- 精华
- 0
- 帖子
- 666
- 威望
- 0 点
- 积分
- 720 点
- 种子
- 0 点
- 注册时间
- 2005-2-21
- 最后登录
- 2016-3-20
|
http://kakaroto.homelinux.net/20 ... the-hermes-payload/
PS3: Why I don’t like the Hermes payload
First things first, the title says “why I don’t like the hermes Payload” so this has nothing to do with Hermes himself. I don’t know him, I never spoke to him, so I don’t know what kind of person he is and so I have no opinion on him personally.
Now, I want to make some things clear, I’ve seen a lot of people criticize me for ‘bashing Hermes’, and many people seem to think that I’m trying to say “I’m better than him” or something. Also, it looks like I created some confusion with my comments from my previous blog post. So I want to apologize and make sure there is no confusion anymore :
When I said that the hermes payload is ‘dangerous’, people misunderstood me.. no it is not specifically dangerous for your PS3, it won’t brick it or anything, the only ‘danger’ there is, is that it could (in some situations) crash… then you’d need to reboot, that’s it.. so don’t freak out about his work being harmful or anything, because as far as I know, it’s not!
Some people also told me “give credit where its due”, and I want to do it, I’ve always given credit to people, every time I achieved something, I gave credit to those who helped me achieve it. I’m not looking for fame here (if I did, I would have announced PL3′s release 3 weeks ago when I created it) I’m just having some fun in my free time doing something that I like. Hermes did contribute some nice things, and I appreciate what he did, mainly he figured out how to fix the controller issues with some games, that was something very diffi*** to fix and I’m surprised at how fast he came up with a solution and it was a smart solution and all I can say is “good job”. The other stuff he did in his payload, I don’t like that much, and that’s what I want to cover in this post..
I may have been ‘harsh’, but I don’t see the point in trying to be diplomatic, I’m a programmer, not a politician. I don’t like his code, and I speak my mind, I’m being honest, and I’m not trying to criticize him without any reason, as far as I know, I’ve stayed respectful and that’s all that matters to me.
To those who don’t want to know about all the technical details, let me ‘conclude’ here by saying that if the hermes payload works for you, then good, use it, I’m not telling people to stop using it, I’m not saying that PL3 works better either, maybe his payload works better in some situations, maybe not, but overall, the user’s choice should always be “whatever works for you”. The PL3 initiative is about having a standard repository for payloads, and having a common code base for everyone to work on, so in the future, PL3 might evolve faster and have more features, or maybe it won’t, the thing to note is that it’s better for payload developers to base their work off PL3. But again, this is meaningless for most users, apart from maybe clearing up the confusion about all these payloads and nobody knows which one to use.
Also, I talked about PL3, which is a common repository for contributors to work on, people seem to have nicknamed it “kakaroto’s payload” or “kakaroto’s pl3″, but I never said it was my payload, PL3 is PL3, it’s not all my work, and if you look at the commit log, you will see that I’m not the only contributor to it. PL3 itself integrates patches and solutions provided by Hermes, Waninkoko and Mathieulh, I improved some of their patches to make sure it works better for non-3.41 firmwares, but it’s still credited to be their work. PL3 is not my payload, PL3 is a payload repository for everyone. Also, PL3 as a project is a repository containing multiple payloads (default one, development one, dump_lv2, dump_elfs, etc..).
PL3 is not perfect, nothing in the world is perfect, so it might have bugs, it might not work for some people, who knows what might happen. But I never said that it was perfect, so people should stop thinking that I said that. It’s written more cleanly, it’s better in terms of the infrastructure behind it, but that’s the only thing I can vouch for.
Also for those complaining about me adding a donate button to the blog post, I don’t see how that’s relevant, I’m not begging people for money (and I haven’t received any donations in the last ~3 weeks just so you know). If you don’t want to donate, then don’t, no reason to bitch about it. I’ve put the button there so that people who appreciate the work and want to donate something have a way of doing it. I asked for donations before because I needed to buy a PS3 for development, I already raised enough money to buy it, so I don’t need any more donations, so I’m not asking people for donation money anymore, as simple as that.
Anyways, here are the more detailed/technical explanations on the reasons why I don’t like his payload :
First, the code is not clean, it’s unmaintainable. The fact that he gives his source code in .rar files instead of a git fork is the biggest issue I have with it. And yes, that does not matter for users, it only matters for developers. The problem with his method of delivery is that you have no way of knowing what he based his code off, so it’s hard to figure out what he changed.. also, when you find out his base, and do a diff, you get a huge diff for everything he did, all in one shot, and then you have to reverse engineer it to understand what he patched. That’s complicated and annoying for developers! For those who follow my ***, you can see how many commits I do, I always like having “small commits” because each commit becomes independent, self explanatory and easy to review. It also makes things easier to integrate, if you want a specific thing, you just merge/cherry-pick that single commit, instead of copy/pasting code, and editing it to remove the clutter. The other reason I like git is that if he used it and I merge a commit from him, then the code stays credited to his name in the commit log, it allows me to have his code without ‘taking ownership’ of his work, it allows everyone to be credited for what they did, and I think it’s the first thing to have for an open source and community project.
The reason why I said his code could crash is because his payload got too big and couldn’t fit anymore in the allocated memory we have in the kernel (1296 bytes), so he decided to just move the code to a random position (0x7fff000 I think). This means that his payload will work as long as no application, game or kernel allocates memory which ‘randomly’ ends up in that area.. if it does, then the payload would get overwritten and the kernel will crash. The proper way of doing it (PL3 does that) is to allocate memory during the initialization of the payload, copy the functions we want in that memory that we own, and write those functions to be position independent so they would work no matter where they are placed in RAM.
Another reason is the way his syscall8 works, I tried to read his assembly and reverse engineer it, and I seriously was lost and couldn’t understand what was happening.. there are no comments (you’ll notice that my payload has a comment on almost every instruction), so how can I integrate his syscall if i don’t even know what it does… if at least it was on git, I could see the commit messages and understand what each chunk of code did, but he doesn’t use git, so…
The way he fixed the controller issue was also not very good, he patched two offsets to jump to a function that decides on some kind of enum on what response to return and you controlled that with his own system call 8.. why do something like that? it makes the fix dependent on people using this new syscall, and it’s useless when you can just patch it directly to return the right value.
I also didn’t like the fact that his code became a mess that is 3.41 independent, and it would have taken a huge amount of work just to try to make it work again on 3.15. I already spent time cleaning up the payloads and making them work for older firmwares, so why fork and write code that doesn’t integrate that, it just makes collaboration harder.
There’s also the whole syscall 35 versus 36 issue, but that has nothing to do with his payload since I added sc35 after he released his payload. It’s not about his payload being bad because it doesn’t support it, it’s simply about PL3 having a ‘superior’ (if I may say so) system call. What it means for users? nothing at the moment, maybe it will be used for doing fancier stuff later on, maybe you can map a game to your bluray and a different game to /app_home, that could be useful for users, but for now, it’s simply more flexible and cleaner code.
There are many other small things that I didn’t like, but it mostly just summarizes to “the code is not clean and it’s unmaintainable” and “he doesn’t use git”. Like I said, if you don’t care about that, then I see no reason for you not to use his payload. It doesn’t mean either that he’s not skilled, it simply means that he may lack experience in code sharing and experience in open source. But that doesn’t make his work any less valuable.
I hope this clears things up a bit. I criticized his work, said what I thought of it and people over reacted, I wanted to make sure people didn’t misunderstand me, and didn’t think I don’t respect Hermes for what he’s done already. Everything else is just drama and people trying to get attention.
If this post stirs up even more trouble, then so be it, I don’t think I have much more to say. I said what I think, people should take it or leave it. I do not however tolerate people insulting me for no reason at all. So please, criticize me all you want, just stay respectful.
Thank you,
KaKaRoTo. |
|
川公网安备 51019002005286号
楼主
发表于 2010-10-17 12:00 · 北京
