In-depth understanding of Android: Wi-Fi, NFC and GPS [excerpt]-Appendix (discussions and thoughts with reviewers)

Thank you for your patience. This book is expected to be available in middle March. From today on, I will include some of the content of this book in my blog. Note that the original article is serialized without being edited by the publisher, so it may look a little non-professional.

In the appendix, Mr. Wu Jinliang, one of the authors and reviewers, discusses the positioning and learning methods of this book. I believe that these discussions can resonate with readers.

The appendix contains several discussion emails from the author and Mr. Wu Jinliang, which are slightly modified here.

[--> Email from the author to Mr. Wu Jinliang]

Brother Wu:

I have several questions to discuss with you in depth, that is, the positioning of this book:

Let me first talk about my views:

1) For Beginners (I .e. those who have no experience in wifi, nfc, or gps), this book must be an entry book, but it is more difficult than the entry book in the general sense.

2) This book is also suitable for intermediate scholars who have been positioning for 1-2 years or who have experience in actually changing bugs but lack global understanding. However, some content may be relatively simple for them. In addition, our statistics show that NFC and GPS have very few problems. According to the interview, it is also important to know about The datasheet of the NFC chip (GPS should be without such public information. However, this book does not consider NFC, GPS, and Wi-Fi HAL. On the one hand, I feel that the Wi-Fi driver layer is very closely integrated with the protocol. readers with a bit of research spirit can do this by adding some driving experience on the basis of this book. The future development trend of the nfc hal layer may be the same as that of wpa_supplicant, that is, there will be no NXP, and Broadcom is too specific. GPS is generally unlikely to let outsiders see the driver code, right? I have read the HAL code of GPS in the QC open-source project Codeaurora [①. It is in the C/S architecture, only the Client content, and it is simple to send some commands, and then receive replies, there is no core thing.

3) for senior scholars, such as those who have relatively good experience and theoretical knowledge, the only advantage of this book may be that it can be used as a reference book, but the content will be relatively simple.

In addition, after communicating with Eva yesterday, I felt that this book did not have much practical experience. I am positioning this book by myself. I still want to get through the whole knowledge. If I have practical experience, I need to talk about it in theory and practice. At present, many engineers only have practice, no theory, or less attention. On the other hand, if we talk about practice, this kind of book is less valuable, because it is too low to operate. Unlike Network Management books, it can be done step by step.

This is my opinion on the positioning of this book. Brother Wu, can you tell me how you feel from front-line engineers, new recruits, and other aspects? The shortcomings are also raised.

Finally, after writing this book, I feel like the core of Wi-Fi, NFC, and GPS is the chip manufacturer. The only thing we can do is to modify bugs and accumulate practical experience, there seems to be very few places to play (there is still much room for play in the nfc ce mode, especially for Secure delivery solutions ). What about this problem, Brother Wu?

We sincerely welcome Brother Wu's speech!

Br

Deng fanping

[--> Reply from Mr. Wu Jiliang]

Brother DENG:

1) As Deng Xiong said, this book will have their own gains for scholars with different depth of knowledge. It can be used as a beginner and provide intermediate scholars with clues to problem analysis, the opportunity for senior scholars to think about knowledge (comparing their own understanding and supplement their knowledge) can be used as a reference for the book. The amount of data collected by different readers depends on the individual, readers who think more can also learn from the book the ideas of Deng's problem analysis and reflect on how to improve their search skills.

2) This book focuses on theoretical analysis and has no specific answers, but I think it is enough. This is not a q & A book. WiFi, NFC, and GPS are the three major parts of the book. Android involves A wide range of major support knowledge. Readers can perform selective in-depth analysis, each individual has different pursuits for knowledge points and is difficult to meet everyone's needs. Personally, I will be interested in android Wi-Fi sleep strategy and Location Network Positioning, this is related to problems encountered in actual work.

3) "NFC and GPS problems are very few", which may be related to whether the function modules are widely used and the application breadth. If more modules are used, more problems may be exposed, many application scenarios will also lead to the development of functions, which will naturally lead to new problems. Gps hal code is not provided by all manufacturers. Broadcom, MTK, and RDA only provide so, which may involve core technologies. It is estimated that some Command implementations are involved. GPS is generally a UART interface, UART is only responsible for data communication between the upper layer and the module.

4) for new users in charge of wireless modules, I arranged for them to do a functional verification test first, so that they can gain a deeper understanding of the functional points in the test, know what is prone to problems, and then let some modules that have been debugged with OK debug to familiarize themselves with the work required to debug a module; in the end, it will gradually assume some actual tasks. The learning arrangement is: learning the analysis of Module Interface drivers such as NL802.11, USB, SDIO, UART, and I2C, and then learning from the kernel to Android, such: wi-Fi driver-> netd-> wpa_supplicant-> HAL-> framework, Android will arrange learning of some core knowledge points, mainly to clarify the working mechanism. In the end, we hope that new users can have an Android network structure and draw it out.

5) because it is necessary to ensure that the relevant Peripheral modules can be used with the master node, this also determines that the focus of the common wireless work will be on the module porting debugging, involving a lot of kernel driver debugging, at present, android is becoming more and more perfect, with few major problems and few minor problems, but it is better to solve the problem. (Wi-Fi on android 4.4 has tested several bugs in the native code, the most serious one is that the socket created by supplicant is not closed when Wi-Fi is disabled, and is created every time Wi-Fi is enabled. When the number of socket opened exceeds 65536, ).

6) wireless module Wi-Fi, BT, NFC and GPS, the core technology is in the chip factory, and in the chip design, driver writing is only a small part of it, even for a relatively complex Wi-Fi Driver, it takes two or three people to write the driver for two months. I also agree that there are few places to play, unless new features are developed from the application perspective or some functional innovations are made. However, from the perspective of work, it is not easy to support these wireless modules. It is like making a new Wi-Fi system, the system needs to be stabilized, and no bugs can be reached the standards for mass production. This will usually take up to two months. Although there are only a few places to play, the demand for technical personnel who are convenient is still quite large.

Br.

Wu Jinliang

[①] Https://www.code=ra.org/, you can download the code of high-definition Reference Design (QC Reference Design.