- l4-hackers - example.com Site

crt0.S and C++
by Alexey Mandrookin 15 Oct '02

15 Oct '02

Hi, Could anybody help me with crt0.S? I have project, which tightly related to L4. Project written in C++ and use abstract classes. The crt0.S was taken from FIASCO L4 examples and works well with C code. Unfortunately, it does not work with C++ code. Calling constructors for global objects does not work. Program crashed when abstract classes or virtual functions are used. It seems, virtual tables is not initialized. Guess a __static_initialization_and_destruction_0 function was not called. Does anybody have a C++ compatible crt0.S? Alexei

2 1

Native OS
by Alexey Mandrookin 09 Oct '02

09 Oct '02

Hi, does anybody interested in native OS implementation on top of a L4 microkernel? Perhaps I can donate a console driver within October. Alexei

5 7

Re: Native OS
by Frederic BOYER 09 Oct '02

09 Oct '02

I am interessed by a such project. In the past, I planed to write an Unix like on top of L4 in Ada. The goal was to port the GNAT Ada Run-Time on top of L4 and then write all the servers needed to make a good O.S. My new job take me a LOT of time, but I plan to realise this a futur day. Regards, Frédéric. "Alexey Mandrookin" <alman(a)mgateinc.com>@os.inf.tu-dresden.de on 09/10/2002 04:38:26 Sent by: l4-hackers-admin(a)os.inf.tu-dresden.de To: <l4-hackers(a)os.inf.tu-dresden.de> cc: (bcc: Frederic BOYER/PARIS/CAI-Cheuvreux) Subject: Native OS Hi, does anybody interested in native OS implementation on top of a L4 microkernel? Perhaps I can donate a console driver within October. Alexei _______________________________________________ l4-hackers mailing list l4-hackers(a)os.inf.tu-dresden.de http://os.inf.tu-dresden.de/mailman/listinfo/l4-hackers "This message, including any attachments may contain confidential and privileged material; it is intended only for the person to whom it is addressed. Its contents do not constitute a commitment by Crédit Agricole Indosuez Cheuvreux except where provided for in a written agreement. Crédit Agricole Indosuez Cheuvreux assumes no liability or responsibility for the consequences arising out of a delay and/or loss in transit of this message, or for corruption or other error(s) arising in its transmission and for any misuse or fraudulent use which may be made thereof. If you are not the intended recipient, please contact us and abstain from any disclosure, use or dissemination. To the extent that this message contains research information and/or recommendations, these are provided on the same basis as Crédit Agricole Indosuez Cheuvreux's published research and the recipient must have regard to all disclosures and disclaimers contained therein."

1 0

Pistachio
by Thomas Quintana 07 Oct '02

07 Oct '02

Hi, I was wondering what the time frame for the Pistachio micro-kernel looks like cuz i'm really interested in working on a multiserver project (ex. sawmill) under the v. 4 spec. the performance gains should be outstanding and good enough to compete against most monolithic kernels (not that the hazelnut kernel doesn't give them a run for their money). If any help is needed in the development pls post cvs and a todo list. thx, Thomas --------------------------------- Do you Yahoo!? Faith Hill - Exclusive Performances, Videos, & more faith.yahoo.com

1 0

Purpose of ping-pong in l4linux
by Jacob Gorm Hansen 02 Oct '02

02 Oct '02

hi, what is the purpose of the ping-pong task in l4linux? I can see that it handles vmalloc requests for the kernel, but why does it need its own task for this? thanks, Jacob

3 7

Tuning L4Linux on Fiasco
by Jacob Gorm Hansen 01 Oct '02

01 Oct '02

hi, the original l4linux papers mention performance about 4% lower than native linux. When running l4linux on top of Fiasco, performance seems to be more like 50% of what we get with native 2.2. For example, encoding a .wav-file to ogg-vorbis takes 2 minutes native, and almost 4 minutes with l4linux. LMBench also gives way-different figures. Is there anything we can do to tune Fiasco for better performance? Are other people experiencing the same speed differences as we are? Best regards, Jacob

2 3

Bugreport
by Alexey Mandrookin 24 Sep '02

24 Sep '02

Hi! I found strange behaviour of l4_i386_ipc_wait(). In case of interrupt IPC, it return wrong value of the thread id of IPC source. *** Version $gcc -v Reading specs from /usr/lib/gcc-lib/i686-pc-linux-gnu/2.95.3/specs gcc version 2.95.3 20010315 (release) $as -v GNU assembler version 2.11 (i686-pc-linux-gnu) using BFD version 2.11 $ld -v GNU ld version 2.11 (with BFD 2.11) *** UUencoded source code begin 644 test.tar.gz M'XL(`-=5CCT``^U:67/;1A+V*Z;*_V%B20XIDR!(4Y:+BE/10<M:4\>*<C;> MV(L"@:&("`<#@#KBTO[V_7IP\U!<J;6=![1*)&:FIZ>[IZ>/`<T@TM3AHR\* M6EO37G2[CS1-:V]O:<5O@N?MYYU'VO96]WFWLZUUM@A_^_F+1US[LFS%,`LC M(^#\D2-NC8?Q1!!^#8:^+JQ98FQ[@NOZ[G#8/]X;O-=UMF9[IC.S!/_!Z;9, MWYW:C@C4R8],88H:B=N(L?[)Q?G[FJZ;0:1;,]>]T_5ZVDD:C=)6TE`<83@* M&N958T.$4Z:8AN,HP_?'>Z<#_63WN%]S#=L#8B`*Y).^UB;?!SJG_TO''QD. M-WTOC(*9&?E!J'*^V9)+*+*-%4:WK-U37/_:46K4JJ//N&6*'S@*/27-W_Y0 M.N.$E\VXR[`L1UGOQC24W]RITAZQ3B_FBR<`AHYG3F3O^7[$)\*P1,";/!2" MN]0]0K<ZB=E:L\?0,13<'[R&;E,*JN'8EQ[G7;8FG%`0Q>]2I`:/)D;$76%X MW%#]6<2-,!3N"+M`%.=)=$#"L^QQIC9WI,<\U7-<Q_<N\:W=MO=V#_8TK<-+ M@.5=X](V2_2S.71BZ:@NS!D[QF78X[NG[R[TMX-W!X?]902:11+-`@\+F$5[ M*(A!2\6/.K8G6+9&:6;)+&FRXQO6RJFZL(S(4!0E180V'T#VK!AU%(8/8I8X MBH\!31->%-Q)>\5A`@V&KW!JF$)I:YTNDT>DQQ@P_S41V'_O+C7Y:R.P#1@! M]X2P;*QA%$\!MT-N"=,Q`F%QV^/[SYXU&-_D=L1O`CL2(=")LQL_H/'(AY$) M_KTJ3\SWL%TSLGV/9MQ,;'-"Y`P^]8$)JTNPC>DT\*?@(A*EI0/8*'D1H-%1 M4IF43TUH\GB-QA/C2>,G$+@<V5'X-S]*K4WEK+,[.#H\J77JL32D/*799E(8 MN4.DK*'MVM"Y<]=8JL]8C8X=1MP?I^H,:2)T98EY#89\C$:RW]A![H]^`Z%P M0:/6$HW^/=1C)>KYUI&M@L^!8^,*"8`CON0:?Y+_=31M.\W_D/A1_M?N:L^K M_.]KP!IG:WS]R.KQU!(:U[RM=CAVI=/2VJWG;=[N]MHO>EO;?.)/W%DTX?W; M*5]G:YB93I)N"QITIWAV$;/T:!+`ER,W,/&-<$&I6[OSDB?]B&L6+?T9X`AX M);BH,#0N$7;N3$=@;<P-37@M<KJ#;B$K'-N7L\"07I(X8_98_,YKZY\&7?W@ MZ/R^03G1H(M'_HJK:LOI,O**A<X,E<6^+DV&UVL2H]XBH6FA$(a)^.;RY=G]MC M"L'B%B$@!#->RLW1R?!B=S`H\*0G77K.1P$IX88U4WY2-E2Y-'*2-4ZJ19`) M0YO2A+/SHY]W+_I$;DB\"N_:#GP7"0C;?SW8/1SR9Z^4YBG;/SM+F\C5C@_8 M6G-L3H1YU?3$#6/#\WT=*/R5HB"+B%1S.HT[A]1ER@*2L<'1GFPW'=?D3<</ MK^Q(#V?3J1]$:#NNB\\K$7CX"MS+@+&+W?/#_@7F$%7PMPS`-4*:@40`5G4% M*XJ3%`9KZO%X'GWVUFNG>_\8HE+`)AW4>?.$-R^H4$&NV='BM-/GZS_Q]?_( M;20E@V':36J#=3DUT7:=,)LA;[I\>VLKW_>1[;78:G-0P\B"+[D4V&<JH>!8 MP"6*D,)Q8&O),OIQ?SC</>S7LQZY<-.2=(N6D*Q;0BN27#6!)>;@^M8,QF") M*0Q(>*8MPMR&UFL'_3,(_ZW=4`7?"-(3_277>#C^M[O;VE82_[<Z+]IQ_.]J M5?S_&H`B1F;Y1V?[TI\B@EE"96T47$Y7MY^_?*';4U._,>RH5N<H-1!87#B3 M"(6(L%0Y]P+A3]8UP6P:24H(?$X8A\5NXJ=L2X]X7.[,`L%O;.01:!H>8OJU MX<Q0^M0T5>/((]J::L0W%V_\&W$M@@8(!N`M7\3T72J`)97("*^X;2%B:LA; M:E?B;N0;<8T;8]<9ZT">_<*-4;',BDLL#@ZYAQ*4"EAA,;;98H6+,,S$@NKD MQ\>E3LOV%_L<>R0['Y<OTL*[D/Z)?DB7:0N#4/1"/X7*%@C2MSII%4?'SNRV MY1KF!'5C:[K`B!R^??FB-3)"H5^[\7!RW8=*CP+SJ.8UKNL*6B-]*I]S%-M+ M,.J*[<GANA0I&3X[/;_0W^X=Z`>[%[L/9&_:[0MMR2Q$JXMWPX=FM?-91^?_ MU&DF<AR%`CKQ87L.#5W[V#:9=06NGNUXC;K![J?'3,E$S<@T9%90WWG,[HE2 MOL[;_ON]T]WS`WWOW>O7_7-]>/3O?I&K6ON''U[65^,?[P[?EO"7$6RVI1[E M]8EBT6V,'D&NDU-]_\WNR6%?45XIM:0?"4%[A\6X9,\\7Q+"P.(Q&(a)_.O!`5 MNH`N)G!G)"DQHZ?H=#ANM<Y.D12V8/]M+",(:9TN5F*F8X0A#XP;G<X1_\2F MLY%CFSVFI'VQ;G>80D04?"+/2\]<<M8SE(PKX"I(C`@Q'633P+Y&90#2U`,5 MX&#X-SDM._@]Q86/CFR3)WA%C+D5RXC36:2/9N/QW"BI"*/$SY)18E510KK9 M(VY+#JRAH.F&EY8_%7!GF_FLD>]3"FZ'DJ0(=,_W=.%.HSNXS4_(Q.#U/%XC MM#H'7=WT9UZTP^\+>BB1(H^KPTL1$SNLQ`66D8*7629A;61XM[#K8C^).==_ M[3OX=J375G)FLO&2,24(E@A__9BCQ$Q2Q-"G=&5D[;![[&G<G9I*KU>4`E8V M-E!J[3#)5XZ4<4Y6^GAA.!.`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`4,T(X:`=))$CMD73H0E]$X2X4 MCM1AID1*'W5I&7`-Q1Q"!O(QKTF\[V"!L<FU-F7V'LZF(KBV0SBIF/7D'DO> MWF$SZ&+.$Y?8BVM!^IS"M0GY.D!!5/.B<>W)1MC;L'H\UU&/GQ\?GJ.$\&Q* M#(+?^<:L)Y6,%&7C]H/WI`%%*(JNOSX:].D5A:X/CD[D4Y'SAIQ#_&-K*"*J MSD2%&^2ORDG2,TJ2TO&)?9F<,_E2F=[$13XWH@AI-#W1A$12DF*)"0;"%)"V M)K=NT?J>TG8_O6EG8Q='Q_W3=Q<UK4%_;?JL`R,SOK+ZDTGG_61:O!NI+HM* M-`WO>QS1U+F`=Q*1]X,`F[5&JM2ZOTAEQFKBL9X@='CGF9/`]^P_Y`L[>L%& M!,C)D\RN4&U+=1(MD*[XZK-(1OS91U";/X&Y%CAQ1L%$D2](P8\-?XO4DBY6 M_2F].Y,G!S4?#B`Y0^D+F4(<4PRM[>R0KM:@3/F6FBG_C\U;X%/)=BO>&26. M]<6-(8M.UH$G09YB81?DU'M62$G1(U_',=9J\=0'Y"8,"YWK+YKN,MEDB?XT MF0">,\_P=>1*3/0\QK"(<_6#]R$:^K/`%'SC5MVXY:CE0[H8CX]Y*AS,C:RO MV$[,K]B5S-6A'[EJNN*'Z`TTTQM`;3#Z7WKTP<^E8'SCE^(ZB1(;966G@4!- M'.\<\3,C,-RV),WE<R=^SN\BXD5X[.<;^0`Q38XZR:2D3>+XX9,:(QPN.''S M2KX(MBFD7<[L<,)'(KH1PN-TX20O4\C7EJY7U)1&`&DB?F/<T5MF^IV.$8AT MCZSL7D1>DFAJ)YGTSL-YB68(H,*YHU<3EB]">?-!B2>0I/M&%G*C\5=90)B/ MD7$FJDA$OKSJJ?,L;2`W432V>R[?1*>6EL?\>.'"4@\125]&XVDA8UE=EE)T M+A2"Z0LI^MG%KWE)_)$2$!K>E"-3JBJ>KL!%D?Z13F46\^4O3U*+6(C\)(=[ MIT\#09JBVRUJH30(=N82@'+\3S,E:*A4%20)B]RGSRP_'J@^2M8@S_>J,D3& MDV)^03JOK4@H_BP=H+ES6<"R!"`)^8@323!#3I,&["F=3D$_KTCB6.P@\VQ( M*6J=Q\'VY.?=P1$2A8-DF+9A<:BPA;JX11BY#*4Y9I<VI<>GY=U]FM+%8V(9 MY>\X"Y`:P=*NV$E;Y4!<S&J2_"5YMUK(6!;YC$G)G"H/Q4MRUSF,Q.QE[VJ! M:'"U3/)'95>41@90UI(UR6\%,\^CN#X2L$;YLR*XPAGUK,C`DD"7B/50%%\9 M[MBRNZ;$3[,D]KW-[Y'MR#8<I$N6/`_S94'QDFNN$LRNN^@,%`YQ[BJ*/F+) MP2][CD*E@*(%[AE3DZIQ58*V6DUIO%J9$OQYCO&7=)^Y`HIM:;!2>1)GD20D M438I!6(V"WJO%?5=E\J[SZ]C%+K6H(V@7XX6KH'3:KNT67`C22DJ;770+;YA M,")S`E\@M;#9HAL!=6E0J>\H"EWHIBM8/J=KY[BTI\)=32U!7E1(#YTZ0S-. MT"=RY)Y^>.>(&F;!;[9E7VJ,P\F,`O6-5[1`+;W!_M8OD2JHH((**JB@@@HJ JJ*"""BJHH((**JB@@@HJJ*"""BJHH((**JB@@@HJ^&;P/[=^;\8`4``` ` end Best regards Alexei

3 3

Portable servers?
by Alexey Mandrookin 19 Sep '02

19 Sep '02

Hi! How to write portable servers for L4 micro-kernel family? I found what the l4_threadid_t type is different in various L4 implementation. Does any wrapper on top of syscalls exist? ==== Fiasco microkernel: ==== typedef struct { unsigned version_low:10; unsigned lthread:7; unsigned task:11; unsigned version_high:4; unsigned site:17; unsigned chief:11; unsigned nest:4; } l4_threadid_struct_t; typedef union { l4_low_high_t lh; l4_threadid_struct_t id; } l4_threadid_t; typedef l4_threadid_t l4_taskid_t; === L4-KA microkernel: ==== #define L4_X0_VERSION_BITS 10 #define L4_X0_THREADID_BITS 6 #define L4_X0_TASKID_BITS 8 typedef union { struct { unsigned version : L4_X0_VERSION_BITS; unsigned thread : (L4_X0_THREADID_BITS + L4_X0_TASKID_BITS); unsigned chief : L4_X0_TASKID_BITS; } id; struct { unsigned version : L4_X0_VERSION_BITS; unsigned thread : L4_X0_THREADID_BITS; unsigned task : L4_X0_TASKID_BITS; unsigned chief : L4_X0_TASKID_BITS; } x0id; dword_t raw; } l4_threadid_t; ========================== Second question is how to distinguish between IRQ messages and ordinary messages. That code makes me confused: irq_th.lh.low = COM_IRQ + 1; irq_th.lh.high = 0; as I understand, it correspond to irq_th.id.version_low = COM_IRQ + 1; irq_th.id.lthread = 0; irq_th.id.task=0; irq_th.id.version_high=0; irq_th.id.site=0; irq_th.id.chief=0; irq_th.id.nest=0; Is lthread = 0 and task = 0 means interrupts? Third question is how to handle memory mapped IO? I found phystokv() function. It is just a simple macro, which placed deep inside of OSKIT Is that right to use that function for MMIO? What will happen if several servers requests same memory range simultaneously? Best regards Alexei

2 1

Clans&Chiefs question
by Jacob Gorm Hansen 16 Sep '02

16 Sep '02

hi, I have the following setup: - rmgr, fiasco, and sigma - a network server 'net' which exports an abstracted ethernet-device, built with oskit10. - a custom pager, 'pager' which is able to spawn new l4linux servers (currently piggybacked on to it). All started via entries in menu.lst, using Grub The spawned l4linuxes would like to talk to 'net' so that they may boot from NFS. However, net and pager are not in the same clan, so that is not allowed (obviously they need to talk to pager as well to get paged). Both net and pager need to talk to rmgr. Is there a way, other than merging net & pager into one task (which works fine, but is not really what I want), in which this can be made to work in fiasco? Thanks, Jacob

2 1

One more problem
by Alexey Mandrookin 13 Sep '02

13 Sep '02

Hi everybody! New problem appeared. It seems to be a make looking for subdirectories and did not found theirs. $gcc -v gcc version 2.95.3 20010315 (release) $make hello make[3]: Entering directory `/usr/src/Fiasco/l4/pkg/l4util/examples/watchdog' set -e ; for d in ; do \ make -C OBJ-$d all ; done /bin/sh: -c: line 1: syntax error near unexpected token `;' /bin/sh: -c: line 1: `set -e ; for d in ; do make -C OBJ-$d all ; done' make[3]: *** [all] Error 2 make[3]: Leaving directory `/usr/src/Fiasco/l4/pkg/l4util/examples/watchdog' make[2]: *** [all] Error 2 make[2]: Leaving directory `/usr/src/Fiasco/l4/pkg/l4util/examples' make[1]: *** [all] Error 2 make[1]: Leaving directory `/usr/src/Fiasco/l4/pkg/l4util' make: *** [hello] Error 2 Best regards Alexei

2 2