Difference between revisions of "Proxmark"

From YobiWiki
Jump to navigation Jump to search
 
(18 intermediate revisions by the same user not shown)
Line 1: Line 1:
  +
 
==Upgrading ELECHOUSE Proxmark3 Easy V3 to 512k==
  
==Upgrading ELECHOUSE Proxmark3 Easy V3 to 512k==
 
===(de)soldering===
  
===(de)soldering===
Line 11: Line 12:
 
* removing the chip with a small [https://www.aliexpress.com/wholesale?SearchText=vacuum+suction+pen+chip suction pen]
  
* removing the chip with a small [https://www.aliexpress.com/wholesale?SearchText=vacuum+suction+pen+chip suction pen]
 
* putting flux on the pads
  
* putting flux on the pads
−
* cleaning the pads with desoldering wire
 +
* cleaning the pads with desoldering wick
 
* putting the new chip and soldering some pins to lock it in place (look for aligning the small dot on the correct corner)
  
* putting the new chip and soldering some pins to lock it in place (look for aligning the small dot on the correct corner)
 
* putting flux on the pins
  
* putting flux on the pins
 
* putting solder on the pins, don't be afraid of bridges...
  
* putting solder on the pins, don't be afraid of bridges...
−
* removing extra solder with desoldering wire
 +
* removing extra solder with desoldering wick
 
* checking carefully for residual solder bridges
  
* checking carefully for residual solder bridges
   
Line 22: Line 23:
 
Note that it's maybe easier to solder the new chip not by using flux+solder+iron but [https://www.aliexpress.com/wholesale?SearchText=mechanic+solder+flux solder flux paste] and heating with air gun, as shown [https://www.youtube.com/watch?v=2Z7nCAxS2Rg&t=9m30s in this video]...
  
Note that it's maybe easier to solder the new chip not by using flux+solder+iron but [https://www.aliexpress.com/wholesale?SearchText=mechanic+solder+flux solder flux paste] and heating with air gun, as shown [https://www.youtube.com/watch?v=2Z7nCAxS2Rg&t=9m30s in this video]...
   
−
===JTAG programming===
 +
===JTAG programming on Proxmark3 Easy V3===
 
Then wire your JTAG programmer to the board. Mine is a Segger J-Link.
  
Then wire your JTAG programmer to the board. Mine is a Segger J-Link.
 
<br>To make it easier, solder a [https://www.aliexpress.com/wholesale?SearchText=breakable+single+row+male+curved+header+2.54 breakable single-row male curved header].
  
<br>To make it easier, solder a [https://www.aliexpress.com/wholesale?SearchText=breakable+single+row+male+curved+header+2.54 breakable single-row male curved header].
Line 42: Line 43:
 
TCK 9
  
TCK 9
 
GND 6
  
GND 6
−
3.3 not connected
 +
3.3 2
   
−
I didn't connect the 3v3 because J-Link Vref is 5v so I prefered to power the PM3 over USB while reprogramming it.
 +
If you prefer you can leave 3.3 not connected and power the PM3 over USB while reprogramming it. But *don't* provide both 3.3 + USB!
   
 
To use the J-Link on Debian:
  
To use the J-Link on Debian:
Line 54: Line 55:
 
ATTRS{idVendor}=="1366", ATTRS{idProduct}=="0101", MODE="664", GROUP="plugdev"
  
ATTRS{idVendor}=="1366", ATTRS{idProduct}=="0101", MODE="664", GROUP="plugdev"
   
−
I created a config file by reusing most of [https://github.com/Proxmark/proxmark3/blob/master/tools/at91sam7s512-buspirate.cfg tools/at91sam7s512-buspirate.cfg], but specific to J-Link instead of buspirate:
 +
I created a config file by reusing most of [https://github.com/Proxmark/proxmark3/blob/master/tools/at91sam7s512-buspirate.cfg tools/at91sam7s512-buspirate.cfg] from Adam Laurie, but specific to J-Link instead of buspirate:
 
<pre>
  
<pre>
 
telnet_port 4444
  
telnet_port 4444
Line 112: Line 113:
 
Launching a telnet:
  
Launching a telnet:
 
<pre>
  
<pre>
−
telnet localhost 4444
 +
$ telnet localhost 4444
 
Connected to localhost.
  
Connected to localhost.
 
Escape character is '^]'.
  
Escape character is '^]'.
Line 122: Line 123:
 
> flash erase_sector 0 0 15
  
> flash erase_sector 0 0 15
 
erased sectors 0 through 15 on flash bank 0 in 0.033260s
  
erased sectors 0 through 15 on flash bank 0 in 0.033260s
  +
> flash erase_sector 1 0 15
  +
erased sectors 0 through 15 on flash bank 1 in 0.033311s
  +
</pre>
  +
Apparently some chips have a bank 1, some don't...
  +
  +
===Backuping firmware===
  ⚫
The chip was new but in case you need to backup the chip content first, this is possible from the telnet session with size = 0x40000 for a 256kb chip and 0x80000 for a 512kb chip. E.g. for 256kb:
  +
<pre>
  +
> dump_image backup.bin 0x100000 0x40000
  +
dumped 262144 bytes in 9.232331s (27.729 KiB/s)
  +
</pre>
  +
  ⚫
===Flashing full image, take 1===
  +
Probably the easiest way, from the telnet session:
  +
<pre>
  +
> flash write_image ./recovery/proxmark3_recovery.bin 0x100000
  +
wrote 217204 bytes from file ./recovery/proxmark3_recovery.bin in 18.417032s (11.517 KiB/s)
  +
</pre>
  +
This works also for restoring a <code>backup.bin</code> image dumped earlier.
  +
===Flashing full image, take 2===
  +
Using the elf images, from the telnet session:
  +
<pre>
  +
> flash write_image ./bootrom/obj/bootrom.elf
  +
> flash write_image ./armsrc/obj/fullimage.elf
  +
</pre>
  +
Note the absence of base address in the command.
  +
===Flashing full image, take 3===
  +
Using the s19 images, from the telnet session:
  +
<pre>
  +
> flash write_image ./bootrom/obj/bootrom.s19 0x100000
  +
> flash write_image ./armsrc/obj/fullimage.s19 0x100000
  +
</pre>
  +
This failed when I tried because the data section seemed to be at a wrong address in <code>fullimage.s19</code>.
  +
  +
Indeed <code>fullimmage.s19</code> loads code and data with offset 0x100000 (<code>flash write_image ./armsrc/obj/fullimage.s19 0x100000</code>):
  +
* code from 0x100000+0x200 to 0x100000+0x33580
  +
* data from 0x100000+0x133580 to 0x100000+0x135200
  +
But data should be at 0x133580 so 0x100000+0x35200
  +
<br>Here is a patch that solved the issue:
  +
<source lang=diff>
  +
diff --git a/common/Makefile.common b/common/Makefile.common
  +
index 95820e66..e697ae14 100644
  +
--- a/common/Makefile.common
  +
+++ b/common/Makefile.common
  +
@@ -99,7 +99,7 @@ $(VERSIONOBJ): $(OBJDIR)/%.o: %.c $(INCLUDES)
  +
# See ldscript.common. -- Henryk Plötz <henryk@ploetzli.ch> 2009-08-27
  +
OBJCOPY_TRANSLATIONS = --no-change-warnings \
  +
--change-addresses -0x100000 --change-start 0 \
  +
- --change-section-address .bss+0 --change-section-address .data+0 \
  +
+ --change-section-address .bss+0 --change-section-address .data-0x100000 \
  +
--change-section-address .commonarea+0
  +
$(OBJDIR)/%.s19: $(OBJDIR)/%.elf
  +
$(OBJCOPY) -Osrec --srec-forceS3 --strip-debug $(OBJCOPY_TRANSLATIONS) $^ $@
  +
</source>
  +
  +
Update: patch applied in https://github.com/iceman1001/proxmark3/commit/d59026518e02c37ab0673a670b9ea14e17ae399c
  +
  +
===Flashing full image, take 4===
  +
In case flashing the full image from JTAG seems to fail, you can flash only the bootloader:
  +
<pre>
 
> flash write_image /tmp/bootrom.s19 0x100000
 
> flash write_image /tmp/bootrom.s19 0x100000
 
wrote 3624 bytes from file /tmp/bootrom.s19 in 0.392169s (9.024 KiB/s)
  
wrote 3624 bytes from file /tmp/bootrom.s19 in 0.392169s (9.024 KiB/s)
 
</pre>
  
</pre>
  +
Then follow the usual recovery procedure:
⚫
The chip was new but in case you need to backup the chip content first, this should be possible with sth like "dump_image original.bin 0x100000 0x80000"
  
⚫
===Flashing full image===
  
−
I also tried to flash the fullimage via JTAG but it failed working afterwards, so once the bootloader is installed, I used the usual recovery procedure:
  
 
* Press button and keep it pressed during the whole procedure
  
* Press button and keep it pressed during the whole procedure
 
* Plug PM3 to USB
  
* Plug PM3 to USB
Line 150: Line 209:
 
Architecture Identifier: AT91SAM7Sxx Series
  
Architecture Identifier: AT91SAM7Sxx Series
 
Nonvolatile Program Memory Type: Embedded Flash Memory
  
Nonvolatile Program Memory Type: Embedded Flash Memory
  +
</pre>
  +
==JTAG connectors==
  +
===Proxmark3 with headers===
  +
[[File:Pm3.jpg|400px]]
  +
===Proxmark3 without headers===
  +
[[File:Pm3naked1.jpg|400px]] [[File:Pm3naked2.jpg|400px]]
  +
<br>Apply some pressure to ensure a proper connection during the whole flashing operation
  +
  +
===Proxmark3 Easy with headers===
  +
[[File:Pm3easy.jpg|400px]] [[File:Pm3easy2.jpg|400px]] [[File:Pm3easy3.jpg|400px]]
  +
<br>See above for the proper pinout.
  +
<br>If you want to solder a header, beware that there is very little space between both PCBs. It's possible to solder curved pins but without their plastic support, see pics.
  +
<br>I used a breakout but you can directly connect to the JLink if you prefer. You can fint the board by searching for "jlink adapter card" on aliexpress. You'll need to add a jumper to connect jlink pin2 and 3.3v.
  +
===Proxmark3 Easy without headers===
  +
See [[Proxmark#JTAG_programming_on_Proxmark3_Easy_V3|above]] for pinouts
  +
  +
[[File:Pm3easy4.jpg|400px]]
  +
<br>Using Dupont individual wires
  +
  +
[[File:Pm3easy6.jpg|400px]] [[File:Pm3easy5.jpg|400px]]
  +
<br>Using the same cable as shown above
  +
  +
==Resources==
  +
* https://github.com/RfidResearchGroup/proxmark3/blob/master/doc/jtag_notes.md
  +
* https://github.com/Proxmark/proxmark3/wiki/Debricking-Proxmark3 using a Bus Pirate
  +
* "Lock Error Bit Detected, Operation Abort" ?
  +
<pre>
  +
flash protect 0 0 15 off
  +
flash protect 1 0 15 off
 
</pre>
  
</pre>

Latest revision as of 16:42, 16 October 2020

Upgrading ELECHOUSE Proxmark3 Easy V3 to 512k

(de)soldering

The popular Proxmark3 Easy has an at91sam7s256 with only 256k and e.g. it's already about 83% full with the current iceman firmware.
So I decided to attempt an upgrade.

at91sam7s512 is about 15€ on Farnell.

The steps I followed to desolder and solder the new chip are basically the same as seen on this youtube video:

  • heating the chip with my desoldering station
  • removing the chip with a small suction pen
  • putting flux on the pads
  • cleaning the pads with desoldering wick
  • putting the new chip and soldering some pins to lock it in place (look for aligning the small dot on the correct corner)
  • putting flux on the pins
  • putting solder on the pins, don't be afraid of bridges...
  • removing extra solder with desoldering wick
  • checking carefully for residual solder bridges

And voila.

Note that it's maybe easier to solder the new chip not by using flux+solder+iron but solder flux paste and heating with air gun, as shown in this video...

JTAG programming on Proxmark3 Easy V3

Then wire your JTAG programmer to the board. Mine is a Segger J-Link.
To make it easier, solder a breakable single-row male curved header.
Choose a curved one so you can leave it in place later and still stack the PM3 daughterboard.

Then using Dupont wires male-female, wire it to the JTAG programmer.

For the J-Link, the pinout is: 

  ---------  ---------
 |1917151311 9 7 5 3 1|
 |201816141210 8 6 4 2|
  --------------------

PM3  JLink
---  -----
TMS   7
TDI   5
TDO  13
TCK   9
GND   6
3.3   2

If you prefer you can leave 3.3 not connected and power the PM3 over USB while reprogramming it. But *don't* provide both 3.3 + USB!

To use the J-Link on Debian: 

$ apt-get install openocd

There is some doc installed locally: file:///usr/share/doc/openocd/openocd.html/index.html

Create /etc/udev/rules.d/60-jlink.rules with 

ATTRS{idVendor}=="1366", ATTRS{idProduct}=="0101", MODE="664", GROUP="plugdev"

I created a config file by reusing most of tools/at91sam7s512-buspirate.cfg from Adam Laurie, but specific to J-Link instead of buspirate: 

telnet_port 4444
gdb_port 3333
interface jlink
transport select jtag
adapter_khz 1000
reset_config srst_only srst_pulls_trst
jtag newtap sam7x cpu -irlen 4 -ircapture 0x1 -irmask 0xf -expected-id 0x3f0f0f0f
target create sam7x.cpu arm7tdmi -endian little -chain-position sam7x.cpu
sam7x.cpu configure -event reset-init {
	soft_reset_halt
	mww 0xfffffd00 0xa5000004	# RSTC_CR: Reset peripherals
	mww 0xfffffd44 0x00008000	# WDT_MR: disable watchdog
	mww 0xfffffd08 0xa5000001	# RSTC_MR enable user reset
	mww 0xfffffc20 0x00005001	# CKGR_MOR : enable the main oscillator
	sleep 10
	mww 0xfffffc2c 0x000b1c02	# CKGR_PLLR: 16MHz * 12/2 = 96MHz
	sleep 10
	mww 0xfffffc30 0x00000007	# PMC_MCKR : MCK = PLL / 2 = 48 MHz
	sleep 10
	mww 0xffffff60 0x00480100	# MC_FMR: flash mode (FWS=1,FMCN=72)
	sleep 100
}
gdb_memory_map enable
sam7x.cpu configure -work-area-virt 0 -work-area-phys 0x00200000 -work-area-size 0x10000 -work-area-backup 0
flash bank sam7x.flash.0 at91sam7 0 0 0 0 sam7x.cpu 0 0 0 0 0 0 0 18432
flash bank sam7x.flash.1 at91sam7 0 0 0 0 sam7x.cpu 1 0 0 0 0 0 0 18432

Launching OpenOCD: 

$ openocd -f at91sam7s512-jlink.cfg
Open On-Chip Debugger 0.9.0 (2017-03-07-13:28)
Licensed under GNU GPL v2
For bug reports, read
	http://openocd.org/doc/doxygen/bugs.html
adapter speed: 1000 kHz
srst_only srst_pulls_trst srst_gates_jtag srst_open_drain connect_deassert_srst
Info : J-Link ARM V8 compiled Dec  1 2009 11:42:48
Info : J-Link caps 0xb9ff7bbf
Info : J-Link hw version 80000
Info : J-Link hw type J-Link
Info : J-Link max mem block 9576
Info : J-Link configuration
Info : USB-Address: 0x0
Info : Kickstart power on JTAG-pin 19: 0xffffffff
Info : Vref = 3.332 TCK = 1 TDI = 0 TDO = 0 TMS = 0 SRST = 1 TRST = 1
Info : J-Link JTAG Interface ready
Info : clock speed 1000 kHz
Info : JTAG tap: sam7x.cpu tap/device found: 0x3f0f0f0f (mfg: 0x787, part: 0xf0f0, ver: 0x3)
Info : Embedded ICE version 1
Info : sam7x.cpu: hardware has 2 breakpoint/watchpoint units

Launching a telnet: 

$ telnet localhost 4444
Connected to localhost.
Escape character is '^]'.
Open On-Chip Debugger
> halt
target state: halted
target halted in ARM state due to debug-request, current mode: Supervisor
cpsr: 0xf00000d3 pc: 0x001c9c60
> flash erase_sector 0 0 15
erased sectors 0 through 15 on flash bank 0 in 0.033260s
> flash erase_sector 1 0 15
erased sectors 0 through 15 on flash bank 1 in 0.033311s

Apparently some chips have a bank 1, some don't...

Backuping firmware

The chip was new but in case you need to backup the chip content first, this is possible from the telnet session with size = 0x40000 for a 256kb chip and 0x80000 for a 512kb chip. E.g. for 256kb: 

> dump_image backup.bin 0x100000 0x40000
dumped 262144 bytes in 9.232331s (27.729 KiB/s)

Flashing full image, take 1

Probably the easiest way, from the telnet session: 

> flash write_image ./recovery/proxmark3_recovery.bin 0x100000
wrote 217204 bytes from file ./recovery/proxmark3_recovery.bin in 18.417032s (11.517 KiB/s)

This works also for restoring a backup.bin image dumped earlier.

Flashing full image, take 2

Using the elf images, from the telnet session: 

> flash write_image ./bootrom/obj/bootrom.elf
> flash write_image ./armsrc/obj/fullimage.elf

Note the absence of base address in the command.

Flashing full image, take 3

Using the s19 images, from the telnet session: 

> flash write_image ./bootrom/obj/bootrom.s19 0x100000
> flash write_image ./armsrc/obj/fullimage.s19 0x100000

This failed when I tried because the data section seemed to be at a wrong address in fullimage.s19.

Indeed fullimmage.s19 loads code and data with offset 0x100000 (flash write_image ./armsrc/obj/fullimage.s19 0x100000):

  • code from 0x100000+0x200 to 0x100000+0x33580
  • data from 0x100000+0x133580 to 0x100000+0x135200

But data should be at 0x133580 so 0x100000+0x35200
Here is a patch that solved the issue: 

diff --git a/common/Makefile.common b/common/Makefile.common
index 95820e66..e697ae14 100644
--- a/common/Makefile.common
+++ b/common/Makefile.common
@@ -99,7 +99,7 @@ $(VERSIONOBJ): $(OBJDIR)/%.o: %.c $(INCLUDES)
 # See ldscript.common. -- Henryk Plötz <henryk@ploetzli.ch> 2009-08-27
 OBJCOPY_TRANSLATIONS = --no-change-warnings \
        --change-addresses -0x100000 --change-start 0 \
-       --change-section-address .bss+0 --change-section-address .data+0 \
+       --change-section-address .bss+0 --change-section-address .data-0x100000 \
        --change-section-address .commonarea+0
 $(OBJDIR)/%.s19: $(OBJDIR)/%.elf
        $(OBJCOPY) -Osrec --srec-forceS3 --strip-debug $(OBJCOPY_TRANSLATIONS) $^ $@

Update: patch applied in https://github.com/iceman1001/proxmark3/commit/d59026518e02c37ab0673a670b9ea14e17ae399c

Flashing full image, take 4

In case flashing the full image from JTAG seems to fail, you can flash only the bootloader: 

> flash write_image /tmp/bootrom.s19 0x100000  
wrote 3624 bytes from file /tmp/bootrom.s19 in 0.392169s (9.024 KiB/s)

Then follow the usual recovery procedure:

  • Press button and keep it pressed during the whole procedure
  • Plug PM3 to USB
  • ./flasher /dev/ttyACM0 fullimage.elf
  • Release button and re-plug the PM3

Done

$ ./proxmark3 /dev/ttyACM0
Proxmark3 RFID instrument
bootrom: iceman/iceman/v2.1.0-1547-gb0df293d 2017-05-03 20:44:34
os: iceman/iceman/v2.1.0-1547-gb0df293d 2017-05-03 20:44:39
LF FPGA image built for 2s30vq100 on 2015/03/06 at 07:38:04
HF FPGA image built for 2s30vq100 on 2015/11/ 2 at  9: 8: 8

uC: AT91SAM7S512 Rev B
Embedded Processor: ARM7TDMI
Nonvolatile Program Memory Size: 512K bytes. Used: 217204 bytes (41%).
Free: 307084 bytes (59%).
Second Nonvolatile Program Memory Size: None
Internal SRAM Size: 64K bytes
Architecture Identifier: AT91SAM7Sxx Series
Nonvolatile Program Memory Type: Embedded Flash Memory

JTAG connectors

Proxmark3 with headers

Pm3.jpg

Proxmark3 without headers

Pm3naked1.jpg Pm3naked2.jpg
Apply some pressure to ensure a proper connection during the whole flashing operation

Proxmark3 Easy with headers

Pm3easy.jpg Pm3easy2.jpg Pm3easy3.jpg
See above for the proper pinout.
If you want to solder a header, beware that there is very little space between both PCBs. It's possible to solder curved pins but without their plastic support, see pics.
I used a breakout but you can directly connect to the JLink if you prefer. You can fint the board by searching for "jlink adapter card" on aliexpress. You'll need to add a jumper to connect jlink pin2 and 3.3v.

Proxmark3 Easy without headers

See above for pinouts

Pm3easy4.jpg
Using Dupont individual wires

Pm3easy6.jpg Pm3easy5.jpg
Using the same cable as shown above

Resources

flash protect 0 0 15 off
flash protect 1 0 15 off
Retrieved from "https://wiki.yobi.be/index.php?title=Proxmark&oldid=10520"