--- Log for 19.10.119 Server: cherryh.freenode.net Channel: #rockbox --- Nick: logbot Version: Dancer V4.16 Started: 27 days and 0 hours ago 00.06.39 # I performed the "dd of=/dev/sdb if="clipzip4GB_RECOVERY_image32.bin" bs=512" command. The output had three lines. Something in, something out, and a line with the speed and total. 00.07.12 # It seemed promising. But, when I repeated the dumb, all I got was zeros again. So, this area seems write protected. 00.07.48 # It seems strange that dd didn't notice the write protection. 00.08.45 # I meant to write dump above not dumb. 00.09.47 # Are there any options to dd that might be worth trying to bypass any write protection? 00.10.36 # The size of the dumps were the same as the clipzip4GB_RECOVERY_image32.bin file. So, I think I was doing things correctly. 00.10.50 # Compressed, those sizes would be about zero, however. 00.11.05 # (For the dumps.) 00.11.45 # My guess is sandisk has a special way of writing to this area. It's handed out in case it's a mater of national security. 00.12.12 # But for us, no go. Perhaps that explains a few things. 00.13.02 # jason666: given that I've wriiten it in the past, it's more likely to be failing 00.13.36 # The paradox remains: how do the boot logos show up (both of them, OF and rockbox) if the flash drive isn't accessible? 00.13.41 # which would explain why dd didn't fail 00.14.16 # most flash chips silently lock as read only when they are close to failure. 00.14.49 # But all sansa flash disks lock at a 30.6 mb zeroed area that's unwritable. 00.14.50 # at that point the only option I know of is to backup and replace. 00.15.41 # Does anyone here know enough about the boot process to resolve the paradox? 00.17.12 # jason have you made a dump of the entire drive? 00.17.49 # I performed the dump that was suggested, and it is only 30.6 mb in size. All zeros. 00.18.48 # The rest isn't accessible, as far as I can tell. I'm happy to try other commands. I used the one bilgus said to use. 00.18.59 # With the F000. 00.19.11 # then i would suggest "dd if= of=backup.dd status=progress" 00.19.39 # that way you can always try to recover your files. 00.20.52 # Mandlelmun, so I will try the same dd command bilgus suggested, except no feeding it of any parameters and request a status bar. Ok. One moment. 00.22.14 # Just remember to give the backup a name where you wont accidentally overwrite it with another one later. 00.22.20 Quit ZincAlloy (Quit: Leaving.) 00.24.07 # jason666: actually can you use pastebin to give the results of lsblk? 00.26.00 Quit petur (Remote host closed the connection) 00.32.37 Quit lebellium (Quit: Leaving) 00.33.47 # There was a slight difference. Firstly my version of dd doesn't allow the status=progress flag. It allowed me to disable status output after the command was done only. 00.34.36 # that shouldn't matter. but how gib is the output file. 00.34.41 # *big 00.34.52 # The only difference was the file size. Your output caused a ~32000000 byte output, while bilgus' was 31000000 bytes. Both outputs were only zeros. 00.36.04 # can you pastebin lsblk? 00.36.27 # I am unable to mount the clip zip recovery mode drive. does this lsblk command work on devices directly? 00.36.40 # yes 00.37.50 # Well, not only am I unable to mount the drive but I don' 00.37.58 # t have lsblk on my system anyways. 00.38.27 # gparted sees it however and is ready to mess around. 00.39.01 # Perhaps inability to mount is an important problem? 00.39.24 # no if you were able to mount it would indicate something wrong. 00.40.00 # are you able to connect it without recovery at all? 00.44.37 # I think lsblk probably gives the same infor that gparted does. Gparted says: device information:Undetermined. storage size = 30.61 MB. Partitions: unrecognized. Heads: 255. Sectors/track = 63. Cylinders = 3. Total sectors = 62688. Secor size = 512. 00.46.01 # When I turn it on it goes into rockbox screen. After 60 seconds it says ~"CONNECT TO USB" . BUt if i connect nothing shows. If I press << while booting it does dual sansa logo spins and halts. 00.47.26 # what happens if you plug it in while powered off? 00.49.59 # There are about 10 different states that I can I can plug it is: when off. when in recovery. When either logo appears, after either logo. during either logo. When It says to connect. before it says to connect, etc. Which time do you want me to describe mendelmun? 00.50.36 # when completely off. 00.50.37 # The wiki says to boot into OF pressing << and then to UN plug it! 00.50.54 # ok. 00.51.02 # I mean plug it into a computer not the wall. 00.51.03 # one moment 00.56.08 # whats the output of lsusb? 00.56.38 # jason sorry ive got to go offline for ~25 hrs good luck. 00.57.12 # thanks for your help emdelum 00.57.59 # fyi there was no offer when plugging in when off. the rockbox logo screen appears and never goes away. (if unplugged it goes to plug into usb but no usb offer is made) 00.58.28 # I also plugged in when the sansa logos appeared but no usb offer is ever made. but I didn't wait that long, either 01.02.19 # jason666, if it is only showing ~32 mb drive its already gone 01.02.44 # they go read only and lock in the corruption when they go 01.03.17 # thats why you see the sansa logo and rb logo the previous data is mostly there but you can't do anything with it 01.11.26 # bilgus you seem to think you know the solution to the paradox, but where is the logo stored? 01.11.36 # Both logos. 01.12.50 # in the firmware image 01.12.54 # on the drive 01.13.54 # you can try a few more times getting into the hidden area maybe you just aren't removing the short quickly enough but I doubt it 01.14.29 # the reason the recovery method works is that the short disable the flash chip at first boot and the processor goes into raw mode 01.15.41 # you should be able to get data from it though if it still boots 01.19.29 # be sure fdisk -l states 'Disk /dev/sdb doesn't contain a valid partition table' 01.19.31 # Ok let me tell you what's what. The rockbox firmware is highly flawed. It needs to not hang if the booting flashdrive goes into read only mode. It needs to recognize that it isn't able to write and go into usb-ready mode, allowing people to get their data off of it. 01.20.03 # and you can rightly fuck off no firmware works after shit goes south 01.20.38 # It's just logical to allow for a failure mode that let's people get their data! 01.21.45 # you can't do anything if your code gets corrupted thats the flaw with the sansa clip series not rockbox 01.22.05 # Regarding fdisk idea, I think gparted basically confirmed that. (See above.) 01.23.47 # No you certainly can write the code so that if it is hanging when first trying to write (or any time) after enough time has elapsed it needs to execute usb-mode code that does not attempt any more write to the usb drive by itself(unless it has to because the host computer tries to write). This is needed in case someone wants to mark a sector as 01.23.47 # bad. 01.24.08 # It's a logical improvement. I know change causes pain but wake up. 01.24.24 Quit beencubed (Ping timeout: 276 seconds) 01.25.06 # ok so try this take your harddrive and erase every third sector now try to boot something off the drive to back it up 01.26.03 # you just need to realize you are expecting the impossible 01.26.16 # There's no evidence that anything is unreadable on the 4gb flash drive chip. 01.26.36 # It has gotten into a read only mode, that's all. 01.26.50 # no its perfectly readable just corrupted and readonly 01.27.14 # The nice thing about open source is that not everyone needs to agree on an idea for it to be implimented. 01.28.01 # ok well once you implement it I'll believe you till then it stays in the realm of impossible 01.29.28 # Bilgus hypothetically if a perfectly fresh installation of rockbox internal flash drive was made unwritable, where would the boot process hang? 01.29.51 # it wouldn't 01.30.09 # It never writes anything during the boot process? 01.30.24 # no 01.30.31 # never ever 01.30.41 # Ok, then I'm wrong. 01.30.49 # the bootloader is separate from the firmware 01.33.03 *** Saving seen data "./dancer.seen" 01.33.03 # So rockbox goes into the resume playing without any writes. I didn't know this. 01.34.02 # like I said the bootloader and the firmware are separate 01.34.59 # I made a multiboot bootloader for the situation that your flash goes readonly but guess what its already too late 01.35.11 # new users get it by default 01.35.48 # Does that bootloader allow you to go into a usb-data mode even if the internal drive is read only? 01.36.29 Join beencubed [0] (~beencubed@209.131.238.248) 01.40.27 # it should.. 01.43.07 # It's strange that you don't know for sure. 01.43.48 # The failure point is the same on all these devices. It goes into read only mode and hangs during the start up process when it first needs to do a write request. 01.45.02 # Because nobody thought of considering that a hang might cause data loss and that a few lines of read-only detect code jump to usb mode code could be of value. 01.45.33 # So by now a few people should have thanked you. 01.45.42 # For saving their data. 01.47.17 # oh I am sure it does go into usb mode when the cbootloader code is intact 02.03.02 Quit PrinceKyle (Read error: Connection timed out) 02.04.40 Join PrinceKyle [0] (~cancerian@cpe-2606-A000-111A-A097-0-0-0-F17.dyn6.twc.com) 02.24.15 Join massiveH [0] (~massiveH@ool-18e4eaeb.dyn.optonline.net) 02.42.40 # Right, rockbox developers were smart enough to think of this idea before. So, after about 60 seconds of "booting" my device says "Plug USB Cable" But, the code is bad. It doesn't connect to windows or linux. It does show up as an unrecognized usb device, but not as a usb storage device. The code being run during "Plug USB Cable" is bad. You 02.42.40 # can verify this by reading this: https://www.google.com/search?q=site:rockbox.org+%22plug+usb+cable% It's a bug. 02.43.30 # corrected url: https://www.google.com/search?q=site:rockbox.org+%22plug+usb+cable%22 02.58.04 # " the cbootloader code is intact" Did you mean "the bootloader code is intact" What's the cbootloader code? 03.33.06 *** Saving seen data "./dancer.seen" 03.33.54 Quit michaelni (Ping timeout: 265 seconds) 03.46.39 Join michaelni [0] (~michael@213-47-68-29.cable.dynamic.surfer.at) 04.22.44 Join dandels [0] (~dandels@unaffiliated/dandels) 04.36.26 # the bootloader code is machine code compiled from c but I meant the bootloader code obviously.. 04.39.24 Join reductum [0] (~weechat@cpe-104-32-77-28.socal.res.rr.com) 05.04.32 Quit dandels (Quit: WeeChat 2.6) 05.11.45 Quit igitoor (Remote host closed the connection) 05.33.09 *** Saving seen data "./dancer.seen" 06.49.38 Quit PrinceKyle (Ping timeout: 250 seconds) 06.53.34 Join PrinceKyle [0] (~cancerian@cpe-2606-A000-111A-A097-0-0-0-F17.dyn6.twc.com) 07.19.29 Quit massiveH (Quit: Leaving) 07.33.12 *** Saving seen data "./dancer.seen" 09.33.16 *** No seen item changed, no save performed. 10.09.24 Quit PrinceKyle (Ping timeout: 250 seconds) 10.14.44 Join PrinceKyle [0] (~cancerian@cpe-2606-A000-111A-A097-0-0-0-F17.dyn6.twc.com) 10.18.50 Quit reductum (Quit: WeeChat 2.6) 10.29.48 Join lebellium [0] (~lebellium@89-92-69-72.hfc.dyn.abo.bbox.fr) 10.36.07 Join ZincAlloy [0] (~Adium@ip5f5acea4.dynamic.kabel-deutschland.de) 10.41.11 Quit ZincAlloy (Ping timeout: 265 seconds) 10.41.30 Join dys [0] (~dys@tmo-112-255.customers.d1-online.com) 11.01.25 Join ZincAlloy [0] (~Adium@2a02:8108:9440:dfc:fd38:f945:cdc7:8a67) 11.28.37 Quit mauved (Ping timeout: 245 seconds) 11.28.46 Join mauved [0] (~q@2602:fff6:f:6713::4873) 11.33.19 *** Saving seen data "./dancer.seen" 11.57.04 Quit Natch (Ping timeout: 265 seconds) 12.00.17 Join Natch [0] (~Natch@h-112-130.A444.priv.bahnhof.se) 12.15.57 Join jdarnley [0] (~J_Darnley@d51A44418.access.telenet.be) 12.16.22 Quit J_Darnley (Ping timeout: 268 seconds) 12.27.31 Quit TheSeven (Ping timeout: 265 seconds) 12.27.44 Join TheSeven [0] (~quassel@rockbox/developer/TheSeven) 12.43.55 Join dandels [0] (~dandels@unaffiliated/dandels) 13.27.52 Quit ZincAlloy (Ping timeout: 276 seconds) 13.32.04 Join petur [0] (~petur@rockbox/developer/petur) 13.33.22 *** Saving seen data "./dancer.seen" 13.45.21 Join krabador [0] (~krabador@unaffiliated/krabador) 14.06.41 Join andris [0] (~andris@83.166.211.9) 14.07.06 Part andris ("Konversation terminated!") 14.07.44 Quit krabador (Ping timeout: 250 seconds) 14.07.57 Join andris [0] (~andris@83.166.211.9) 14.28.23 Join ZincAlloy [0] (~Adium@2a02:8108:9440:dfc:fd38:f945:cdc7:8a67) 15.33.25 *** Saving seen data "./dancer.seen" 16.09.16 # Arg. SH toolchain build fails on current debian (gcc 9.2.1) 16.09.18 Join J_Darnley [0] (~J_Darnley@d51A44418.access.telenet.be) 16.10.48 Quit jdarnley (Ping timeout: 265 seconds) 16.34.35 # ...same for m68k toolchain 17.09.13 Join igitoor [0] (igitur@2a00:d880:3:1::c1ca:a648) 17.15.45 Quit igitoor (Changing host) 17.15.45 Join igitoor [0] (igitur@unaffiliated/contempt) 17.22.47 # ...and arm-eabi... 17.26.43 # amiconn 17.27.27 # https://pastebin.com/4B8gXbZ9 17.27.27 # only thing that might mess you up is it is silent unless errors are present I got REALLY tired of having to parse errors 17.31.04 # <__builtin> Bilgus: push that to Gerrit? 17.31.09 # In what way is this different from current git? 17.31.30 # it works? 17.31.32 # lol 17.31.33 # Btw, it's always gcc that fails to build; binutils is building ok (albeit with warnings) 17.32.59 # __builtin, will do 17.33.29 *** Saving seen data "./dancer.seen" 17.35.43 # Bilgus: The only differences I can see is that (1) it logs to a file instead of stdot and (2) it gets MPC from a different source (but uses the same version) 17.36.00 # So how will this fix gcc build? 17.36.22 # * amiconn prefers stdout over logfile 17.36.28 # Less file clutter 17.42.14 Quit andris (Ping timeout: 250 seconds) 18.01.42 Quit ZincAlloy (Quit: Leaving.) 18.04.37 # hmm might not be the right one I remember making more changes last time 18.04.51 # Building mipsel-elf works. So that's the only freestanding toolchain that worked :( 18.07.26 # * amiconn should be able to transfer the (working) toolchains from his old box, but rockboxdev.sh definitively needs fixing 18.34.15 # amiconn try that one g#2273 18.34.16 # Gerrit review #2273 at http://gerrit.rockbox.org/r/2273 : rockboxdev.sh updated for newer versions of gcc host (don't push yet) by William Wilgus 18.40.32 # IIRC you'll need to use sudo to delete the old working dirs 18.41.35 # in ~/tmp/ 18.48.43 Join PimpiN8 [0] (~textual@2001:1c04:3305:b700:70d3:fa7d:328d:64e2) 18.57.54 # Sorry i'm not sure WTH that first file was maybe where I started and I grabbed the wrong one on accident I had to grab this one from backups so hopefully it has all the changes I made to get it compiled 19.09.35 Join Soap_ [0] (~Soap@rockbox/staff/soap) 19.12.22 Quit Soap (Ping timeout: 276 seconds) 19.17.59 Quit PimpiN8 (Quit: My MacBook has gone to sleep. ZZZzzz…) 19.31.23 Quit pR0Ps (Ping timeout: 265 seconds) 19.31.47 Join pR0Ps [0] (~pR0Ps@orlnon0608w-lp130-01-70-53-100-171.dsl.bell.ca) 19.33.30 *** Saving seen data "./dancer.seen" 20.29.07 Quit Natch (Remote host closed the connection) 20.44.10 Join ZincAlloy [0] (~Adium@2a02:8108:9440:dfc:19:7bac:ecfe:aa09) 20.46.41 Join Natch [0] (~Natch@h-112-130.A444.priv.bahnhof.se) 20.49.12 Quit ZincAlloy (Ping timeout: 276 seconds) 20.50.49 Join reductum [0] (~weechat@cpe-104-32-77-28.socal.res.rr.com) 21.30.16 Quit dys (Ping timeout: 245 seconds) 21.33.32 *** Saving seen data "./dancer.seen" 21.41.51 Join ZincAlloy [0] (~Adium@2a02:8108:9440:dfc:19:7bac:ecfe:aa09) 22.16.06 Quit dandels (Quit: WeeChat 2.6) 23.33.36 *** Saving seen data "./dancer.seen" 23.58.34 Quit lebellium (Quit: Leaving)