defragster
Senior Member+

LFS stores a signature at the start. If that is missing then it will be created 'new'. So if upload format ( in the mandatory minimum area ) where code storage is not needed is used for LFS then losing those first bytes...

In order to commit data on LFS it must finalize a write once block - then that block has to be reformatted. Any updates are done to a freshly formatted block. Doing that 50 to 100 times per second the update to...

As far as p#23 - doing the formatUnused process 'at some non time critical point' will prevent the format on use slowdown as the dirty blocks have to be used again. Also if files are migrated off and deleted, that...

If I understand what you are hitting ... LFS on FLASH does have drawbacks. Each committed update has to make an updated DIR entry - which leaves behind a dirty set of flash blocks from the prior DIR info. So limiting...

good to find a non .ld solution. Edits to that .ld are non-trivial and have systemic ties

Suspect that concern was resolved with the attachInterrupt() code in that file as it has been in use for some long time now without issue ... the problem with comments ... ??? > no comment, no clue beyond what the...

This thread pjrc.com/threads/69185-FLASHMEM-for-all-functions-in-a-file Has tips on .LD file edits that keep ALL code in FLASH - and notes on pulling select code to RAM.

yield() is the one 'runtime' feature that goes with Arduino. When called directly, and called with calls to delay() and on each exit of loop() before it is called again. What yield does is check an existing...

Yes, that should be the case - with the exception of some conditional and instruction time. While in FAST I/O mode all pins route through a single common interrupt routine given the 1062 MCU design. There is then...

Putting "FLASHMEM" as a prefix to the function prototype will have it reside in FLASH. This example from PJRC's Startup.c for example: FLASHMEM void configure_external_ram() { ... } Start with any one time...

Seems details on the hardware controllers, motors and connections may be needed for related answers. Wondering: If Master, Slave, DC unpowered:: > Plug in DC motor power: results? > Then power Slave Teensy:...

Detail dithering catches me at times too - I did take a minute to find the 'quote from code' ... millis() and micros() are used for each of the respective 'elapsed' types. Indeed T_4.x uses a slower timer for...

It does not rely on any timers, but "class elapsedMillis" uses millis() for time reference: ... elapsedMillis(void) { ms = millis(); } ...

Mike, are these (other) p#1299 edits in place for SD code: pjrc.com/threads/68139-Teensyduino-File-System-Integration-including-MTP-and-MSC?p=308206&viewfull=1#post308206 ... I got Win_github to sync - but not...

@kurtE - Nice - will look at the usage and see if it helps MakeFiles - except that one wanted to pick a single 'DISK' for testing.

Does the SD card work right on a computer? Or is it really damaged? Some T_4.1's with many card swaps here in testing and no such issues yet as far as the SD Socket failing - though I have an sd card or two in...

Yes, that part of p#2was correct: "will provide generous Serial output of any events as they happen" With that enabled, remove the HUB and try a device individually and that should then show since the device connects...

Below is the code that resolves micros() given the last milli second timer update, and the cycle count stored then. If the math were bumped to all 64 bit and the scaling to nano adjusted from...

No, I use android - but not much for posting. Also not noticed a trend of that. At this point this old BBulletin forum is basically 'what it is' - ideally PJRC was looking to move to a new platform - but time/work...

Post restored ...

See this: github.com/TeensyUser/doc/wiki/Implementing-a-high-resolution-Teensy-clock Teensy has a processor clock based count of cycles: ARM_DWT_CYCCNT At 600 MHz it overflows 32 bits after 7+ seconds. That...

Yes, sorry - indeed, in that .h file : ...\USBHost_t36\USBHost_t36.h // Uncomment this line to see lots of debugging info! //#define USBHOST_PRINT_DEBUG

That looks to be an unpowered hub? With T_4.0 connection wires had to be soldered to the bottom Host D+ and D- pins, those lines have to be done well. Good sign both devices work individually. The T_4.0 unlike the...

No complaint on the track width - one more PCB than I have done :) Was just noting so many traces that it didn't seem like many were unused, and not all were visible in images. Good note on what is available from...

Very cool. Edited p#1 link to work. What few pins are unused? Though PCB trace space looks tight to pull many out anywhere.

That's how I read it as AFAIK it would make sense in context to help with sorting FS coding issues.

@luni I started a reply like yours ... if not there then #else bring it in, but did a refresh and saw the @WMXZ note that the has_include doesn't seem to perform the desired test?

That is a good ref @WMXZ : gcc.gnu.org/onlinedocs/cpp/_005f_005fhas_005finclude.html#g_t_005f_005fhas_005finclude @Kurt - hopefully this link goes there as used that is not portable when has_include isn't supported ...

Updated Defragster/T4LockBeta/tree/main/MakeFiles to include features and functions of SD_MTP_debug.ino Edited 2nd SD to only ADD when this works: if (sdSPI.begin(CS_SD2)) All code from SD_MTP_debug.ino at bottom...

Found it: diff in SD versus LFS_Flash mkdir() UPDATED >> github.com/Defragster/T4LockBeta/tree/main/MakeFiles When a folder in a folder is needed {and neither exists} the SD FS will create both "ManyF" and subfolder...

My Bad for not checking enough jumping between LFS Flash and SD ... > just off a 2+ hour telecon and got my outside project moved along ... There must be something unique in my 'string' presentation in that and...

There was a note way back about depth support on 'some media' : perhaps mis-remembered, or it changed ... in scanning it seemed others had depth, but not seen just then. Though still the same 'creation' works on SD...

@KurtE: would having an 'Unmount/Safe Remove' option allow closing disk files and then pulling the MTP drive and doing MTP.reset be a cleaner or good option? Cool there was an answer! The Off by 1: maybe the...

This is with 'b'ig create - not designed to work on LFS as written. Updated with added process debug notes noting fail to create DIR. Detected fail to create DIR and exited without streaming cascading failures....

This isn't right - in moving path define this one got mismanaged: Works - wrong on SD - should be making unique files in a folder - not re-using file name: ================================ Media...

Problem was using LFS FLASH that was too small. With command 'b'ig? > except testing here used size is only 9MB of 64MB ???? Back on 64MB LFS Flash I see those errors - that do NOT repro on SD using the same 'b'...

@mjs513 : That was running on 16MB FLASH: Media Size=16,777,216 ================================ Media Size=16777216 Used Size=16777216 us=112552 KB/sec=34706.18 Bytes written=4000000 It ran out of ROOM! ...

Mike, ... Opps - seems it was fixed here and working - just not updated. :rolleyes: github.com/Defragster/T4LockBeta/tree/main/MakeFiles TESTING: 't', 'b', 's' cmds results in no create errors and no 'v'erify...

@KurtE - okay will got back to MAIN - misunderstood that code was edited since that last update and newest was that linked branch. Darn .. I think I saw that the other day and torn with other tasks - I didn't get...

Need to go to other task ... But glancing at github it looks like this is the updated branch with changes?: github.com/KurtE/MTP_Teensy/tree/try_auto_detect_sd_media

Odd not repro at all? >> will hope to have time to reproduce with updated code. The file set on SD card was 'probably': makefiles(2206): but unknown cmds :( > should create and add to a "mfLog.txt" for tracking...

@luni got you to a solution! But what the suggestion in post #2 was : replace ALL _isr code with just the indicated lines, not ADD the delay. At 1 MHz there are less than 600 cycles available - and then adding in...

just to see ...Change: volatile uint16_t writeIdx; // ... void loop() { cli(); int16_t widx = (int16_t)writeIdx; sei(); //...

The USB connector on T_4.1 is the Device connector to HOST for programming, etc. If using the inboard USB_Host pins the the Teensy 4.1 (like the T_3.6) is the Host. Using a HUB allows multiple devices to connect...

Ran with the updated MTP__MAIN code and SD_MTP_DEBUG.INO 1: First without debug=2, pulled copy MTP to HDD, did 'c' and 'p' - no debug output of course 2: Built SD_MTP_DEBUG.INO with debug=2, pulled another copy MTP...

Thanks Kurt - meant to get that page linked ... shows 14% error on clock math @1200 baud for T_3.6 ???

May be a clock setting problem not tested to an external device at that rate? Can you cross Serial2 Rx/Tx to another port on the {same or other} T_3.6 and set them both up at 1200 baud and try send and receive on...

Sounds good Kurt ... will sync and run some.

Yes, if held low and released in the measured 13-17 second window, the PGM pin is wired in common with the Button, it would have the same effect.

IDE 2.0 often and regularly asks to save an unchanged file, even one the IDE presented on opening and never altered. KurtE noted this and has made an issue. Indeed that Defragster's install of Teensy package...

Back to this laptop - opened IDE 2 In board Mgr selected The Teensy 1.56.1 marked installed Clicked INSTALL and it proceeded to reinstall: ... Upgrading platform teensy:avr@1.56.53 with teensy:avr@1.56.1...

Thanks Kurt, it works - yes 27MB zip to download was fairly large! Nice to have it recognized and work seamlessly with TyComm again!

Thanks Kurt. It was late and I ran the sdmtpDebug and saw that and until I found the code it was too late to discover the logic behind it. And my head hurt after seeing 'skiped' on one screen and re-trying to find...

@KurtE > did @koromix publish an updated TyComm for PID_VID::MTP+Serial? I missed it is so ... if not can you point to one? - got the updated code. Inserted SD card running the sd_mtp_debug not sure what this is...

Very cool Kurt, will see about updating makefiles to allow it - and the debug. Maybe add a #define in MTP_storage.h to ID when that version of the lib is present ... so it can do the right thing when present or not...

Pushed Update to github.com/Defragster/T4LockBeta/tree/main/MakeFiles Added LFS options for: F - Format Disk Low Level S - SAFE Unused block Format The main write loop of the Verifiable data will exit on...

This is exactly what .formatUnused() does. It does take 'user code' "during Down Time ... or before Active Time" when write speed needs to be optimal. Usage was included in the integrity test code for LFS and I just...

That is puzzling that whole files got missed. That was showing before on SD card, but it went away. All recent SD copies had at worst a missing packet or so, but all files were present after MTP read from T_4.1. >...

Odd they did not copy ... Assume makefiles: 'l'ist shows them so they are properly findable and the "C"opy would put them on SD. Also 'v'erify. Done minimal testing on Flash - SD no fun transferring the card all the...

Should that be: " 1 directory of 5 files did NOT copy when I did the transfer from the flash to the PC" > CodeCompare looks to show them missing Don't forget the fun - but, perhaps useless here - given the files...

MTP Explorer in Win 11 select All Files copied from last MakeFiles.ino create copied to the PC from the SD card: 14.2 MB in 1540 files and 40 folders SD card starts empty and freshly ExFat formatted by official SD...

SD card corruption follow up. <EDIT>: This SD card is formatted ExFAT Emptied drive with "W"ipe > PC found no errors but added the 'Sys Vol Info" folder back SD into MakeFiles.ino > 'r', 't', 'b' SD returned to...

QUESTION: With recent code there was a better way to RESTART MTP to catch Teensy side changes? Currently MakeFiles offers this: case 'U': // USB Reset #if defined(USB_MTPDISK) || defined(USB_MTPDISK_SERIAL)...

Playing with MakeFiles.ino ROOT files count 490 - wondering if their size of 48 was too small - bumped to 992 and renamed so they would trigger content size during Verify. To see it work ( Files that Verify are NOT...

Here using KurtE's Debug demo above and the Example_3_Simple... BOTH work properly - with the set of 490 files on the root. Check the end of the Console Build output - make sure there isn't an local old library in...

UPDATE - based on sleep and post in the 'fail to enumerate thread': Repeated test without LFS Program disk for mtpindex.dat storage and the old DIR and data confusion seems to be improved? >> This looks like...

@KurtE - hi - you're welcome. Indeed, using TD 1.57b3 or newer is important. Reading Kurt's post#7 suggested the LFS Prog disk isn't needed - which I didn't try at 4 AM ... Just rebuilt with: //#define...

TeensyDuino 1.57 Beta 3 just released. In this post above code was incorporated (mostly) as posted: pjrc.com/threads/68139-Teensyduino-File-System-Integration-including-MTP-and-MSC?p=307790&viewfull=1#post307790 ...

Post #1 not updated: TD 1.57 b3(+) and MAIN BRANCH >> github.com/KurtE/MTP_Teensy This 'user post' having an issue not repro when test case added to MakeFiles, suggested use of LFS PROGRAM for storage of mtpindex.dat...

Just did a mod to ongoing test sketch on above thread - not yet posted there or updated. Using IDE 1.8.19 with TeensyDuino 1.57 Beta 3 on a T_4.1 with 128MB SD card Writing 490 files on SD card root "Mrt_###.txt",...

If read correctly ... the 490+ root files pre-exist and have been collected over time? And when presented or viewed not all are presented. Looking at Repro in test code for thread where DEV work is in progress:...

Good luck @clinker8. Losing TyComm is a pain - having it take your whole machine offline is disaster. Wonder if the build or lib tools @koromix uses are causing the trouble after some change. Not sure where he is...

Will give EaseUS P.M. a look. Have the Pro edition and just running it now to move a HDD to a SSD. Not planning to jump on Ext4 right away given I have no linux boxes active.

Thanks, grabbed a copy - those disk* tools do suggest they'd work!

Are encoder interrupts going to be over 7 seconds apart? If not using the cycles64() is wasted effort, if neither triggers in that ~7.15 second windows - then the cycles64() will not update correctly. Having...

Strange indeed ... Win 11 Desktop and Win 10 laptop both downloaded that .tar. and extracted tools folder. Indeed that teensy_secure.exe is present on both, and is executable as extracted. So nothing seems to be...

win10 laptop unzipped ide 2.0.rc7, added board mgr link p#1 Installed the Teensy boards Opened example BlinkWithoutDelay and selected board T_4.1 and did Verify with no Teensy connected, maybe missed error popup...

Yes, that was the reason behind the second line in my crosspost. @KurtE - has there ever been cases of the RED blink after running started? p#1 says: "executes the code for around a second and then it stops...

Odd Windows doesn't seem to have a way to R/W ext4. Search shows some tools to read only or one to converts ext4 partition to Fat32 or NTFS So funny that Teensy can do ext4 I/O ... ???

Once enabled the ARM_DWT_CYCCNT ticks on each clock cycle of the MCU without pause. In doing the micros() based offset to millis() code - or other testing - I did something like. uint32_t anArray then in some...

Awesome indeed!

Indeed, using/adding libraries is way too common {even for the newest of 'noobs' the IDE was meant to support} and not having a clear path to using the 'desired one' is a showstopper. The PJRC console output for...

Crashing into T_4.0 bootloader is an odd sign, in fact first report? > is there a pattern/count with a pause in the RED bootloader LED blinking? Generally the first easy thing would be to look into 'CrashReport'....

Glad to point it out. It is of course the MCU's cycle clock counter and at 600 MHz has 600 times more resolution than micros() - and on T_4.x it is used to resolve micros() between millis(). It is generally very...

@markkimball: using uint32_t time0; and ARM_DWT_CYCCNT for timing should have even less error than micros(): // Benchmark for modifying GPIO6_DR time0 = ARM_DWT_CYCCNT; for(i = 0; i < 100000;...

The UNO is running at 5V? Is the Display producing 5V? Or switched to 3.3V where it can work? The Teensy 4 is not 5V tolerant and 5V will destroy it.

Dynamic allocation is with sm_alloc() given external QSPI RAM Some info here: forum.pjrc.com/threads/63695-malloc-free-for-EXTMEM-and-DTCM?highlight=sm_malloc It has some smarts about maintaining the memory pool...

@KurtE - Cool - indeed, that does seem to match your post #196: Missed reading that as such, but had seen the sketchbook issue recently when a private lib took precedence. Wow - this is a complex process...

Not seeling files at groups.io/g/SoftwareControlledHamRadio ... is membership required to view?

@mjs513 - open a known SD sketch and build - then check console to see if it is using the right SD library for Teensy? If IDE 1.8.19 for instance links to a lib from 'sketchbook\libraries' it won't show examples for...