Development mailing list

Syndicate content
Archive of posts for haiku-development at FreeLists
Updated: 1 hour 28 min ago

[haiku-development] Re: replacing runtime_loader

Mon, 2016-04-04 15:45
On16.03.201616:07,TimKellywrote: IngoWeinholdwrote: Thefileshouldbeatthepackageroot.Pleasealsocheckthesyslog whetheryourpackagehasbeenactivated. ~/tools/pcc-current/packages/rtlpackagelistruntime_loader.hpkg ...
Categories: Development

[haiku-development] Re: _kern_reserve_address_range

Mon, 2016-04-04 15:45
Ok, as Adrien suggested, the problem appears to occur when two child processes of a parent process request the same exact address (displaying relevant parts of strace -t): [ 1044] _kern_normalize_path(/boot/system/non-packaged/lib/cpp, true, 0x72fee2f8) = 0x00000000 No error (21 us) [1044]_kern_read(0x4,0x0,0x72fed2c4,0x34)=0x00000034(4us) [1044]_kern_read(0x4,0x34,0x72fed2f8,0xa0)=0x000000a0(2us) ...
Categories: Development

[haiku-development] Re: _kern_reserve_address_range

Sun, 2016-04-03 11:45
Followinguponthis, system/kernel/vm/VMUserAddressSpace.cpp VMUserAddressSpace::ReserveAddressRange(size_tsize, constvirtual_address_restrictions*addressRestrictions, uint32flags,uint32allocationFlags,void**_address) { //checktoseeifthisaddressspacehasenteredDELETEstate ...
Categories: Development

[haiku-development] Re: replacing runtime_loader

Sun, 2016-04-03 11:45
IngoWeinholdwrote: The file should be at the package root. Please also check the syslog whether your package has been activated. ~/tools/pcc-current/packages/rtlpackagelistruntime_loader.hpkg package-attributes: name:runtime_loader summary:modifiedruntime_loader ...
Categories: Development

[haiku-development] Re: replacing runtime_loader

Sun, 2016-04-03 11:45
On16.03.201614:06,TimKellywrote: IngoWeinholdwrote: Youdon'thavetoreplacethecompletehaikupackage.Youcanbuilda packagethatcontainsyourruntimeloaderanddeployitin /system/packages.Whenyourruntime_loaderfile'smodifiedtimeis newerthantheoneofthesystempackage,itwillreplacethelatter. ...
Categories: Development

[haiku-development] Re: replacing runtime_loader

Sun, 2016-04-03 11:45
IngoWeinholdwrote: You don't have to replace the complete haiku package. You can build a package that contains your runtime loader and deploy it in /system/packages. When your runtime_loader file's modified time is newer than the one of the system package, it will replace the latter. In the package, where do I put runtime_loader? I have tried at the top level and within a folder system, but neither approach appears to ...
Categories: Development

[haiku-development] Re: replacing runtime_loader

Sun, 2016-04-03 04:45
IngoWeinholdwrote: You don't have to replace the complete haiku package. You can build a package that contains your runtime loader and deploy it in /system/packages. When your runtime_loader file's modified time is newer than the one of the system package, it will replace the latter. That sounds good. I might as well get started on learning how to make a package anyways. Thank you to waddlesplash and Adrien for your suggestions. ...
Categories: Development

[haiku-development] Re: replacing runtime_loader

Sat, 2016-04-02 12:45
On16.03.201601:23,TimKellywrote: Again,thankyoutoeveryonethathascontributedinformationand knowledgeinresponsetomyquestions.Iwouldliketoreplacethe stockruntime_loaderwithonethatreturnsthespecificerrorcodefrom _kernel_reserve_address_range.However,thisdoesnotappeartobe trivial.I'vecompiledruntime_loaderwiththechangeIwouldlike,but Iamnotabletodiscernhowtoreplacethesystemonewiththecompiled one.I'dprefertoavoidbuildinganentireimageandreinstalling.Is ...
Categories: Development

[haiku-development] Re: replacing runtime_loader

Fri, 2016-04-01 06:45
OnMar16,20162:14AM,AdrienDestuguespulkomandy@xxxxxxxxxxxxx wrote: that,butyougetitasabyproductifyoubuildanimage) IIRC,jamhaiku.hpkgworks. -waddlesplash ...
Categories: Development

[haiku-development] Re: replacing runtime_loader

Fri, 2016-04-01 00:45
OnTue,Mar15,2016at08:23:19PM-0400,TimKellywrote: Again,thankyoutoeveryonethathascontributedinformationandknowledge inresponsetomyquestions.Iwouldliketoreplacethestock runtime_loaderwithonethatreturnsthespecificerrorcodefrom _kernel_reserve_address_range.However,thisdoesnotappeartobetrivial. I'vecompiledruntime_loaderwiththechangeIwouldlike,butIamnotable todiscernhowtoreplacethesystemonewiththecompiledone.I'dprefer toavoidbuildinganentireimageandreinstalling.Istheresome documentationonreplacingindividualsystemfiles(notadd-ons)? ...
Categories: Development

[haiku-development] Re: _kern_reserve_address_range

Fri, 2016-04-01 00:45
OnTue,Mar15,2016at05:59:42PM-0400,TimKellywrote: Theoddthingisthatthiserroroccursabouteveryotherbuildofan executable.Icanrunthecompileimmediatelyagainanditwillsucceed. Runitagainanditfails,withanoutofmemoryissue.Runitagainandit succeeds,etc. WearerunningwithASLR(addressspacelayoutrandomization)enabledby default,whichmeanstheavailablememoryaddresseschangeateachrun. ...
Categories: Development

[haiku-development] replacing runtime_loader

Fri, 2016-04-01 00:45
HiEveryone, Again, thank you to everyone that has contributed information and knowledge in response to my questions. I would like to replace the stock runtime_loader with one that returns the specific error code from _kernel_reserve_address_range. However, this does not appear to be trivial. I've compiled runtime_loader with the change I would like, but I am not able to discern how to replace the system one with the compiled one. I'd prefer to avoid building an entire image and reinstalling. Is there some documentation on replacing individual system files (not add-ons)? ...
Categories: Development

[haiku-development] Re: _kern_reserve_address_range

Thu, 2016-03-31 16:45
In tracking this further, I find that _kern_reserve_address_range can return one of the following values: directly B_BAD_ADDRESS B_BAD_VALUE fromvm_reserve_address_range: B_BAD_VALUE ...
Categories: Development

[haiku-development] Re: _kern_reserve_address_range

Thu, 2016-03-31 16:45
Adrien,Urias,Jerome,andRene,thankyoufortheinformation. HereiswhereIampuzzled: strace-t(stuff) [119613] _kern_reserve_address_range([0x200000], 0x1, 0x67000) = 0x80000005 Invalid Argument (3 us) [119613] _kern_debug_output(runtime_loader: ...
Categories: Development

[haiku-development] Re: _kern_reserve_address_range

Thu, 2016-03-31 16:45
OnTue,Mar15,2016at2:12PM,TimKellygtkelly@xxxxxxxxxxxxxxxxxwrote: somewhere? Sincethe_kern_*isgeneratedatbuildtime,Iguessthecodeyou're lookingforisthis: -Urias ...
Categories: Development

[haiku-development] Re: _kern_reserve_address_range

Thu, 2016-03-31 16:45
OnTue,Mar15,2016at5:12PM,TimKellygtkelly@xxxxxxxxxxxxxxxxxwrote: _user_reserve_address_rangeand_kernel_reserve_address_rangeeffectively thesame? Andregardlessofifitisprivateornot,isn'tthecodeavailable somewhere? _kernistheuserspacestubthatmapstothecorrespondingsystemcall _user_reserve_address_range,whichiswhatJeromepointedto.The ...
Categories: Development

[haiku-development] Re: _kern_reserve_address_range

Thu, 2016-03-31 16:45
OnTue,Mar15,2016at05:12:11PM-0400,TimKellywrote: _kernel_reserve_address_range,whichisstubbedoutin src/system/libroot/stubbed/libroot_stubs.c._kernel_reserve_address_range getscalledbyruntime_loader,whichisofinteresttome. Retractthatstatement-tryingtogetclarification.Are _user_reserve_address_rangeand_kernel_reserve_address_rangeeffectively thesame? ...
Categories: Development

[haiku-development] Re: _kern_reserve_address_range

Thu, 2016-03-31 16:45
TimKellywrote: That's actually _user_reserve_address_range. I'm looking for _kernel_reserve_address_range, which is stubbed out in src/system/libroot/stubbed/libroot_stubs.c. _kernel_reserve_address_range gets called by runtime_loader, which is of interest to me. Retract that statement - trying to get clarification. Are ...
Categories: Development

[haiku-development] Re: _kern_reserve_address_range

Tue, 2016-03-29 22:45
Trying to find out why the call to _kernel_reserve_address_range is failing in runtime_loader with executables compiled by pcc. In conjunction with Jerome's answer, are _user_reserve_address_range and _kernel_address_range going to be basically about the same thing? ReneGollentwrote: ...
Categories: Development

[haiku-development] Re: _kern_reserve_address_range

Tue, 2016-03-29 20:45
That's actually _user_reserve_address_range. I'm looking for _kernel_reserve_address_range, which is stubbed out in src/system/libroot/stubbed/libroot_stubs.c. _kernel_reserve_address_range gets called by runtime_loader, which is of interest to me. JérômeDuvalwrote: Am15.03.20169:25nachm.schriebTimKellygtkelly@xxxxxxxxxxxxxxxxx: ...
Categories: Development