search

avrdudes / avrdude

AVRDUDE is a utility to program AVR microcontrollers
GNU General Public License v2.0
696 stars 136 forks source link

avrdude can not write usersig of XMEGA part in terminal mode #1361

Closed mcuee closed 1 year ago

mcuee commented 1 year ago

Somehow avrdude can only read but not write user signature row of the XMEGA part, tested with AVRISP mkII and ATxmega32A4U.

avrdude version 7.1-20230504 (e6a2f3c0), URL: <https://github.com/avrdudes/avrdude>
PS C:\work\avr\avrdude_test\avrdude_bin> .\avrdude -c avrispmkii -p x32a4u -t
avrdude: AVR device initialized and ready to accept instructions
avrdude: device signature = 0x1e9541 (probably x32a4u)
avrdude> dump prodsig
Reading | ################################################## | 100% 0.17 s
0000  0d 40 7b 0a 40 7f ff 01  38 55 30 31 35 32 ff ff  | @{ @...8U0152..|
0010  03 ff 05 00 16 00 ff ff  ff ff cf 07 24 40 ff ff  |............$@..|
0020  00 00 00 ff ff ff ff ff  ff ff ff ff ff ff 5e 09  |..............^ |
0030  ff ff                                             |..              |
avrdude> dump usersig
Reading | ################################################## | 100% 0.41 s
0000  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
0010  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
0020  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
0030  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
0040  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
0050  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
0060  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
0070  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
0080  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
0090  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
00a0  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
00b0  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
00c0  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
00d0  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
00e0  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
00f0  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
avrdude> write usersig 0 0xaa
Writing | -------------------------------------------------- | 0% 0.01 s
avrdude error: (write) verification error writing 0xaa at 0x00000 cell=0xff
avrdude> flush
avrdude> dump usersig 0 0x10
Reading | ################################################## | 100% 0.05 s
0000  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
avrdude> quit

avrdude done.  Thank you.
mcuee commented 1 year ago

I am not so sure if #379 is similar or not. I do not have access to RF AVR parts.

mcuee commented 1 year ago

@MCUdude

Please help to check with your programmers and XMEGA parts to see if you have the same issues or not. Thanks.

MCUdude commented 1 year ago

I can confirm that this is an issue with the Power Debugger as well:

$ ./avrdude -cpowerdebugger_pdi -patxmega128a3u -t
avrdude: AVR device initialized and ready to accept instructions
avrdude: device signature = 0x1e9742 (probably x128a3u)
avrdude> read usersig
Reading | ################################################## | 100% 0.42 s 
0000  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
0010  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
0020  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
0030  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
0040  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
0050  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
0060  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
0070  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
0080  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
0090  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
00a0  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
00b0  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
00c0  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
00d0  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
00e0  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
00f0  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
avrdude> write usersig 0 0x10
Writing | -------------------------------------------------- | 0% 0.00 s 
avrdude error: (write) verification error writing 0x10 at 0x00000 cell=0xff
avrdude>
MCUdude commented 1 year ago

However, userrow/usersig works with UPDI targets:

$ ./avrdude -cpkobn_updi -patmega4809 -t
avrdude: AVR device initialized and ready to accept instructions
avrdude: device signature = 0x1e9651 (probably m4809)
avrdude> read usersig
Reading | ################################################## | 100% 0.26 s 
0000  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
0010  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
0020  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
0030  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
avrdude> write usersig 0 "Hello World!"
Writing | ################################################## | 100% 0.19 s 
avrdude> read usersig
Reading | ################################################## | 100% 0.26 s 
0000  48 65 6c 6c 6f 20 57 6f  72 6c 64 21 00 ff ff ff  |Hello World!....|
0010  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
0020  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
0030  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
avrdude> quit

avrdude done.  Thank you.
mcuee commented 1 year ago

@stefanrueger has the comment here, which may explain why UPDI parts like ATmega4809 works.

#------------------------------------------------------------
# AVR8X mega family common values
#------------------------------------------------------------

part parent ".avr8x"
    desc                   = "AVR8X mega family common values";
    id                     = ".avr8x_mega";
    family_id              = "megaAVR";
    # Dedicated UPDI pin, no HV
    hvupdi_variant         = 1;

    memory "userrow"
        size               = 64;
        page_size          = 64;
        offset             = 0x1300;
        readsize           = 256;
    ;

    memory "usersig"
        alias "userrow";
    ;
;
mcuee commented 1 year ago

However, the simple patch from the comment by @stefanrueger does not work yet becaue of the cache size.

PS C:\work\avr\avrdude_test\avrdude_sr> git diff
diff --git a/src/avr.c b/src/avr.c
index ea80f4c..070ad30 100644
--- a/src/avr.c
+++ b/src/avr.c
@@ -1375,6 +1375,8 @@ int avr_memtype_is_flash_type(const char *memtype) {
      strcmp(memtype, "flash") == 0 ||
      strcmp(memtype, "application") == 0 ||
      strcmp(memtype, "apptable") == 0 ||
+        strcmp(memtype, "usersig") == 0 ||
+        strcmp(memtype, "userrow") == 0 ||
      strcmp(memtype, "boot") == 0);
 }

PS C:\work\avr\avrdude_test\avrdude_bin> .\avrdude_pr1360mod -C .\avrdude_pr1360.conf -c avrispmkii -p x32a4u -t
avrdude_pr1360mod: AVR device initialized and ready to accept instructions
avrdude_pr1360mod: device signature = 0x1e9541 (probably x32a4u)
avrdude> dump usersig
Reading | -------------------------------------------------- | 0% 0.00 s
avrdude_pr1360mod error: usersig cache address 0xe0400 out of range [0, 0x8fff]
avrdude_pr1360mod error: (dump) error reading usersig address 0x00000 of part ATxmega32A4U
                         read operation not supported on memory type usersig
avrdude> write usersig 0 0xaa
Caching | -------------------------------------------------- | 0% 0.00 s
avrdude_pr1360mod error: usersig cache address 0xe0400 out of range [0, 0x8fff]
avrdude_pr1360mod error: (write) error writing 0xaa at 0x00000, rc=-1
                          write operation not supported on memory type usersig
avrdude> quit

avrdude_pr1360mod done.  Thank you.
stefanrueger commented 1 year ago

@mcuee The curiosity is getting the better of me: If you do similar changes for the ATxmega32A4U's usersig memory, can it be written to then?

mcuee commented 1 year ago

@mcuee The curiosity is getting the better of me: If you do similar changes for the ATxmega32A4U's usersig memory, can it be written to then?

I think it has already the memory definition.


#------------------------------------------------------------
# ATxmega32A4U
#------------------------------------------------------------

part parent ".xmega-a"
    desc                   = "ATxmega32A4U";
    id                     = "x32a4u";
    variants               =
        "ATxmega32A4U-AN:  TQFP44,  Fmax=32 MHz, T=[0 C,   105 C], Vcc=[1.6 V, 3.6 V]",
        "ATxmega32A4U-ANR: TQFP44,  Fmax=32 MHz, T=[0 C,   105 C], Vcc=[1.6 V, 3.6 V]",
        "ATxmega32A4U-AU:  TQFP44,  Fmax=32 MHz, T=[-40 C,  85 C], Vcc=[1.6 V, 3.6 V]",
        "ATxmega32A4U-AUR: TQFP44,  Fmax=32 MHz, T=[-40 C,  85 C], Vcc=[1.6 V, 3.6 V]",
        "ATxmega32A4U-CU:  BGA49,   Fmax=32 MHz, T=[-40 C,  85 C], Vcc=[1.6 V, 3.6 V]",
        "ATxmega32A4U-CUR: VFBGA49, Fmax=N/A,    T=[N/A,     N/A], Vcc=[N/A,     N/A]",
        "ATxmega32A4U-M7:  VQFN44,  Fmax=32 MHz, T=[-40 C, 105 C], Vcc=[1.6 V, 3.6 V]",
        "ATxmega32A4U-MH:  QFN44,   Fmax=32 MHz, T=[-40 C,  85 C], Vcc=[1.6 V, 3.6 V]",
        "ATxmega32A4U-MHR: VQFN44,  Fmax=32 MHz, T=[-40 C,  85 C], Vcc=[1.6 V, 3.6 V]";
    mcuid                  = 239;
    n_interrupts           = 127;
    boot_section_size      = 4096;
    signature              = 0x1e 0x95 0x41;
    usbpid                 = 0x2fe4;

    memory "eeprom"
        size               = 1024;
        page_size          = 32;
        offset             = 0x8c0000;
        readsize           = 256;
    ;

    memory "flash"
        size               = 0x9000;
        page_size          = 256;
        offset             = 0x800000;
        readsize           = 256;
    ;

    memory "application"
        size               = 0x8000;
        page_size          = 256;
        offset             = 0x800000;
        readsize           = 256;
    ;

    memory "apptable"
        size               = 4096;
        page_size          = 256;
        offset             = 0x807000;
        readsize           = 256;
    ;

    memory "boot"
        size               = 4096;
        page_size          = 256;
        offset             = 0x808000;
        readsize           = 256;
    ;

    memory "fuse4"
        initval            = 0xfe;
    ;

    memory "usersig"
        size               = 256;
        page_size          = 256;
        offset             = 0x8e0400;
        readsize           = 256;
    ;
;
mcuee commented 1 year ago

@stefanrueger

I think we only need to fix the following issue. But I do not know how to fix it myself.

usersig cache address 0xe0400 out of range [0, 0x8fff]
stefanrueger commented 1 year ago

becaue of the cache size

Good shout! So, for the terminal this needs more programming to bring usersig into the flash memory cache regime. What about the command line? -U usersig:w:file

stefanrueger commented 1 year ago

userrow/usersig works with UPDI targets

Interesting. The terminal does not cache these memories, so the UPDI programmer probably is able to do bytewise r/w?

mcuee commented 1 year ago

becaue of the cache size

Good shout! So, for the terminal this needs more programming to bring usersig into the flash memory cache regime. What about the command line? -U usersig:w:file

That works fine. So I change the title to terminal mode.

PS C:\work\avr\avrdude_test\avrdude_bin> .\avrdude -c avrispmkii -p x32a4u -U usersig:w:entest_64B.hex:i
avrdude: AVR device initialized and ready to accept instructions
avrdude: device signature = 0x1e9541 (probably x32a4u)
avrdude: Note: programmer supports page erase for Xmega devices.
         Each page will be erased before programming it, but no chip erase is performed.
         To disable page erases, specify the -D option; for a chip-erase, use the -e option.
avrdude: reading input file entest_64B.hex for usersig
         with 64 bytes in 1 section within [0, 0x3f]
         using 1 page and 192 pad bytes
avrdude: writing 64 bytes usersig ...
Writing | ################################################## | 100% 0.02 s
avrdude: 64 bytes of usersig written
avrdude: verifying usersig memory against entest_64B.hex
Reading | ################################################## | 100% 0.01 s
avrdude: 64 bytes of usersig verified

avrdude done.  Thank you.

PS C:\work\avr\avrdude_test\avrdude_bin> .\avrdude -c avrispmkii -p x32a4u -t
avrdude: AVR device initialized and ready to accept instructions
avrdude: device signature = 0x1e9541 (probably x32a4u)
avrdude> dump usersig
Reading | ################################################## | 100% 0.41 s
0000  54 68 65 20 71 75 69 63  6b 20 62 72 6f 77 6e 20  |The quick brown |
0010  66 6f 78 20 6a 75 6d 70  73 20 6f 76 65 72 20 74  |fox jumps over t|
0020  68 65 20 6c 61 7a 79 20  64 6f 67 0a 54 68 65 20  |he lazy dog The |
0030  71 75 69 63 6b 20 62 72  6f 77 6e 20 66 6f 78 20  |quick brown fox |
0040  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
0050  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
0060  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
0070  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
0080  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
0090  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
00a0  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
00b0  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
00c0  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
00d0  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
00e0  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
00f0  ff ff ff ff ff ff ff ff  ff ff ff ff ff ff ff ff  |................|
avrdude> quit

avrdude done.  Thank you.
stefanrueger commented 1 year ago

The memories below all have more than 9 bytes, so are potentially of the kind that might not deal gracefully with bytewise read/write. AVRDUDE has implemented a terminal cache for only 5 of these. Of the other five memories only two are writeable and need special cache treatment b/c somehow programmers cannot do bytewise read/write.

Number of avrdude parts that have this particular memory

Memory exotic TPI ISP PDI UPI Cache
application 50 ✓
apptable 50 ✓
boot 50 ✓
eeprom 2 186 50 91 ✓
flash 4 8 188 50 91 ✓
prodsig 50 91 read only
sernum 91 read only
sigrow 2 read only
userrow 91 ❌
usersig 50 ❌
mcuee commented 1 year ago

@stefanrueger

In this case, using -c dryrun is not able to reproduce the issue. I am not sure if you can enhance -c dryrun so that it can reproduce the issue.

PS C:\work\avr\avrdude_test\avrdude_bin> .\avrdude -c dryrun -p x32a4u -t
avrdude: AVR device initialized and ready to accept instructions
avrdude: device signature = 0x1e9541 (probably x32a4u)
avrdude> dump usersig
Reading | ################################################## | 100% 0.30 s
0000  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
0010  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
0020  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
0030  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
0040  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
0050  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
0060  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
0070  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
0080  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
0090  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
00a0  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
00b0  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
00c0  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
00d0  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
00e0  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
00f0  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
avrdude> write usersig 0 0xaa
Writing | ################################################## | 100% 0.00 s
avrdude> dump usersig
Reading | ################################################## | 100% 0.27 s
0000  aa 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
0010  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
0020  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
0030  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
0040  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
0050  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
0060  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
0070  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
0080  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
0090  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
00a0  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
00b0  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
00c0  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
00d0  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
00e0  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
00f0  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
avrdude> quit

avrdude done.  Thank you.
stefanrueger commented 1 year ago

-c dryrun is not able to reproduce the issue

That's fair enough. Terminal code uses bytewise access to memories only, and was traditionally not able to deal with paged memories. When the terminal implemented a cache that translates bytewise access to paged access for paged memories, the usersig memory was not dealt with. It is "simply" a case of providing a cache for that memory in terminal. There is no way a programmer (which is what dryrun is) can have that level of detail about AVRDUDE working.