xba00, ver: 0x3) Info : JTAG tap: stm32f1x.bs tap/device found: 0x06420041 (mfg: 0x020 (STMicroelectronics), part: 0x6420, ver: 0x0) Info : stm32f1x.cpu: hardware has 6 breakpoints, 4 watchpoints ^C $ jtag -q jtag> cable ASPO ppdev /dev/parport0 Initializing ppdev port /dev/parport0 jtag> frequency 10 ... jtag> detect IR length: 9 Chain length: 2 Device Id: 001110111010010001110111 (0x3BA00477) Unknown manufacturer! (01000111011) (/usr/local/share/urjtag/MANUFACTURERS) Device Id: 011001100101 (0x06420041) Unknown manufacturer! (010) (/usr/local/share/urjtag/MANUFACTURERS) jtag> quit free(): invalid size Aborted Regards, /Karl Hammar ___ UrJTAG-development mailing list UrJTAG-development@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/urjtag-development
