为什么MHC不允许CPU速度低于20兆赫？

我只是让我的小指脚趾沾湿了MHC，所以我不会试图帮助它。我邀请来自真正和谐体验的人的其他评论。然而，我可以向您保证，PIC32 MX795可以以低于20 MHz的速度运行。例如，用一个8 MHz的晶体，我得到10 MHz的系统时钟和外围总线时钟与这些（在其他配置语用）：80兆赫的PLL DI。VIDIN由8提供10 MHz系统时钟，DIVIG1给出相同频率的外围总线时钟。在我的PIC32以太网启动器工具箱II上测试了XC32版本1.43（裸金属测试项目，不和谐）。问候，戴夫



      以下为原文

   
 
I'm just getting my pinkie-toe wet with MHC, so I won't try to help with that.  I invite other comments from people with real Harmony experience.
 
However...
I can assure you that the PIC32MX795 can run at speeds lower than 20 MHz.
 
For example, with an 8 MHZ crystal, I get 10 MHz System clock and Peripheral Bus clock with these (among the other configuration pragmas):
#pragma config FPLLIDIV = DIV_2         // PLL Input Divider (2x Divider)
#pragma config FPLLMUL = MUL_20         // PLL Multiplier (20x Multiplier)
#pragma config FPLLODIV = DIV_8         // System PLL Output Clock Divider (PLL Divide by 8)
#pragma config FPBDIV = DIV_1           // Peripheral Clock Divisor (Pb_Clk is Sys_Clk/1)


80 MHz PLL divided by 8 gives 10 MHz System clock, and DIV_1 gives same frequency Peripheral Bus clock.
 
Tested on my PIC32 Ethernet Starter Kit II with XC32 version 1.43 (bare-metal test project, not Harmony).
 
Regards,

Dave

嗨，你使用哪种和声和MHC版本？它实际上可能是MHC XML时钟配置……的问题。



      以下为原文

    Hi,
Which Harmony and MHC versions do you use ?
It could in fact be an issue with MHC xml clock configuration...
Regards

哦，对不起！粉红色：忘记提到最重要的是：和谐1.09和MHC 1.0.92*）我想我也观察到了以前版本中的错误，但是在错误的版本中无法辨别。也许可以追溯到1.06岁？是的，它工作在10兆赫，在5兆赫，结束了。没有抱怨5兆赫肯定！*我将使用和声1.10，但Microchip拒绝公布它们相对于先前版本的更改。在1.10个简短的“几个微小变化”之类的描述中，对任何使用大脑和认真工作的人来说都是一种冒犯。你不能追踪你的变化吗？使用Git！尼克



      以下为原文

    Oh sorry! pink:
Forgot to mention the most important:
Harmony 1.09 and MHC 1.0.9.2 *)
I think I observed that error in previous version too, but can't tell in wich ones. Maybe it dates back to 1.06?
 
 
Yes, it does work with 10 MHz and at 5 MHz, it is over. No complaining about the 5 MHz for sure!
 
 
*) I would use Harmony 1.10, but Microchip refuses to publish what they changed in respect to a previous version. The description in the 1.10 brief "several minor changes" or the like is just an offense to anyone using his brain and doing serious work.
Can't you track your changes? Use git!
 
 
Nick

这是不正确的，微芯片拒绝公布他们的变化。发行简报不过是一个总结，却忽略了一些细微的变化。因此，如果只有一个小的变化在一个版本，这是一个正确的总结。因此，我个人从来没有真正的使用那些和声发布简报。如果你想要细节，读错了，而不是MPLAB和谐发布笔记和内容（V1.10）？它甚至是由主题构成的，还列出了已知的问题。



      以下为原文

   
It's not true, that Microchip refuses to publish their changes. The release brief was always just a summary, which left out minor changes. So if there are only minor changes in a release, that is a correct summary. Therefore, I personally had never any real use for those Harmony Release Briefs.
 
If you want details, what's wrong with reading instead the MPLAB Harmony Release Notes and Contents (v1.10)? It is even structured by topic and also lists the known issues.

这是不正确的，微芯片拒绝公布他们的变化。发行简报不过是一个总结，却忽略了一些细微的变化。因此，如果只有一个小的变化在一个版本，这是一个正确的总结。因此，我个人从来没有真正的使用那些和声发布简报。如果你想要细节，读错了，而不是MPLAB和谐发布笔记和内容（V1.10）？它甚至是由主题构成的，还列出了已知的问题。即使是“综合”的文档也有半真半假和误导性的陈述。例如，“USB设备堆栈已经用RTOS进行了有限的容量测试。”从我的测试中，通过FreeRTOS和Hyg设备通过Hyror配置器在启动时就崩溃了调度程序。这意味着它没有被测试，甚至是最有限的测试（它开始了吗？）会把这个评论改为“USB设备堆栈与FreeRTOS无关”



      以下为原文

   
It's not true, that Microchip refuses to publish their changes. The release brief was always just a summary, which left out minor changes. So if there are only minor changes in a release, that is a correct summary. Therefore, I personally had never any real use for those Harmony Release Briefs.
 
If you want details, what's wrong with reading instead the MPLAB Harmony Release Notes and Contents (v1.10)? It is even structured by topic and also lists the known issues.




Even that 'comprehensive' documentation has half truths and misleading statements.  E.g. "The USB Device Stack has been tested in limited capacity with RTOS."  From my testing the USB device stack combined with FreeRTOS via the Harmony configurator just crashes the scheduler on startup.  
 
This means it's untested as even the most limited testing (does it start?) would have changed this comment to "The USB Device Stack is not functional with FreeRTOS"

这是不正确的，微芯片拒绝公布他们的变化。发行简报不过是一个总结，却忽略了一些细微的变化。因此，如果只有一个小的变化在一个版本，这是一个正确的总结。因此，我个人从来没有真正的使用那些和声发布简报。如果你想要细节，读错了，而不是MPLAB和谐发布笔记和内容（V1.10）？它甚至是由主题构成的，也列出了已知的问题。抱歉的双重答复，但论坛被打破，不让我编辑我的帖子。我想补充的是，文档中没有涉及的另一个问题是，Bootloader打破了EBI/SQI映射的一个古老而持久的问题，这里的其他用户复制了Gt:HTTP://www. MigCHIP.COM/FUMMS/M91764 7.ASPXIT，这在我可以看到的发行说明中也没有记载。无论是EBI还是Bootloader。问题在于，用于协调的理由是，它意味着“无缝”地处理旧的应用笔记方法的问题，允许多个函数组合在一起。不幸的是，它只给人一种印象，它实际上可以像许多情况下那样做（例如RTOS +任何USB功能，或引导加载程序+任何外部映射的内存函数（我怀疑这是否包括图形？））它似乎并没有被测试，更不用说功能了。



      以下为原文

   
It's not true, that Microchip refuses to publish their changes. The release brief was always just a summary, which left out minor changes. So if there are only minor changes in a release, that is a correct summary. Therefore, I personally had never any real use for those Harmony Release Briefs.
 
If you want details, what's wrong with reading instead the MPLAB Harmony Release Notes and Contents (v1.10)? It is even structured by topic and also lists the known issues.




Sorry for the double reply but the forums are broken and not letting me edit my post.  I wanted to add that another issue not covered in the documentation is the year old and persistent issue with the Bootloader breaking EBI/SQI mappings, reproduced by other users here> http://www.microchip.com/forums/m917647.aspx
 
This is also not documented in the release notes that I can see under either EBI or bootloader.
 
The problem is that the justification used for Harmony is that it is meant to work around the problems of the old application notes approach by 'seamlessly' allowing multiple functions to be combined together.  Unfortunately it only gives the impression that it can actually do that as in many cases (e.g. RTOS+ any USB functions, or bootloader+ any external mapped memory functions (I wonder if this includes graphics?) ) it doesn't seem to actually be tested let alone functional.
 
 
 
 

不幸的是，最好的和最简单的方法是文件比较实用程序，如无法比拟的。改变列表通常是浪费词。它不涵盖什么是固定的或为什么。甚至有人改变了公共变量名。从V1.09到V 1.10的变化是相当的，因为没有太大的变化。它在维护中。而是原来的问题。获取V1.10和2.xx并尝试它们。你可以改变你的程序。



      以下为原文

    Unfortunately the best and easiest way is a file compare utility like beyond compare.
Either change list is usually wasted words.
It does not cover what was fixed or why. Or even is someone changed public variable names.
The change from v1.09 to v 1.10 is fairly some , since not much changed. It is in maintainance.

But to the original question.
Get v1.10 and 2.XX and try them. You can do it with out changing you program.

我同意。虽然发行说明中的变化清单比产品简报要好，但还远远不够完善。问题是非常不完整的。并不是所有的变化都能进入这个列表。有时甚至会出现一些关键的变化，但是至少这些变化笔记会让你对他们正在研究的部分和主题做一个粗略的概述。它仍然使我意识到一些问题，我会遇到这些问题。我经常在升级时使用Orthor框架文件夹中的文件比较工具，所以我知道我应该更仔细地测试我的功能的哪些部分。



      以下为原文

    I agree. Although the list of changes in the Release Notes is better than the Product Brief, it is far from perfect. Issues are very incomplete. And not all changes make it into that list. Sometimes even some critical changes are missing.
 
But at least the change notes give you a rough overview about the parts and topics which they were working on. It still made me aware of some problems, which I would have run into.
 
I often use a file compare utility by myself on the harmony framework folder when upgrading, so I know which parts of my functionality I should probably test more carefully.

那么，他们在哪里描述他们对前一版本的改变呢？无处可去！没用！与前一版本相比，他们可以简单地描述他们添加或修复的内容。我不打算对他们以前的版本有所不同。这是他们的工作，而不是许多用户的工作，浪费他们的时间试图弄清楚以前的废话是如何工作的，以及它现在可能有什么更好的工作。Nick。



      以下为原文

   
 
So where do they describe what they changed in regard to the previous version?
Nowhere! It is useless! They simply could describe what they added or fixed, compared to (at least) the previous version.
 
I don't intend to make a diff to their previous version. It is THEIR job and not the job of many users that waste their time trying to figure out how that crap previously was intended to work and how it now might kinda work better.
 
Nick

啊，好吧！我看了1.10的描述，看到他们没有解决这个问题，所以尝试一个不同的版本只是浪费时间，只是发现它不是固定的。或者说整个和声只是实验性的，用户应该只是用不同的版本试试运气。我敢打赌，他们也没有重新实现的USAT内关于Tx缓冲区空，一个字符等没有，不固定在1.9-2，尽管我报告了它。哦，好吧，“我们监视论坛”哈哈哈！尼克



      以下为原文

   
 
Ah, OK! I look at the description of 1.10 and see, that they didn't fix that issue.
So it would only be wasted time to try a different version just to find out it wasn't fixed.
Or is that whole Harmony just experimental and users should simply try their luck with a different version?
 
And I bet, they also didn't re-implement the USART INTs in regard to Tx-buffer empty, one char etc.
No, not fixed in 1.9.2, despite I reported it. Oh well, "we do monitor the forum" hahaha!
 
Nick

哦，我忘了：50%以上的系统服务在和谐1.10是阿尔法或贝塔。不到50%是生产。和声2。x是完全beta。真的！巨大的进步！我想把它画在项目经理的额头上。反过来。所以他每天早上都能读。Nick



      以下为原文

    Oh, I forgot:
 
More than 50% of the System Services in Harmony 1.10 are Alpha or Beta. Less than 50% are Production.
 
Harmony 2.X is completely Beta. Wow! A huge improvement!
 
I'd like to tatoo that on the forehead of the project manager. In reverse. So he can read it every morning.
 
Nick

2.0x将很快被释放，V1.1x将在50%阿尔法和beta中死亡。仅仅因为它是或不在音符中并不意味着它在V1.10中没有改变，但是你是我第一次可以用这个抱怨来回忆。既然如此，你可能需要手动设置振荡器选项。这是一个耻辱，那就是O。NE的和谐的好部分。



      以下为原文

    2.0X will be Released soon and V1.1X will die in 50% Alpha and Beta.
 
Just because it is or isn't in the Note does not mean it has not changed in V1.10.
But you are the First I can recall with this complaint.
 
That being the case you may need to manual set the Oscillator options.
It is a shame, that is one of the Good parts of Harmony.

那么，他们在哪里描述他们对前一版本的改变呢？无处可去！没用！与前一版本相比，他们可以简单地描述他们添加或修复的内容。他们在第3页到第9页，“添加和更新”栏目中描述了它们。正如你所看到的，对于V1.10，该列大多是空的，除了两个点与图形库和两个点与DMA。对于其他的和声版本，该栏目通常描述一个巨大的事物列表。是的，正如其他人和我自己所说的：不幸的是，这个列表通常是不完整的。但考虑到大多数空的“添加和更新”列，我预计V1.10中的变化不会太大。一方面，我希望从V1.09升级没有什么问题。另一方面，升级可能并没有真正的好处。我只想说：他们发布了他们在发行说明中所改变的内容（尽管列表既不完美也不完整）。属性“无用”更适合产品简介。但这可能是我个人的看法。



      以下为原文

   
 
So where do they describe what they changed in regard to the previous version?
Nowhere! It is useless! They simply could describe what they added or fixed, compared to (at least) the previous version.
They describe it at page 3 to 9, column "Additions and Updates". As you can see, for v1.10 the column is mostly empty, except for two points with the graphics libraries and two points with the DMA. For other Harmony versions, that column typically describes a huge list of things. 
 
And yes, as already said by others and also myself: Unfortunately that list typically is not complete. But given the mostly empty "Additions and Updates" column I would expect that were not many changes in v1.10. On the one hand I would expect little problems for an upgrade from v1.09. On the other hand there probably is no real benefit from the upgrade.
 
I just wanted to say: They are publishing what they changed in their Release Notes (although the list is neither perfect nor complete). And the attribute "useless" fits more likely to the Product Brief. But that might be my personal view.

再次（我固执）。他们在我用的图书馆里什么也没变。但是，我还是要尝试一个新的版本，因为他们可能已经改变了他们害羞的东西（？）承认所以我切换到一个新的版本，我的程序不再工作，我没有线索在哪里开始狩猎。-不要跟我在一起！或者他们的变化报告是完整的，我没有看到任何改进，我使用。没有理由切换。Nick



      以下为原文

   
 
Again (I'm stubborn). They changed nothing within the libraries I use. But still, I'd have to try a newer version, because they might have changed something they are shy (?) to admit. So I switch to a new version, my program no longer works and I have no clue where to start hunting. -> Not with me!
Or their change report is complete and I see no improvements in anything I use. -> No reason to switch.
 
Nick