USB Over IP Documentation
in [Device Drivers]
|
Prev: Thomas F. Divine
Next: WinDDK is not building samples
From: anshul makkar on 4 May 2010 01:31 Thanks Pavel for support. After suffering so much of criticism , guilt and humiliation , I finally have someone who understood the true spirit behind me supporting this project. We have this kind of solution in market. This project (USBOIP) has a great appeal for masses, is good for my growth , is doable, then why should I drop it ? I pose this question to all the experienced people linked with the group. The answer that project should be dropped because "MS can't provide with the documentation/support or I will require lot of money for getting MS support" is unacceptable . I know I can implement it and I will do it. Thanks Anshul Makkar www.justkernel.com anshul_makkar(a)justkernel.com On May 4, 2:49 am, "Pavel A." <pave...(a)12fastmail34.fm> wrote: > Don, > but this is exactly what I've complained about: after so many years of > Windows > on the market, how come that they still does not feel confident to make > a commitment? And no matter how they explain the lack of commitment, > they are going to be blamed. > > Besides of this, commercial USB over LAN products already exist, so the OP > knows for sure this is possible. He would be a bit late on the market, > though, > but I really like his attitude -- MS has their own plans, we have ours, and > we'll succeed, with them or without. > > Regards, > -- pa > > "Don Burn" <b...(a)stopspam.windrvr.com> wrote in message > > news:E5AB73A81D714D4B82E9CF70757C38C9(a)Destiny... > > > If Microsoft say YES to giving you the documentation they face the > > situation of locking down an interface. This is a challenge that most > > OS vendors of the last 50 years or so have faced. Microsoft may feel > > they do not have a stable interface for stack, remember right now this > > is a contract between Microsoft drivers only. > > > Yes you can reverse engineer it, but do not blame Microsoft for the > > problems of your trying to reproduce an undocumented interface that can > > change out from under you at any bug fix or update. THE DECISION TO > > MAKE A DRIVER THAT IS UNLIKELY TO EVER BE STABLE IS YOURS, AND A > > PROFESSIONAL RECOGNIZES THEIR RESPONSIBILITES. So if you persist in > > trying to blame Microsoft, I can only assume that you are someone whose > > firm should be avoided at all costs. > > > Don Burn (MVP, Windows DKD) > > Windows Filesystem and Driver Consulting > > Website:http://www.windrvr.com > > Blog:http://msmvps.com/blogs/WinDrvr > > >> -----Original Message----- > >> From: anshul makkar [mailto:anshul.makkar.maill...(a)gmail.com] > > >> Don, If Microsoft says NO, that can't stop someone to do what one > > wants to do. > > >> What will be the end result , MS will again suffer when substandard > > drivers > >> (built due to lack of support) will be running along with Windows > > which may > >> lead to problems/crashes in OS. > > >> Thanks > >> Anshul Makkar > >>www.justkernel.com > >> anshul_mak...(a)justkernel.com
From: anshul makkar on 4 May 2010 01:38 Thanks Doron for this alternative approach. I will look into this and will get back. Thanks Anshul makkar www.justkernel.com anshul_makkar(a)justkernel.com On May 3, 9:58 pm, "Doron Holan [MSFT]" <doron.ho...(a)online.microsoft.com> wrote: > do not create a virtual host controller, just write your own hub driver. > that way you do not have to deal with reverse engineering an interface > between the port and miniport, you just have to implement the public URB > interface and all of the various IOCTLs and QIs. no small task either > though. > > d > > -- > > This posting is provided "AS IS" with no warranties, and confers no rights. > > "anshul makkar" <anshul.makkar.maill...(a)gmail.com> wrote in message > > news:a9ae2b39-c5b8-4883-9aaf-6e0556c3c086(a)6g2000prg.googlegroups.com... > > > Hi, > > > As per the documentation available (a single PDF file) to implement > > USBOIP we have to implement a virtual host controller driver . > > > - Virtual Host control driver will be equivalent to USB Host > > Controller driver and will be responsible for handling of URBs. > > - USBs will be converted to USB/IP packets and sent to remote machine > > - At remote machine another application will extract URBs from USB/IP > > packets and submit them to local USB drivers. > > > Now according to earlier postings and other documentation , Microsoft > > does not provide any support / documentation for writing host > > controller driver. I may have to do lot of reverse enginnering to > > decipher the functionality of USB Host Controller driver. > > > So just wondering whethe the above design approach will be correct, > > considering lack of support/documentation from Microsoft. > > > Is there any other design approach that can be suitable ? > > > Please share your views and experiences. > > > -- > > Thanks > > Anshul Makkar > > justkernel.com > > anshul_mak...(a)justkernel.com
From: Chris on 4 May 2010 11:23 On May 4, 12:31 am, anshul makkar <anshul.makkar.maill...(a)gmail.com> wrote: > We have this kind of solution in market. This project (USBOIP) has a > great appeal for masses, is good for my growth , is doable, then why > should I drop it ? I pose this question to all the experienced people > linked with the group. No, it doesn't have a great appeal for "masses". Facebook has mass appeal. USB over IP, on the other hand, is a niche product that will get you a handful of customers at most. And they'll inevitably find some device that won't work with your solution, and you'll spend a ton of time trying to mess with it, and then ultimately refund the money because of some stupid timing issue or undocumented call that the client driver is making. Think about it, are you prepared to support: - OHCI add-on controller cards? - PCMCIA USB cards? - Win2k? - Vista and Win7 64-bit? - all the URB and hub IOCTLs, documented and undocumented? - misbehaving client drivers that will try to evade your solution for fear of BSOD'ing some antivirus or Nero filter driver, etc.? - external hubs? what about three hubs chained together? - VMs? > The answer that project should be dropped because "MS can't provide > with the documentation/support or I will require lot of money for > getting MS support" is unacceptable . No it isn't. > I know I can implement it and I will do it. You know you can implement it? Then why are you even asking if it can be done and if so how?
From: anshul makkar on 4 May 2010 21:09
- OHCI add-on controller cards? > - PCMCIA USB cards? > - Win2k? > - Vista and Win7 64-bit? > - all the URB and hub IOCTLs, documented and undocumented? > - misbehaving client drivers that will try to evade your solution for > fear of BSOD'ing some antivirus or Nero filter driver, etc.? > - external hubs? what about three hubs chained together? > - VMs? Yes, I have to support all these issues. > You know you can implement it? Then why are you even asking if it can > be done and if so how? Anshul : I am not asking whether it can be implemented or not. I am just asking the alternative approaches for its implementation. Thanks Anshul Makkar www.justkernel.com anshul_makkar(a)justkernel.com On May 4, 8:23 pm, Chris <chris.aselt...(a)gmail.com> wrote: > On May 4, 12:31 am, anshul makkar <anshul.makkar.maill...(a)gmail.com> > wrote: > > > We have this kind of solution in market. This project (USBOIP) has a > > great appeal for masses, is good for my growth , is doable, then why > > should I drop it ? I pose this question to all the experienced people > > linked with the group. > > No, it doesn't have a great appeal for "masses". Facebook has mass > appeal. USB over IP, on the other hand, is a niche product that will > get you a handful of customers at most. And they'll inevitably find > some device that won't work with your solution, and you'll spend a ton > of time trying to mess with it, and then ultimately refund the money > because of some stupid timing issue or undocumented call that the > client driver is making. > > Think about it, are you prepared to support: > > - OHCI add-on controller cards? > - PCMCIA USB cards? > - Win2k? > - Vista and Win7 64-bit? > - all the URB and hub IOCTLs, documented and undocumented? > - misbehaving client drivers that will try to evade your solution for > fear of BSOD'ing some antivirus or Nero filter driver, etc.? > - external hubs? what about three hubs chained together? > - VMs? > > > The answer that project should be dropped because "MS can't provide > > with the documentation/support or I will require lot of money for > > getting MS support" is unacceptable . > > No it isn't. > > > I know I can implement it and I will do it. > > You know you can implement it? Then why are you even asking if it can > be done and if so how? |