#11 Standard integers (stdint.h) and mixing integer types

[@classyjohn1923]

quick comment: 24:13 it would be better to just typecast to the lvalue datatype since if u16c > 32,767, you will get the incorrect value regardless in the implementation you showed.

[@StateMachineCOM]

Yes, there is a problem at 24:13. The correct version of the code is "s32 = 10 - (int32_t)u16c;" I have added an info card to the video at this time stamp and the code has also been corrected in the code downloads for this lesson ( www.state-machine.com/quickstart ) and on GitHub ( github.com/QuantumLeaps/modern-embedded-programming-course ). Thank you for pointing this out! --MMS

[@sandpaper-egg-2020]

I recently got hired by a company for a mixed programming position. I had NO experience in embedded systems programming whatsoever and was struggling to catch up. This course single-handedly brought me up to speed with quality and clarity unrivaled by any other KZbin training course (and I have watched quite a few). If you have just found this course, keep going. You have come to the right place. The fact that Miro Samek has posted this course for free is the height of scholastic generosity. I have learned more from this course than any computer-science college class.

[@birnodin]

This video should be viewed from every C-Programmer once a day!

[@RandallLowetx]

Once someone has learned from any of the other tutorials out there how to build something basic in C or C++, this series is where everyone should go to learn what's actually happening. So freaking clear and valuable.

[@leohuang990]

Great to know that in C, small integers in computation are converted to the platform's "int" instead of the left-hand side's datatype, and a signed integer used with unsigned ones is converted to unsigned int. Did not notice that stdint was in the C99 standard and ARM & RISV-V can support both endianness. A very informative lecture.

[@StateMachineCOM]

Karel, Here is the explanation for the promotion (from MISRA-C:2012, Appendix C): "An integer promotion conversion to signed int or unsigned int takes place whenever a bit-filed or object with (un)signed char, (un)signed short, or enum type is used arithmetically as described in C90 Section 6.2.1.1, C99 Section 6.3.1.1..." So, in summary, the integer promotion happens before the expression is evaluated (before the ~ operand in this case) and also before the result of the expression is balanced to the right-hand side of the equals operand. --MMS

[@ismaeelahmed]

Hello Dr. Samek, Thanks for these wonderfully taught lessons .. however still waiting for lessons on interrupt sections ..... Please add them as well... thanks

[@StateMachineCOM]

@@ismaeelahmed Interrupts are explained in Lessons-16,17,18 (kzbin.info/www/bejne/oIGUe6yjoa2IqsU ). --MMS

[@mx4evaNZ]

Just plain excellent. Great tutorials on both C and Arm. Thanks a lot

[@clemolk]

This course is simply great because it does not focus only on basic and standard topics but it instead explains very well several practical problems you can encounter in embedded programming. Thank you Dr. Samek.

[@KarelHeyse]

Can you elaborate on why u8a in "~u8a == 0x00u" (26:40 in the video) is promoted to (unsigned) int before inversion is applied, and not afterwards. I understand that in the case of "u8a == 0x00u" the left hand side is promoted to unsigned int to match the type of the right hand side.

[@miguelorrrr]

Seven years ago but jus in case you still have the question. 22:19 When you have an unsigned and signed operands, both are promoted to unsigned int and the result is unsigned int: 10 - 100 = -90 = FFA6 (unsigned and 16 bits of length)

[@MEKTOR900]

I'm really glad about this kind of courses help out to undesrtand thing that sometimes seems to be imposible when you are just reading a book, this lessons had been so useful, thank you very much for giving your time and dedication to others.

[@rubenspatterson]

The type of the result will be equal to the type of the operand that allows a larger maximum value, so the other operand will first be converted to a value of the appropriate type. It should be noted that in the C programming language, not a single calculation operation is performed on data smaller than int. (on MSP430 not smaller than 16 bit int)

[@theemelex]

Dear Dr. MS, I really have learned a lot from these lessons. I am a so-so embedded C programmer, not good but not bad either. I have been trying to learn C++ and OO concepts have me stumped. I'd like this same course in C++ for ARM. It would really help the learning curve a lot. Besides, I don't think anyone else's doing anything embedded related in C++. I would really appreciate something in C++. Thanks a lot for these lessons. theemelex

[@StateMachineCOM]

There are still many things to learn about C in this course, so this will remain the main focus. However, I do plan to talk about object-oriented programming and I will show how to achieve encapsulation, single inheritance, and polymorphism (virtual functions) in C. The way I will present these concepts is exactly how they are implemented internally in C++. In fact, I will compare machine code generated by the C++ compiler with the code generated by the C compiler. Again, my main guidance for this course is that when students see how the concepts are ultimately implemented, they will understand the concepts better and will be able to use them more efficiently and with greater confidence.

[@keithevans6450]

Great job Miro! You have probably saved someone hours or days of head scratching with this video. Also wanted to mention that I think your videos could easily be used in a high-school level course with appropriately motivated or mentored students. Even the BBC micro:bit can be programmed in C (among other things) and uses an ARM Cortex-M0. Fry's Electronics has been promoting it over the summer.

[@StateMachineCOM]

Yes, I think that interested high schoolers should be able to follow along and learn from this course. I'd also love to use something like the BBC micro:bit, but it does not seem to have a built-in debugger interface that would allow easy stepping through the code like I so frequently do in these videos. In this respect, BBC micro:bit resembles Arduino. You can easily download the compiled code, but you can't step through. If I were to select a board for this course today, I would probably choose one of the STM32 NUCLEO boards (perhaps a Cortex-M0+ part). NUCLEOs allow easy code upload (they show up as USB drives), but at the same time they allow single-step debugging. --MMS

[@devkhan1]

Very Informative , Desperately waiting for next in series lectures !!!

[@FRED82627]

Excellent course. Million thanks and expect the next.

[@김남규-k3k]

MISRA C standard guides some rules about implicit & explicit casting in integral expression. I've compling with that rules, however didn't fully understand what it requires. Thanks to this brilliant lecture, i can finally figure out purpose of the guidelines and differences of integral promotion in various machine.

[@basemanq]

When will you post the next tutorial? This course is just get started. There are still many other important topics to cover, interrupts, hardware timers, uart, i2c etc. This is a great tutorial. I have been using Keil uVision quite some time. I am always interested in IAR embedded workbench. The simulation tools are great assets in terms of debugging. If you can make some more advanced tutorials (hardware timeing, i2c bus) on using those simulation tools, that would be really helpful.

[@ElectronicsGate]

This was an important lesson. Thanks. Please keep doing this great series.

[@04mfjoyce]

Thank you for the detailed explanations! This is a fantastic series.

[@vivektodmal1]

This is hell lot of information in one video...very nice man keep it up

[@dongolahmed]

First, I really do appreciate the effort and time you put sir into these awesome and god-send lessons. Second, I wonder if you sir shorten the period between the lessons ... 'cause with this pace, it would take years to reach more advanced topics such as RTOS. Thanks sincerely,

[@lemon6521]

It did take years but we got there eventually!

[@MohammedNoureldin]

Waiting for the next lesson Dr. Semek, could it be soon released? I'm just excited to go on with it

[@StateMachineCOM]

The next lesson has just been posted on KZbin. Thanks for the encouragement to keep making these lessons :-) --MMS

[@MohammedNoureldin]

Thank's a lot doctor! :-)

[@Steve-gt5nb]

Thanks so much...This is excellent course...Keep up the good work!!!!

[@GalinaMalakhova]

:D Thanks so much for making this series! Very educational.

[@bhavaysen]

Hi @Quantum Leaps, LLC, Dr. Samik, In the example, where you demonstrated the comparison, there will be an issue. if ( (uint32_t) u32e > -1 ){ u8a = 1U; } else { u8a = 0U; } what will be the result when u32e = 0xFFFFFFFF ? This becomes a bug in the code, or there is some solution to this? please share and thanks for the lectures, these are too good and helpful.

[@MuhammadtahirRana-nn1cb]

God bless you sir.

[@devkhan1]

Eagerly waiting for Embedded Programming Lesson12 video

[@jaimef.sanchez7462]

Hi Dr. Are you taliking about Floating variables at any specific session?

[@dongolahmed]

Sir, when will you make lessons on serial communications (UART, SPI, I2C) ?

[@StateMachineCOM]

I plan to talk about communications, but this course is not there yet. People still need to learn about interrupts, for example. But to learn about interrupts, I need to explain the interrupt vector table. This, in turn, leads to explaining pointers to functions and the startup code. So, as you can see we have quite a few subjects to cover before we will be ready for communications. --Miro

[@ahmedwahdan672]

I saw this series about 10 times, and I get back to it to refresh my knowledge, Thanks a lot. Unfortunately, I get warning while downloading the code using Cortex Core m0 and the assembly code not changed either, it still uses the M4 instructions? where is the problem?

[@VuNguyen-ws6nt]

Is this the last tutorial in the series? Well, I would expect some more tutorials :D

[@shashankhebbar1921]

great video and in depth explanation. can you show us how to interface a sensor with microcontroller using i2c protocol.

[@hakuhakuji]

Why did we put a header definition in delay.c? Isn't it redundant since it is linked during compiling i.e. gcc -o -c main.c delay.c?

[@StateMachineCOM]

There are two files: delay.h header file contains the *declaration* of the function delay() and delay.c source file contains the *definition* of the function delay(). The declaration is just the function prototype and is needed for the compiler, not the linker. For example, the header file delay.h is also included in the main.c source file. This inclusion allows the code inside mani.c to call the delay() function, because the compiler needs to see the function prototype (declaration) in order to property call a function with all its parameters. On the other hand, the linker is not concerned with generating code for calling functions or any other code. Linker is only concerned with placing code and data into various memories (ROM, RAM) and with fixing the addresses. You might also want to watch lesson #14, where you will see explanation of the build process and the roles of the compiler and linker . --MMS

[@saurabhsoni5047]

Many thanks to you Sir for this excellent series on Programming in Embedded Systems ! This will always be referenced by the one interested in Embedded Programming. I have been experimenting on the mixing types you explained above. I ran into a doubt - was hoping if you could provide some hints. Many Thanks ! #include #include int main() { uint8_t __VAR = 0xFF; // Printing some values to give the reader an idea about the underlying system printf("%u ", sizeof(uint8_t)); // 1 printf("%u ", sizeof(uint32_t)); // 4 printf("%08x ", (uint32_t) __VAR); // 000000FF printf("%08x ", (uint32_t) -1); // FFFFFFFF // Doubt is the below code snippet if(__VAR > -1) printf("OOPS"); else printf("CHECK"); return 0; } This code finally prints OOPS and I wonder why. Since, 000000FF is less than FFFFFFFF, I was expecting it to print CHECK. Please note this is a similar code which you used in the video, just that I have used the datatype uint8_t and you have used uint32_t. Thank You Again for this amazing series of video lectures !

[@navinreddy88]

Sir, can you please explain the concept of storage of global const variables and local const variables.

[@StateMachineCOM]

The correct use of the 'const' keyword and the issue of 'const-correctness' of programs is a big subject, actually. I would have to devote at least one whole lesson to it. I will keep it in mind as a subject for the future. --MMS

[@kevinchu9601]

27:23 u8a will be promoted to int, do you mean unsigned int ? i'm a little confused about this. Thanks

[@StateMachineCOM]

All small integer types, no matter signedness, get implicitly converted to (signed) int when used in most expressions. But yes, the implicit integer promotion rules in C are murky and it's difficult to explain them simply. One good discussion of the issues is found on StackOverflow: stackoverflow.com/questions/46073295/implicit-type-promotion-rules

[@kevinchu9601]

@@StateMachineCOM Thanks very much. So what is the type of 0x00U? In my opinion, the type of 0x00U should be unsigned int, then u8a should be promoted to unsigned int. Am i wrong?

[@Nikita-wh2vc]

26:04 It should be noted that such explicit cast (unsigned integer to signed integer) is also dangerous, because it works incorrectly for unsigned integers greater than or equal to 0x80000000U. Obviously, the unsigned integer 0x80000000U is greater than -1 in common sense, but after explicit cast it becomes negative number -2,147,483,648. Maybe the better way is to make an explicit cast to int64_t

[@StateMachineCOM]

Mixing signed and unsigned representations is tricky, and its correctness depends on the context. Therefore, explicit casting is often necessary. Using a bigger integer (like your suggested int64_t) only kicks the can down the road, because precisely the same argument can be made at the end of the dynamic range of the bigger integer (0x8000000000000000). --MMS

[@ahsanalirafaq971]

can you tell me at 13:30 , how can you assign a 16 bit variable which can hold only 16 bit data a 20 bit data which is beyond its capacity ????

[@StateMachineCOM]

The values assigned to the u16c and u16d variables at 13:30 are within range of 16-bis. Please note that these are decimal constants: 40000==0x9C40 and 30000==0x7530. The problem is only later, when the two variables are added together, because the result (70000==0x11170) indeed exceeds the 16-bit boundary. --MMS

[@ahsanalirafaq971]

@@StateMachineCOM thank you very much

[@peteys2255]

Thank you for the interesting content on your channel! It helps a lot both beginners and advanced persons in Embedded Programming. I guess I misunderstood the following Code at 27:50 : Why will the comparison (~u8a == 0x00U) never become true? Without casting u8a will be promoted to int, but 0xFFU=255, and as signed int -1 respectively 1111 1111 binary (signed, 2er complement) or am I wrong? Therefore the operation ~u8a results in 0x00? I would look forward to a feedback :)

[@jongxina3595]

u8a (which is 0xff) gets promoted to 0x000000ff (8 to 32 bit promotion). therefore ~u8a will be 0xffffff00 which is not 0

[@StateMachineCOM]

The only thing that I'd like to add to Ben's explanation is that by C Standard, an integer must be at least 16-bit wide, so even in a 8-bit CPU, such as AVR, the promotion will result in 0x00FF. This means that ~u8a==0x00U is always false for any CPU type.

[@mingwong9233]

I'm a little confused here.. why is it that the 16-bit machine be able to perform a 32-bit calculation? I don't understand how the 16-bit machine can produce 70000 when it is bigger than 2^16 -1

[@StateMachineCOM]

Ming Wong 32-bit computations require more instructions on 16-bit CPU (or 8-bit CPU). In fact, you can see in the video that 32-bit addition requires two 16-bit additions plus taking into account any carry-over from the lower 16-bits.

[@bareminimum-mrtz]

Appreciate the effort 😀

[@amandeepbedi1280]

Really Helpful..:)

[@veraj5494]

Very helpful. Thanks

[@yashesvi-raina]

This was challenging.

[@liuliu-vo2tf]

厉害

[@kapkunal]

Thank you

[@jp_engineer]

It's the first time I attend this class. Thank you a lot for teaching. I'm using ARM_GCC_541 compiler and I have tested a code. Could someone help me? code: unsigned char a=9, b=10; int c = 0; case 1 -> c = (unsigned char)(a - b); //( this gives 255) case 2 -> c = (unsigned char)a - (unsigned char)b; //( this gives -1) case 3 -> c = (unsigned char)a - b;//( this gives -1) shouldn't all the case give 255 since I casted all of them?

[@StateMachineCOM]

In all cases you have the same promotion to `int`, regardless of the types of the operands. But only in the first case, you cast the *result* on (unsigned char). Therefore, only the first case produces 255. Because the type of the operands does not matter, you will get the same result when you declare int a=9,b=10. --MMS

[@muhammadmaruf1896]

Can anyone please explain, how does 0x70000 in 32 bit converted to 0x4464 in 16 bit?? (17:20 in video).....

[@StateMachineCOM]

First, this is NOT 0x70000 but 70000 decimal. Please check with the calculator that this 70000 is 0x11170, which truncated to 16 bits is 0x1170, which is 4464 decimal. --MMS

[@jamesmaxwell381]

Thank you Dr. Samek. About MISRA-C, there are discussion about whether it is actually helpful. Some people say that it is just a corporate standard for large organizations to force devs. adhere to a writing style and it is "The Way of Big Companies" doing things which may not be the best way of solving problems. They even point to a study done by TU Delft which suggests adhering to MISRA-C actually causes more faults (NOTE: I haven't read that paper myself though). What is your though on this? link to that paper: ieeexplore.ieee.org/document/4658076/

[@StateMachineCOM]

CMSIS is actually a set of several standards and components, including CMSIS-Core, CMSIS-RTOSv1, CMSIS-RTOSv2, CMSIS-Driver, CMSIS-DSP, CMSIS-Pack, CMSIS-Zone, etc. This video course uses only the most basic part (CMSIS-Core) for the CPU core and built-in peripherals. In my opinion, this part of CMSIS is not too bad and is actually helpful. The other parts of the CMSIS might not be as useful. --MMS

[@jamesmaxwell381]

I guess you've mistaken MISRA-C with CMSIS. I was asking about MISRA-C. (But I also had a similar question about CMSIS but didn't ask and I'm happy you answered that too :D)

[@wassimdhokar731]

Check course online to learn embedded C programming on arm cortex-m processor: www.udemy.com/course/embedded-arm-cortex-m3-programming-guide/

[@batuhanmete775]

everything is just perfect except the feeling that your breath in my ear at about minute 16. Just kidding thanks :)

[@joelinacio]

Only your videos don't play audio in my Chrome...in Firefox is ok.

[@StateMachineCOM]

What can I do about it? Do you have any constructive ideas? --MMS

[@joelinacio]

just to inform ... maybe someone has the same problem ... I'm trying to figure out what's going on with Chrome on my computer.

[@bluehornet6752]

Joel Inacio Works just fine for me here in Chrome. I'm watching it and posting here now from Chrome. No problems here.

[@matthewszklany101]

Has anyone else gotten the error "Error 65: access violation at 0x400FE070 : no 'write' permission" when trying to run the code in the debugger? It seems to have something to do with line 587 in system_TM4C123.c which is: SYSCTL->RCC2 = 0x078027810 I don't think I observed this when I had the board connected, rather I think it may have something to do with the simulator?

[@StateMachineCOM]

Projects that involve real peripherals, such as the GPIO, generally don't run in the simulator. Perhaps the simulator could be configured to accept reads/writes to the address range starting at 0x40000000, but I haven't investigated it. For these lessons, please just use the real board. --MMS