00:20.27 | *** join/#edev RobotGuy (~n7pkt@c-24-21-60-36.hsd1.or.comcast.net) |
00:44.01 | *** join/#edev SpeedEvil (~user@tor/regular/SpeedEvil) |
00:50.50 | *** join/#edev djerome (~djerome@ip24-251-158-189.ph.ph.cox.net) |
01:41.21 | *** join/#edev SpeedEvil (1000@tor/regular/SpeedEvil) |
02:14.06 | *** join/#edev kergoth (~kergoth@ip24-251-170-95.ph.ph.cox.net) |
03:53.06 | *** join/#edev sn9 (~danielg4@66.218.54.163) |
03:54.35 | sn9 | anyone awake? |
03:55.28 | sn9 | i seem to be in an unfortunate timezone for finding people awake |
04:02.05 | kergoth | it's getting late in the US (e.g. 9pm mountain), and still early in europe (e.g. 4am UTC) |
04:02.09 | kergoth | heh |
04:02.13 | sn9 | exactly |
04:02.22 | sn9 | it's 8pm here |
04:03.16 | sn9 | i don't suppose you have any experience flashing a bootloader on when jtag is not exposed... |
04:06.48 | sn9 | kergoth: ever needed to do that? |
04:08.04 | kergoth | fraid not. its possible you could via the mtd from linux, assuming that flash is unlocked, but it'd be awfully dangerous, I'd guess. but it's been a while since i've had to touch any hardware, stuck in buildsystem land nowadays |
04:08.09 | kergoth | shrugs |
04:08.26 | sn9 | linux requires a working bootloader |
04:08.34 | kergoth | oh, it's not functional? |
04:08.49 | kergoth | heh, no idea |
04:09.00 | sn9 | that's why i need to flash the bootloader on; i wrote to the wrong mtd |
04:09.09 | kergoth | ahh |
04:10.49 | sn9 | i have attempted to get at the jtag signals from the back of the processor itself: http://www.gimpelevich.san-francisco.ca.us/danielg4/got_jtag.jpg |
04:11.26 | kergoth | i'm sure there are guys in here that could help, hardware folk, but as you say, bad timing :\ |
04:11.27 | sn9 | i based it on this diagram: http://oldwiki.openwrt.org/attachments/AR7Port/ar7_jtag.jpg |
04:12.04 | sn9 | the diagram turned out to contain a couple of mistakes |
05:18.56 | solar | bricks ;/ |
05:20.01 | sn9 | yeah, those |
05:22.54 | sn9 | the setup in the pic was very precarious and did not survive redoing once i got the correct chip pinout. 4 of the 7 traces i need are damaged to the point where catching them would require microscopic contacts now |
05:27.03 | sn9 | my question is: what's the least impossible way to put the bootloader on now? |
05:42.01 | dalias | does rob ever come around anymore? |
06:41.20 | *** join/#edev destenson (~destenson@mail.digitalforcetech.com) |
07:09.49 | sn9 | ... |
08:46.57 | davidc__ | sn9: oi; you soldered them to the top of the BGA package? that must have been... fun. |
08:47.04 | davidc__ | I've never actually heard of someone doing that before. |
08:47.24 | davidc__ | and I guess those testpoints right next to the BGA aren't connected to the jtag IOs? |
08:56.17 | dalias | anyone know why modern linux pre-maps 132k of stack? |
08:56.36 | dalias | it makes it really hard to measure bloat at a glance |
08:57.18 | dalias | because 128k or so of that 132k is probably untouched zero pages |
08:58.30 | dalias | for instance i have a process that's 144k of vss, but 132k of that it stack, and only 8k is dirty (1 page of stack, 1 page of data) |
08:59.50 | dalias | and the real vss (minus 128k of useless stack mapping) is 16k |
10:54.07 | *** join/#edev pirho (debian-tor@gateway/tor-sasl/pirho) |
11:22.57 | *** join/#edev Roland (~roland@195.50.208.35) |
11:27.43 | SpeedEvil | sn9: well - you could unsolder and flash the ROM |
11:29.23 | SpeedEvil | sn9: That is ... ingenious. |
11:29.27 | SpeedEvil | (the pic) |
12:24.09 | sn9 | davidc__: no, i did not solder; look at the pic |
12:26.40 | sn9 | and no, the jtag io's are not connected, at all |
12:27.13 | sn9 | (and if they were, they wouldn't be now) |
13:28.26 | *** join/#edev pirho (debian-tor@gateway/tor-sasl/pirho) |
13:52.48 | *** join/#edev GPSFan (~kenm@64.92.145.112) |
13:52.48 | *** mode/#edev [+o GPSFan] by ChanServ |
14:40.15 | sn9 | SpeedEvil: so, is removing the flash chip really the easiest way now? |
15:13.38 | SpeedEvil | Possibly, yes. |
15:14.11 | SpeedEvil | If you can't expose the JTAG bus, or get it to cooperate and load over some other interface, or get a new one on ebay. |
15:22.55 | sn9 | other interface? |
15:23.29 | sn9 | SpeedEvil: like what? |
16:04.45 | SpeedEvil | For example, if the CPU supports a serial bootloader. |
16:05.01 | SpeedEvil | Or if the firmware tries to load over tftp |
16:09.11 | *** join/#edev prpplague (~danders@nat/ti/x-buqcqdyhkezubzzv) |
16:17.28 | sn9 | SpeedEvil: those are functions of the bootloader, which is what i want to put on |
16:21.15 | SpeedEvil | Not always |
16:21.34 | SpeedEvil | for example, the stm32 has a hardware serial bootloader activated by pulling one line high |
16:23.02 | sn9 | wow |
16:23.55 | sn9 | failing that, what are the chances of still using jtag, after all that? like maybe with the help of a conductive ink pen? |
16:32.14 | *** join/#edev pirho (debian-tor@gateway/tor-sasl/pirho) |
16:47.51 | SpeedEvil | I diddn't see exactly the damage you did |
16:49.02 | sn9 | i'd need a very high-rez camera to show that |
16:49.21 | sn9 | use your imagination |
18:54.54 | *** join/#edev sndcrb (~sndcrb@easy-ff00c200-144.dhcp.inet.fi) |
19:54.55 | *** join/#edev flo_lap (~fuchs@Maemo/community/contributor/florian) |
20:08.20 | *** join/#edev SpeedEvil (1000@tor/regular/SpeedEvil) |
20:24.31 | *** join/#edev NightMonkey (~NightMonk@pdpc/supporter/professional/nightmonkey) |
20:29.50 | *** join/#edev SpeedEvil (1000@tor/regular/SpeedEvil) |
22:03.56 | *** join/#edev Crofton (~balister@adsl-75-37-22-143.dsl.pltn13.sbcglobal.net) |
22:50.12 | *** join/#edev djerome (~djerome@ip24-251-158-189.ph.ph.cox.net) |
23:11.21 | *** join/#edev JViz (~JViz@cpe-024-163-014-080.triad.res.rr.com) |
23:32.01 | *** join/#edev SpeedEvil (~user@tor/regular/SpeedEvil) |