Re: Base registers
On 6/1/12 10:25 AM, "Scott Ford" wrote: > LRR12,R15 program addressability > LAR12,2048(R11) > LAR12,2048(R12) > USING LOGEVX01,R11,R12 set base register another possibility, lets you keep the base regs in "order" R11,R12: LRR11,R15 program addressability LAR12,2048(R11) LAR12,2048(R12) USING LOGEVX01,R11,R12 set base register
Re: Base registers
Scott, your reply-to on your list posts is set to your email address instead of the list. As I said privately (not on purpose), change the LR R12,R15 to LR R11,R12 or change your other instructions to load R11 and use R12,R11. Chuck Arney Arney Computer Systems -Original Message- From: IBM Mainframe Assembler List [mailto:ASSEMBLER-LIST@LISTSERV.UGA.EDU] On Behalf Of Scott Ford Sent: Friday, June 01, 2012 9:26 AM To: ASSEMBLER-LIST@LISTSERV.UGA.EDU Subject: Base registers Guys/Gals: I am the process of modifying our IRREVX01 and have a fundamental question...Currently when the exit is called the prolog looks like this: LOGEVX01 CSECT , LOGEVX01 AMODE 31 LOGEVX01 RMODE ANY YREGS SAVE (14,12) LRR12,R15 program addressability USING LOGEVX01,R12set base register LRR10,R1 save parameter address USING EVXPL,R10 base parameter map L R3,EVXFLAGS exit flag address TM0(R3),EVXPOST post-exit ? BNO GOBACK if not, get out Which works fine . I have added new code and exceeded the initial register cover range of 4k ..I tried to add an addition base register like this: LOGEVX01 CSECT , LOGEVX01 AMODE 31 LOGEVX01 RMODE ANY YREGS SAVE (14,12) LRR12,R15 program addressability LAR12,2048(R11) LAR12,2048(R12) USING LOGEVX01,R11,R12 set base register LRR10,R1 save parameter address USING EVXPL,R10 base parameter map L R3,EVXFLAGS exit flag address TM0(R3),EVXPOST post-exit ? BNO GOBACK if not, get out Assembled and liked the exit, T PROG=xx , where xx is the parmlib member to delete the exit Reloaded the exit via T PROG=yy to activate the exit . Ran a test through the exit ( a RACF command ) and a S0C1-1 Without the addition of the new lines for the second base register. The exit works fine.. Either I blew it on the second base or IRREVX01 has a size restriction ... Can someone point me the right way ? Thanks in advance.. Regards Scott J Ford Software Engineer http://www.identityforge.com
Re: Base registers
Why not just change the code to baseless? I have done that with various things I touched over the last decade to good effect. Now that I am back in a group that does assembler, I look forward to doing more. Start with subroutines using LOCTR * at top of code after CSECT BASED LOCTR , . . . JAS R14,SUB1 . . BASELESS LOCTR , * begin subroutine SUB1 DS 0H . . . BR R14 * end of subroutine BASED LOCTR , . . . IBM Mainframe Assembler List wrote on 06/01/2012 10:44:15 AM: > From: Chuck Arney > > Scott, your reply-to on your list posts is set to your email address instead > of the list. > > As I said privately (not on purpose), change the LR R12,R15 to LR R11,R12 or > change your other instructions to load R11 and use R12,R11. > > Chuck Arney > Arney Computer Systems > > -Original Message- > From: IBM Mainframe Assembler List [ mailto:ASSEMBLER-LIST@LISTSERV.UGA.EDU] > > Guys/Gals: > > I am the process of modifying our IRREVX01 and have a fundamental > question...Currently when the exit is called the prolog looks like this: > > LOGEVX01 CSECT , > LOGEVX01 AMODE 31 > LOGEVX01 RMODE ANY > YREGS > SAVE (14,12) > LRR12,R15 program addressability > USING LOGEVX01,R12set base register > LRR10,R1 save parameter address > USING EVXPL,R10 base parameter map > L R3,EVXFLAGS exit flag address > TM0(R3),EVXPOST post-exit ? > BNO GOBACK if not, get out > > Which works fine . I have added new code and exceeded the initial register > cover range of 4k ..I tried to add an addition base register like this: > > LOGEVX01 CSECT , > LOGEVX01 AMODE 31 > LOGEVX01 RMODE ANY > YREGS > SAVE (14,12) > LRR12,R15 program addressability > LAR12,2048(R11) > LAR12,2048(R12) > USING LOGEVX01,R11,R12 set base register > LRR10,R1 save parameter address > USING EVXPL,R10 base parameter map > L R3,EVXFLAGS exit flag address > TM0(R3),EVXPOST post-exit ? > BNO GOBACK if not, get out > > Assembled and liked the exit, T PROG=xx , where xx is the parmlib member to > delete the exit Reloaded the exit via T PROG=yy to activate the exit . > > Ran a test through the exit ( a RACF command ) and a S0C1-1 Without the > addition of the new lines for the second base register. The exit works > fine.. > > Either I blew it on the second base or IRREVX01 has a size restriction ... > > Can someone point me the right way ? > > Thanks in advance.. > > Regards > Scott J Ford - The information contained in this communication (including any attachments hereto) is confidential and is intended solely for the personal and confidential use of the individual or entity to whom it is addressed. If the reader of this message is not the intended recipient or an agent responsible for delivering it to the intended recipient, you are hereby notified that you have received this communication in error and that any review, dissemination, copying, or unauthorized use of this information, or the taking of any action in reliance on the contents of this information is strictly prohibited. If you have received this communication in error, please notify us immediately by e-mail, and delete the original message. Thank you
Re: Base registers
Best approach is to use IEABRC and use baseless coding. Move all your data to the end and use LARL to set a base register. In your case you are telling the assembler that R11 is a base but you never set it. On Fri, 1 Jun 2012 07:25:45 -0700 Scott Ford wrote: :>Guys/Gals: :> :>I am the process of modifying our IRREVX01 and have a fundamental question...Currently when the exit is called the prolog looks like this: :> :>LOGEVX01 CSECT , :>LOGEVX01 AMODE 31 :>LOGEVX01 RMODE ANY :> YREGS :> SAVE (14,12) :> LRR12,R15 program addressability :> USING LOGEVX01,R12set base register :> LRR10,R1 save parameter address :> USING EVXPL,R10 base parameter map :> L R3,EVXFLAGS exit flag address :> TM0(R3),EVXPOST post-exit ? :> BNO GOBACK if not, get out :> :>Which works fine . I have added new code and exceeded the initial register cover range of 4k ..I tried to add an addition base :>register like this: :> :>LOGEVX01 CSECT , :>LOGEVX01 AMODE 31 :>LOGEVX01 RMODE ANY :> YREGS :> SAVE (14,12) :> LRR12,R15 program addressability :> LAR12,2048(R11) :> LAR12,2048(R12) :> USING LOGEVX01,R11,R12 set base register :> LRR10,R1 save parameter address :> USING EVXPL,R10 base parameter map :> L R3,EVXFLAGS exit flag address :> TM0(R3),EVXPOST post-exit ? :> BNO GOBACK if not, get out :> :>Assembled and liked the exit, T PROG=xx , where xx is the parmlib member to delete the exit :>Reloaded the exit via T PROG=yy to activate the exit . :> :>Ran a test through the exit ( a RACF command ) and a S0C1-1 :>Without the addition of the new lines for the second base register. The exit works fine.. :> :>Either I blew it on the second base or IRREVX01 has a size restriction ... :> :>Can someone point me the right way ? :> :> :>Thanks in advance.. :> :>Regards :>Scott J Ford :>Software Engineer :>http://www.identityforge.com -- Binyamin Dissen http://www.dissensoftware.com Director, Dissen Software, Bar & Grill - Israel Should you use the mailblocks package and expect a response from me, you should preauthorize the dissensoftware.com domain. I very rarely bother responding to challenge/response systems, especially those from irresponsible companies.
Re: Base registers
Thanks guys, I am looking at Ed's Share presentation. ty a bunch.wow I missed a lot. Like Kirk my assembler brain and eyes are coming back. We work in Java,Cobol,C and Assembler...So i have go back and forth ... Thanks again Scott J Ford Software Engineer http://www.identityforge.com From: Binyamin Dissen To: ASSEMBLER-LIST@LISTSERV.UGA.EDU Sent: Friday, June 1, 2012 11:11 AM Subject: Re: Base registers Best approach is to use IEABRC and use baseless coding. Move all your data to the end and use LARL to set a base register. In your case you are telling the assembler that R11 is a base but you never set it. On Fri, 1 Jun 2012 07:25:45 -0700 Scott Ford wrote: :>Guys/Gals: :> :>I am the process of modifying our IRREVX01 and have a fundamental question...Currently when the exit is called the prolog looks like this: :> :>LOGEVX01 CSECT , :>LOGEVX01 AMODE 31 :>LOGEVX01 RMODE ANY :> YREGS :> SAVE (14,12) :> LR R12,R15 program addressability :> USING LOGEVX01,R12 set base register :> LR R10,R1 save parameter address :> USING EVXPL,R10 base parameter map :> L R3,EVXFLAGS exit flag address :> TM 0(R3),EVXPOST post-exit ? :> BNO GOBACK if not, get out :> :>Which works fine . I have added new code and exceeded the initial register cover range of 4k ..I tried to add an addition base :>register like this: :> :>LOGEVX01 CSECT , :>LOGEVX01 AMODE 31 :>LOGEVX01 RMODE ANY :> YREGS :> SAVE (14,12) :> LR R12,R15 program addressability :> LA R12,2048(R11) :> LA R12,2048(R12) :> USING LOGEVX01,R11,R12 set base register :> LR R10,R1 save parameter address :> USING EVXPL,R10 base parameter map :> L R3,EVXFLAGS exit flag address :> TM 0(R3),EVXPOST post-exit ? :> BNO GOBACK if not, get out :> :>Assembled and liked the exit, T PROG=xx , where xx is the parmlib member to delete the exit :>Reloaded the exit via T PROG=yy to activate the exit . :> :>Ran a test through the exit ( a RACF command ) and a S0C1-1 :>Without the addition of the new lines for the second base register. The exit works fine.. :> :>Either I blew it on the second base or IRREVX01 has a size restriction ... :> :>Can someone point me the right way ? :> :> :>Thanks in advance.. :> :>Regards :>Scott J Ford :>Software Engineer :>http://www.identityforge.com -- Binyamin Dissen http://www.dissensoftware.com Director, Dissen Software, Bar & Grill - Israel Should you use the mailblocks package and expect a response from me, you should preauthorize the dissensoftware.com domain. I very rarely bother responding to challenge/response systems, especially those from irresponsible companies.
Re: Base registers
At 07:25 -0700 on 06/01/2012, Scott Ford wrote about Base registers: Guys/Gals: I am the process of modifying our IRREVX01 and have a fundamental question...Currently when the exit is called the prolog looks like this: LOGEVX01 CSECT , LOGEVX01 AMODE 31 LOGEVX01 RMODE ANY YREGS SAVE (14,12) LRR12,R15 program addressability USING LOGEVX01,R12set base register LRR10,R1 save parameter address USING EVXPL,R10 base parameter map L R3,EVXFLAGS exit flag address TM0(R3),EVXPOST post-exit ? BNO GOBACK if not, get out Which works fine . I have added new code and exceeded the initial register cover range of 4k ..I tried to add an addition base register like this: LOGEVX01 CSECT , LOGEVX01 AMODE 31 LOGEVX01 RMODE ANY YREGS SAVE (14,12) LR R12,R15 program addressability LA R12,2048(R11) LAR12,2048(R12) This should be: LAR11,2048(R12) LAR11,2048(R11) You were destroying R12 by loading it with whatever was in R11+4096 USING LOGEVX01,R11,R12 set base register LR R10,R1 save parameter address USING EVXPL,R10 base parameter map L R3,EVXFLAGS exit flag address TM0(R3),EVXPOST post-exit ? BNO GOBACK if not, get out Assembled and liked the exit, T PROG=xx , where xx is the parmlib member to delete the exit Reloaded the exit via T PROG=yy to activate the exit . Ran a test through the exit ( a RACF command ) and a S0C1-1 Without the addition of the new lines for the second base register. The exit works fine.. Either I blew it on the second base or IRREVX01 has a size restriction ... Can someone point me the right way ? Thanks in advance.. Regards Scott J Ford Software Engineer http://www.identityforge.com
Re: Base registers
Robert A. Rosenberg wrote: At 07:25 -0700 on 06/01/2012, Scott Ford wrote about Base registers: Guys/Gals: I am the process of modifying our IRREVX01 and have a fundamental question...Currently when the exit is called the prolog looks like this: LOGEVX01 CSECT , LOGEVX01 AMODE 31 LOGEVX01 RMODE ANY YREGS SAVE (14,12) LRR12,R15 program addressability USING LOGEVX01,R12set base register LRR10,R1 save parameter address USING EVXPL,R10 base parameter map L R3,EVXFLAGS exit flag address TM0(R3),EVXPOST post-exit ? BNO GOBACK if not, get out Which works fine . I have added new code and exceeded the initial register cover range of 4k ..I tried to add an addition base register like this: LOGEVX01 CSECT , LOGEVX01 AMODE 31 LOGEVX01 RMODE ANY YREGS SAVE (14,12) LR R12,R15 program addressability LA R12,2048(R11) LAR12,2048(R12) This should be: LAR11,2048(R12) LAR11,2048(R11) This will work only if you adjust the USING to match it. You were destroying R12 by loading it with whatever was in R11+4096 USING LOGEVX01,R11,R12 set base register LR R10,R1 save parameter address USING EVXPL,R10 base parameter map L R3,EVXFLAGS exit flag address TM0(R3),EVXPOST post-exit ? BNO GOBACK if not, get out Assembled and liked the exit, T PROG=xx , where xx is the parmlib member to delete the exit Reloaded the exit via T PROG=yy to activate the exit . Ran a test through the exit ( a RACF command ) and a S0C1-1 Without the addition of the new lines for the second base register. The exit works fine.. Either I blew it on the second base or IRREVX01 has a size restriction ... Can someone point me the right way ? Thanks in advance.. Regards Scott J Ford Software Engineer http://www.identityforge.com Bob
Re: Base registers
At 18:44 -0400 on 06/01/2012, Bob Rutledge wrote about Re: Base registers: > This should be: LAR11,2048(R12) LAR11,2048(R11) This will work only if you adjust the USING to match it. You were destroying R12 by loading it with whatever was in R11+4096 >> USING LOGEVX01,R11,R12 set base register OOPS you're right. His LAs were correct BUT the LR R12,R15 after the SAVE should have been a LR R11,R15. Otherwise my reply was correct if the USING was updated as you point out.
Re: Base registers
From: Robert A. Rosenberg Sent: Saturday, 2 June 2012 7:04 AM LAR11,2048(R12) LAR11,2048(R11) LA 11,4095(0,12) requires only one instruction (adjust USING accordingly).
Re: Base registers
Why not: ALFI R11,4096 John -Original Message- From: IBM Mainframe Assembler List [mailto:ASSEMBLER-LIST@LISTSERV.UGA.EDU] On Behalf Of robin Sent: Saturday, June 02, 2012 9:12 AM To: ASSEMBLER-LIST@LISTSERV.UGA.EDU Subject: Re: Base registers From: Robert A. Rosenberg Sent: Saturday, 2 June 2012 7:04 AM > LAR11,2048(R12) > LAR11,2048(R11) LA 11,4095(0,12) requires only one instruction (adjust USING accordingly).
Re: Base registers
I don't think you really want your base register pointing to an odd address. You need to add 1 more to make it right and that requires another instruction. Chuck Arney Arney Computer Systems -Original Message- From: IBM Mainframe Assembler List [mailto:ASSEMBLER-LIST@LISTSERV.UGA.EDU] On Behalf Of robin Sent: Saturday, June 02, 2012 8:12 AM To: ASSEMBLER-LIST@LISTSERV.UGA.EDU Subject: Re: Base registers From: Robert A. Rosenberg Sent: Saturday, 2 June 2012 7:04 AM > LAR11,2048(R12) > LAR11,2048(R11) LA 11,4095(0,12) requires only one instruction (adjust USING accordingly).
Re: Base registers
Only one instruction is required: LAY R11,4096(,R12) But really, for over 15 years base registers are only required for data areas. Just use LA, LAY or LARL to address the first data item and base the USING from there. Also switch from branches (Bc, BAS, BCT...) to jumps (Jc, JAS, JCT...). On Sat, 2 Jun 2012 15:53:59 -0500, Chuck Arney wrote: >I don't think you really want your base register pointing to an odd address. >You need to add 1 more to make it right and that requires another >instruction. > >Chuck Arney >Arney Computer Systems > >-Original Message- >From: IBM Mainframe Assembler List [mailto:ASSEMBLER-LIST@LISTSERV.UGA.EDU] >On Behalf Of robin >Sent: Saturday, June 02, 2012 8:12 AM > >From: Robert A. Rosenberg >Sent: Saturday, 2 June 2012 7:04 AM > >> LAR11,2048(R12) >> LAR11,2048(R11) > > LA 11,4095(0,12) > >requires only one instruction (adjust USING accordingly).
Re: Base registers
I agree with Chuck, in fact this is on our interview test! anyone who sets up the 2nd -> nn base registers with 4095, has obviously never had to prepare maint. via a zap! So they would never be hired! - Original Message - From: "Chuck Arney" To: Sent: Saturday, June 02, 2012 1:53 PM Subject: Re: [ASSEMBLER-LIST] Base registers I don't think you really want your base register pointing to an odd address. You need to add 1 more to make it right and that requires another instruction. Chuck Arney Arney Computer Systems -Original Message- From: IBM Mainframe Assembler List [mailto:ASSEMBLER-LIST@LISTSERV.UGA.EDU] On Behalf Of robin Sent: Saturday, June 02, 2012 8:12 AM To: ASSEMBLER-LIST@LISTSERV.UGA.EDU Subject: Re: Base registers From: Robert A. Rosenberg Sent: Saturday, 2 June 2012 7:04 AM LAR11,2048(R12) LAR11,2048(R11) LA 11,4095(0,12) requires only one instruction (adjust USING accordingly).
Re: Base registers
David Bond writes: But really, for over 15 years base registers are only required for data areas. Just use LA, LAY or LARL to address the first data item and base the USING from there. Also switch from branches (Bc, BAS, BCT...) to jumps (Jc, JAS, JCT...). Indeed! There is a case--a strong one--for using traditional methods in patching old code; but there is none to be made for doing so in writing even a new single RSECT. The two-base-register case for something just a bit over 4096 bytes in length is a particularly egregious, hidebound use of obsolete technology. Let's stop it. John Gilmore, Ashland, MA 01721 - USA
Re: Base registers
Certainly I'm not suggesting this is the best approach to handle new code updates. I'm only pointing out that the previous advice on how to correct the described problem is incorrect. I'm sure we could all come up with many other ways to load base registers or even eliminate them. Chuck Arney Arney Computer Systems -Original Message- From: IBM Mainframe Assembler List [mailto:ASSEMBLER-LIST@LISTSERV.UGA.EDU] On Behalf Of David Bond Sent: Saturday, June 02, 2012 5:13 PM To: ASSEMBLER-LIST@LISTSERV.UGA.EDU Subject: Re: Base registers Only one instruction is required: LAY R11,4096(,R12) But really, for over 15 years base registers are only required for data areas. Just use LA, LAY or LARL to address the first data item and base the USING from there. Also switch from branches (Bc, BAS, BCT...) to jumps (Jc, JAS, JCT...). On Sat, 2 Jun 2012 15:53:59 -0500, Chuck Arney wrote: >I don't think you really want your base register pointing to an odd address. >You need to add 1 more to make it right and that requires another >instruction. > >Chuck Arney >Arney Computer Systems > >-Original Message- >From: IBM Mainframe Assembler List >[mailto:ASSEMBLER-LIST@LISTSERV.UGA.EDU] >On Behalf Of robin >Sent: Saturday, June 02, 2012 8:12 AM > >From: Robert A. Rosenberg >Sent: Saturday, 2 June 2012 7:04 AM > >> LAR11,2048(R12) >> LAR11,2048(R11) > > LA 11,4095(0,12) > >requires only one instruction (adjust USING accordingly).
Re: Base registers
> but there is none to be made for doing so in > writing even a new single RSECT. How about this reason. We have several customers running our software that are on pre-MP3000 machines that don't even support relative instructions. They still pay us for support and that includes software upgrades. Other vendors may not care about existing customers, but we do. Almost all of these customers also can't upgrade their VSE. They fell into the ESL trap with IBM many years ago and now can't get their budget dollars back to get current because of the IBM monthly charges. As faithful customers for many years, we do not want to kick them while they are down just so we can do 'fancier' code. Tony Thigpen -Original Message - From: John Gilmore Sent: 06/02/2012 06:57 PM David Bond writes: But really, for over 15 years base registers are only required for data areas. Just use LA, LAY or LARL to address the first data item and base the USING from there. Also switch from branches (Bc, BAS, BCT...) to jumps (Jc, JAS, JCT...). Indeed! There is a case--a strong one--for using traditional methods in patching old code; but there is none to be made for doing so in writing even a new single RSECT. The two-base-register case for something just a bit over 4096 bytes in length is a particularly egregious, hidebound use of obsolete technology. Let's stop it. John Gilmore, Ashland, MA 01721 - USA
Re: Base registers
At 15:49 -0700 on 06/02/2012, Trivers Software wrote about Re: Base registers: I agree with Chuck, in fact this is on our interview test! anyone who sets up the 2nd -> nn base registers with 4095, has obviously never had to prepare maint. via a zap! So they would never be hired! Once I have a 3rd base register, I use this sequence (in lieu of the two instruction per register LA RX,2048... version) to load. If my BASE is R12, and I want to add R9-R11 as extra bases: LA R9,4095 Prime R9 with 4096 LA R11,1(R9,R12) R11=R12+4096 LA R10,1(R9,R11) R10=R11+4096 LA R9,1(R9,R10)R9=R10+4096
Re: Base registers
On 6/2/2012 6:57 PM, John Gilmore wrote: The two-base-register case for something just a bit over 4096 bytes in length is a particularly egregious, hidebound use of obsolete technology. Let's stop it. As they say, all generalities are false. Some of us write dual mode code, to be used both on modern systems and on traditional ones. The Hercules groups for DOS, MVT, and MVS have hundreds of users, and legacy systems and software are alive and well. Running old programs as a hobby is considerably less expensive than many others, and may even help delay the onset of Alzheimer's. Gerhard Postpischil Bradford, VT
Re: Base registers
On Sun, 3 Jun 2012 02:52:11 -0400 Gerhard Postpischil wrote: :>On 6/2/2012 6:57 PM, John Gilmore wrote: :>> The two-base-register case for something just a bit over 4096 bytes in :>> length is a particularly egregious, hidebound use of obsolete :>> technology. Let's stop it. :>As they say, all generalities are false. Some of us write dual :>mode code, to be used both on modern systems and on traditional :>ones. The Hercules groups for DOS, MVT, and MVS have hundreds of :>users, and legacy systems and software are alive and well. :>Running old programs as a hobby is considerably less expensive :>than many others, and may even help delay the onset of Alzheimer's. The eternal separation between practical business programming and pie-in-the-sky elitism. -- Binyamin Dissen http://www.dissensoftware.com Director, Dissen Software, Bar & Grill - Israel Should you use the mailblocks package and expect a response from me, you should preauthorize the dissensoftware.com domain. I very rarely bother responding to challenge/response systems, especially those from irresponsible companies.
Re: Base registers
Tony Thigpen wrote: We have several customers running our software that are on pre-MP3000 machines that don't even support relative instructions. They still pay us for support and that includes software upgrades. Other vendors may not care about existing customers, but we do. Gerhard Postpischil wrote: As they say, all generalities are false. Some of us write dual mode code, to be used both on modern systems and on traditional ones. The Hercules groups for DOS, MVT, and MVS have hundreds of users, and legacy systems and software are alive and well. Running old programs as a hobby is considerably less expensive than many others, and may even help delay the onset of Alzheimer's. Binyamin Dissen wrote: The eternal separation between practical business programming and pie-in-the-sky elitism. Let me deal with them in reverse order. Binyamin's argument is ad hominem, name-calling; and that is enough to say about it. For the record, I will, however, note that, while I like 'pie in the sky', resonant as it is of Joe Hill and the IWW, it combines badly with 'élitist'. Gerhard's point is not so easily dismissed; but 1) it is self-defeating for the technology if pushed very far: it freezes software in a posture that is now more than fifteen years behind the state of the art; and 2) the market for even moderately-priced software among MVT-using hobbyists is very small. Still, seriously pursued dual-path software is is certainly viable. Tony's argument surprised me a little in two ways. I was surprised by the apparent size of the market he describes and by his vehemence, both of which may reflect special characteristics of VSE users, about whom I know too little. In his circumstances dual paths would also seem to be appropriate. As a now certified élitist, I am nevertheless unrepentent. Urgings about compatibility requirements and the like usually hide Luddite impulses, reluctance to accommodate the (not very) new, which account for a good many of the troubles that the mainframe community is experiencing. John Gilmore, Ashland, MA 01721 - USA
Re: Base registers
Robert, Thank you I appreciate the help been writing assembler a long time but not consistently every day. I have had to wear many function hats, sysprog vm,VSE,CICS,vtam ,tcpip, network engineer, etc. So I got rusty in assembler, I think everyone on here are very helpful..and it is appreciated. Scott ford www.identityforge.com On Jun 3, 2012, at 12:30 AM, "Robert A. Rosenberg" wrote: > At 15:49 -0700 on 06/02/2012, Trivers Software wrote about Re: Base registers: > >> I agree with Chuck, in fact this is on our interview test! >> anyone who sets up the 2nd -> nn base registers with 4095, has obviously >> never had to prepare maint. via a zap! >> So they would never be hired! > > Once I have a 3rd base register, I use this sequence (in lieu of the > two instruction per register LA RX,2048... version) to load. If my > BASE is R12, and I want to add R9-R11 as extra bases: > > LA R9,4095 Prime R9 with 4096 > LA R11,1(R9,R12) R11=R12+4096 > LA R10,1(R9,R11) R10=R11+4096 > LA R9,1(R9,R10)R9=R10+4096
Re: Base registers
John, In am far from a Luddite , but I think you have to have need, i.e. project or task to want to change code. My need is a rather large exit we use for RACF. I inherited the exit , so I am trying to Incorporate modern methods and functionality in our code. When your making money off your code , like we do as a vendor , you look at things differently and IMHO you have a responsibility to your customers, yes we are very customer sensitive. Scott ford www.identityforge.com On Jun 3, 2012, at 8:52 AM, John Gilmore wrote: > Tony Thigpen wrote: > > > We have several customers running our software that are on pre-MP3000 > machines that don't even support relative instructions. They still pay > us for support and that includes software upgrades. > > Other vendors may not care about existing customers, but we do. > > > Gerhard Postpischil wrote: > > > As they say, all generalities are false. Some of us write dual mode > code, to be used both on modern systems and on traditional ones. The > Hercules groups for DOS, MVT, and MVS have hundreds of users, and > legacy systems and software are alive and well. > Running old programs as a hobby is considerably less expensive than > many others, and may even help delay the onset of Alzheimer's. > > > Binyamin Dissen wrote: > > > The eternal separation between practical business programming and > pie-in-the-sky elitism. > > > Let me deal with them in reverse order. > > Binyamin's argument is ad hominem, name-calling; and that is enough to > say about it. For the record, I will, however, note that, while I > like 'pie in the sky', resonant as it is of Joe Hill and the IWW, it > combines badly with 'élitist'. > > Gerhard's point is not so easily dismissed; but 1) it is > self-defeating for the technology if pushed very far: it freezes > software in a posture that is now more than fifteen years behind the > state of the art; and 2) the market for even moderately-priced > software among MVT-using hobbyists is very small. Still, seriously > pursued dual-path software is is certainly viable. > > Tony's argument surprised me a little in two ways. I was surprised by > the apparent size of the market he describes and by his vehemence, > both of which may reflect special characteristics of VSE users, about > whom I know too little. In his circumstances dual paths would also > seem to be appropriate. > > As a now certified élitist, I am nevertheless unrepentent. Urgings > about compatibility requirements and the like usually hide Luddite > impulses, reluctance to accommodate the (not very) new, which account > for a good many of the troubles that the mainframe community is > experiencing. > > John Gilmore, Ashland, MA 01721 - USA
Re: Base registers
On Sun, 3 Jun 2012 12:38:07 -0400 Scott Ford wrote: :>John, :> :>In am far from a Luddite , but I think you have to have need, i.e. project or task to want to change code. My need is a rather large exit we use for RACF. I inherited the exit , so I am trying to :>Incorporate modern methods and functionality in our code. When your making money off your code , like we do as a vendor , you look at things differently and IMHO you have a responsibility to your customers, yes we are very customer sensitive. Very well put. You cannot force customers to upgrade, and you must enhance if you want the maintenance revenue stream. Academics are free to postulate otherwise. :>On Jun 3, 2012, at 8:52 AM, John Gilmore wrote: :> :>> Tony Thigpen wrote: :>> :>> :>> We have several customers running our software that are on pre-MP3000 :>> machines that don't even support relative instructions. They still pay :>> us for support and that includes software upgrades. :>> :>> Other vendors may not care about existing customers, but we do. :>> :>> :>> Gerhard Postpischil wrote: :>> :>> :>> As they say, all generalities are false. Some of us write dual mode :>> code, to be used both on modern systems and on traditional ones. The :>> Hercules groups for DOS, MVT, and MVS have hundreds of users, and :>> legacy systems and software are alive and well. :>> Running old programs as a hobby is considerably less expensive than :>> many others, and may even help delay the onset of Alzheimer's. :>> :>> :>> Binyamin Dissen wrote: :>> :>> :>> The eternal separation between practical business programming and :>> pie-in-the-sky elitism. :>> :>> :>> Let me deal with them in reverse order. :>> :>> Binyamin's argument is ad hominem, name-calling; and that is enough to :>> say about it. For the record, I will, however, note that, while I :>> like 'pie in the sky', resonant as it is of Joe Hill and the IWW, it :>> combines badly with 'élitist'. :>> :>> Gerhard's point is not so easily dismissed; but 1) it is :>> self-defeating for the technology if pushed very far: it freezes :>> software in a posture that is now more than fifteen years behind the :>> state of the art; and 2) the market for even moderately-priced :>> software among MVT-using hobbyists is very small. Still, seriously :>> pursued dual-path software is is certainly viable. :>> :>> Tony's argument surprised me a little in two ways. I was surprised by :>> the apparent size of the market he describes and by his vehemence, :>> both of which may reflect special characteristics of VSE users, about :>> whom I know too little. In his circumstances dual paths would also :>> seem to be appropriate. :>> :>> As a now certified élitist, I am nevertheless unrepentent. Urgings :>> about compatibility requirements and the like usually hide Luddite :>> impulses, reluctance to accommodate the (not very) new, which account :>> for a good many of the troubles that the mainframe community is :>> experiencing. :>> :>> John Gilmore, Ashland, MA 01721 - USA -- Binyamin Dissen http://www.dissensoftware.com Director, Dissen Software, Bar & Grill - Israel Should you use the mailblocks package and expect a response from me, you should preauthorize the dissensoftware.com domain. I very rarely bother responding to challenge/response systems, especially those from irresponsible companies.
Re: Base registers
In my original post I noted that old code should be maintained using the technology in which it was written, typically using base registers for code addressability. Moreover, it is clear enough that ISVs must be customer-sensitive if they are to survive. These things said, there is still far too much HLASM code being written using methods that have been, at best, obsolescent for many, many years. With this post I have said enough, maybe too much, about this topic. John Gilmore, Ashland, MA 01721 - USA
Re: Base registers
At 12:21 -0400 on 06/03/2012, Scott Ford wrote about Re: Base registers: Robert, Thank you I appreciate the help been writing assembler a long time but not consistently every day. I have had to wear many function hats, sysprog vm,VSE,CICS,vtam ,tcpip, network engineer, etc. So I got rusty in assembler, I think everyone on here are very helpful..and it is appreciated. Scott ford www.identityforge.com Some times you just run into a cute (or non-standard) trick with assembler. My favorite is the use of the TR command to do a patterned move (when the source and target are both under 256 bytes in length). You fill in the target area with offsets into the source and then TR the target using the source as the TR Table. This was a trick to flip the order of the bytes in a field (so it reads right to left) before MVCI was added to the instruction set. The major effort is in generating the offset mask in the first place. For the invert the order case you just assume the mask as: MASK DC 256AL1(MASK+255-*) 255,254,253,...,2,1,0
Re: Base registers
Robert, I seem remember something similar..I worked a vm/VSE shop for awhile, the good ole card days, we would want a jcl card deck listing, originally written in COBOL it did a perform 1 by 1 varying, I changed it to something similar to what you mentioned, worked as well as as fast..that was a macro CICS shop too, man, a lot of storage violations...boy Scott ford www.identityforge.com On Jun 3, 2012, at 3:48 PM, "Robert A. Rosenberg" wrote: > At 12:21 -0400 on 06/03/2012, Scott Ford wrote about Re: Base registers: > >> Robert, >> >> Thank you I appreciate the help been writing assembler a long time >> but not consistently every day. I have had to wear many function >> hats, sysprog vm,VSE,CICS,vtam ,tcpip, network engineer, etc. So I >> got rusty in assembler, I think everyone on here are very >> helpful..and it is appreciated. >> >> Scott ford >> www.identityforge.com > > Some times you just run into a cute (or non-standard) trick with > assembler. My favorite is the use of the TR command to do a patterned > move (when the source and target are both under 256 bytes in length). > You fill in the target area with offsets into the source and then TR > the target using the source as the TR Table. This was a trick to flip > the order of the bytes in a field (so it reads right to left) before > MVCI was added to the instruction set. The major effort is in > generating the offset mask in the first place. For the invert the > order case you just assume the mask as: > > MASK DC 256AL1(MASK+255-*) 255,254,253,...,2,1,0
Re: Base registers
From: Chuck Arney Sent: Sunday, 3 June 2012 6:53 AM I don't think you really want your base register pointing to an odd address. You need to add 1 more to make it right and that requires another instruction. There's no need to be scared of an odd value. It is, after all, the assembler that calculates displacements. If it bothers you, make it 4092. Still no extra instruction needed. -Original Message- From: IBM Mainframe Assembler List [mailto:ASSEMBLER-LIST@LISTSERV.UGA.EDU] On Behalf Of robin Sent: Saturday, June 02, 2012 8:12 AM From: Robert A. Rosenberg Sent: Saturday, 2 June 2012 7:04 AM LAR11,2048(R12) LAR11,2048(R11) LA 11,4095(0,12) requires only one instruction (adjust USING accordingly).
Re: Base registers
On Mon, Jun 4, 2012 at 2:45 AM, robin wrote: > There's no need to be scared of an odd value. > It is, after all, the assembler that calculates displacements. > If it bothers you, make it 4092. Still no extra instruction needed. Since most of the time you just new a few more bytes anyway ;-)
Re: Base registers
On 6/3/2012 8:52 AM, John Gilmore wrote: Gerhard's point is not so easily dismissed; but 1) it is self-defeating for the technology if pushed very far: it freezes software in a posture that is now more than fifteen years behind the state of the art; and 2) the market for even moderately-priced software among MVT-using hobbyists is very small. Still, seriously pursued dual-path software is is certainly viable. Supporting obsolete systems is a labor of love; I see no reason to impute financial objectives to their support. For example, Dave Cole has offered the MVS version of XDC for free, except no copy of this version of it has been found. As a now certified élitist, I am nevertheless unrepentent. Urgings about compatibility requirements and the like usually hide Luddite impulses, reluctance to accommodate the (not very) new, which account for a good many of the troubles that the mainframe community is experiencing. If I were responsible for a Fortune 500 company, I would be leery of making changes solely to avoid consideration as a Luddite. And for a mission critical business application, I certainly would prefer a machine optimized compiler over assembler (programmers should provide global optimization). My approach is pragmatic - use the appropriate tools best suited for the task at hand. If that means consciously deciding on old code reuse rather than new development, that may prove faster and cheaper in the long run. But I agree that old code should not be used simply because a programmer is not up to speed on new features. Gerhard Postpischil Bradford, VT
Re: Base registers
>> Odd (or nonconsecutive) Base registers Please- If there is ever a need to calculate a ZAP to this code with crossing the boundary between two registers- it will cause major headaches- The PLS compiler does it- but this is not an argument to do it. -- Martin Pi_cap_CPU - all you ever need around MWLC/SCRT/CMT in z/VSE more at http://www.picapcpu.de
Re: Base registers
>> ... said enough, maybe too much, about this topic Since the thread is still not dead- here is my point: Use of relative & immediate feature is the way to go. The time it takes to convert a module from Bxx type to Jxxx is neglectable in comparison to finding an unused register (or restructuring to get one). The major roadblock is codealterations (switches in code) = not reentrant. Which I assume is not the case for a IRREVX01 exit (shudder: and sold as a part of a product). And to those refusing to use "new" techniques. How old has a feature to be to be usable for code? This is more than 17 years old (*). I have yet to find that VSE-customer that pays for Vendor-software and has no money for hardware or operating-system-software AND demands new features. (*) I have been bitten by a DR machine that did not have the feature some 17 years ago- it might be around longer. -- Martin Pi_cap_CPU - all you ever need around MWLC/SCRT/CMT in z/VSE more at http://www.picapcpu.de
Re: Base registers
From: Rob van der Heij Sent: Monday, 4 June 2012 4:41 PM Since most of the time you just new a few more bytes anyway ;-) ?? Doesn't make sense.
Re: Base registers
>> ?? Doesn't make sense. Since most of the time you just NEED a few more bytes anyway Better now? -- Martin Pi_cap_CPU - all you ever need around MWLC/SCRT/CMT in z/VSE more at http://www.picapcpu.de
Re: Base registers
From: Rob van der Heij Sent: Monday, 4 June 2012 4:41 PM On Mon, Jun 4, 2012 at 2:45 AM, robin wrote: There's no need to be scared of an odd value. It is, after all, the assembler that calculates displacements. If it bothers you, make it 4092. Still no extra instruction needed. Since most of the time you just new a few more bytes anyway ;-) Even with the nonsense word changed a la Martin, your response still doesn't make sense.
Re: Base registers
On Mon, Jun 4, 2012 at 12:20 PM, robin wrote: > Even with the nonsense word changed a la Martin, > your response still doesn't make sense. Sorry, as Martin noticed my head and fingers had disconnected... Typically I find the module just slightly extend the 4K after a small change... 4092 or 4094 is a lot compared to what you need at that moment. But for future debugging it is much more obvious when the two base registers are really 4K apart. Seems to me that saving that single instruction is not worth the trouble. Rob
Re: Base registers
> I have yet to find that VSE-customer that pays for Vendor-software and > has no money for hardware or operating-system-software AND demands > new features. We have several that meet this requirement. Think about the VSE 2.6 on P390 ESL boxes. Tony Thigpen -Original Message - From: Martin Trübner Sent: 06/04/2012 03:27 AM ... said enough, maybe too much, about this topic Since the thread is still not dead- here is my point: Use of relative& immediate feature is the way to go. The time it takes to convert a module from Bxx type to Jxxx is neglectable in comparison to finding an unused register (or restructuring to get one). The major roadblock is codealterations (switches in code) = not reentrant. Which I assume is not the case for a IRREVX01 exit (shudder: and sold as a part of a product). And to those refusing to use "new" techniques. How old has a feature to be to be usable for code? This is more than 17 years old (*). I have yet to find that VSE-customer that pays for Vendor-software and has no money for hardware or operating-system-software AND demands new features. (*) I have been bitten by a DR machine that did not have the feature some 17 years ago- it might be around longer. -- Martin Pi_cap_CPU - all you ever need around MWLC/SCRT/CMT in z/VSE more at http://www.picapcpu.de
Re: Base registers
I think it may have been an parody of a joke:
Question: "How much money is enough?"
Answer: "Just a little bit more."
In assembler, it is "how many bytes do you need to be resolvable to a valid
offset?" "just a few more."
It's why I (as a customer only), love doing "baseless" programming. I separate
the code and data sections, using a base register only for the data section. I
load the base register using an LAY (if contained in the same CSECT). I now
also use "pure" coding techniques. In this case "pure" means that I never store
into the CSECT itself. Actually, I use an RSECT. This is a requirement for
writing DLLs in HLASM. Which I have successfully done ("beause I can").
--
John McKown
Systems Engineer IV
IT
Administrative Services Group
HealthMarkets(r)
9151 Boulevard 26 * N. Richland Hills * TX 76010
(817) 255-3225 phone *
john.mck...@healthmarkets.com * www.HealthMarkets.com
Confidentiality Notice: This e-mail message may contain confidential or
proprietary information. If you are not the intended recipient, please contact
the sender by reply e-mail and destroy all copies of the original message.
HealthMarkets(r) is the brand name for products underwritten and issued by the
insurance subsidiaries of HealthMarkets, Inc. -The Chesapeake Life Insurance
Company(r), Mid-West National Life Insurance Company of TennesseeSM and The
MEGA Life and Health Insurance Company.SM
> -Original Message-
> From: IBM Mainframe Assembler List
> [mailto:ASSEMBLER-LIST@LISTSERV.UGA.EDU] On Behalf Of robin
> Sent: Monday, June 04, 2012 5:20 AM
> To: ASSEMBLER-LIST@LISTSERV.UGA.EDU
> Subject: Re: Base registers
>
> From: Rob van der Heij
> Sent: Monday, 4 June 2012 4:41 PM
>
> On Mon, Jun 4, 2012 at 2:45 AM, robin wrote:
>
> >> There's no need to be scared of an odd value.
> >> It is, after all, the assembler that calculates displacements.
> >> If it bothers you, make it 4092. Still no extra
> instruction needed.
>
> >Since most of the time you just new a few more bytes anyway ;-)
>
> Even with the nonsense word changed a la Martin,
> your response still doesn't make sense.
>
>
Re: Base registers
Tony, >> > I have yet to find that VSE-customer that pays for Vendor-software and > has no money for hardware or operating-system-software AND demands > new features. We have several that meet this requirement. Think about the VSE 2.6 on P390 ESL boxes. I can imagine customers running that- but asking for new features? This is like asking for an aircondition in a Ford T. -- Martin Pi_cap_CPU - all you ever need around MWLC/SCRT/CMT in z/VSE more at http://www.picapcpu.de
Re: Base registers
Martin, As the say options are like ...everyone had one...your entitled to yours Scott ford www.identityforge.com On Jun 4, 2012, at 8:27 AM, Martin Truebner wrote: > Tony, > I have yet to find that VSE-customer that pays for Vendor-software and >> has no money for hardware or operating-system-software AND demands >> new features. > We have several that meet this requirement. Think about the VSE 2.6 on > P390 ESL boxes. > > I can imagine customers running that- but asking for new features? > > This is like asking for an aircondition in a Ford T. > > -- > Martin > > Pi_cap_CPU - all you ever need around MWLC/SCRT/CMT in z/VSE > more at http://www.picapcpu.de
Re: Base registers
That was a typo, opinionsour customers usually dictate needs many times Scott ford www.identityforge.com On Jun 4, 2012, at 9:00 AM, Scott Ford wrote: > Martin, > > As the say options are like ...everyone had one...your entitled to yours > > Scott ford > www.identityforge.com > > On Jun 4, 2012, at 8:27 AM, Martin Truebner wrote: > >> Tony, >> > I have yet to find that VSE-customer that pays for Vendor-software and >>> has no money for hardware or operating-system-software AND demands >>> new features. >> We have several that meet this requirement. Think about the VSE 2.6 on >> P390 ESL boxes. >> >> I can imagine customers running that- but asking for new features? >> >> This is like asking for an aircondition in a Ford T. >> >> -- >> Martin >> >> Pi_cap_CPU - all you ever need around MWLC/SCRT/CMT in z/VSE >> more at http://www.picapcpu.de
Re: Base registers
Thanks Gerhard, I feel the same way, especially when you work for a small company and your the only one writing Assembler, with customers asking for changes Scott ford www.identityforge.com On Jun 4, 2012, at 2:49 AM, Gerhard Postpischil wrote: > On 6/3/2012 8:52 AM, John Gilmore wrote: >> Gerhard's point is not so easily dismissed; but 1) it is >> self-defeating for the technology if pushed very far: it freezes >> software in a posture that is now more than fifteen years behind the >> state of the art; and 2) the market for even moderately-priced >> software among MVT-using hobbyists is very small. Still, seriously >> pursued dual-path software is is certainly viable. > > Supporting obsolete systems is a labor of love; I see no reason > to impute financial objectives to their support. For example, > Dave Cole has offered the MVS version of XDC for free, except no > copy of this version of it has been found. > >> As a now certified élitist, I am nevertheless unrepentent. Urgings >> about compatibility requirements and the like usually hide Luddite >> impulses, reluctance to accommodate the (not very) new, which account >> for a good many of the troubles that the mainframe community is >> experiencing. > > If I were responsible for a Fortune 500 company, I would be > leery of making changes solely to avoid consideration as a > Luddite. And for a mission critical business application, I > certainly would prefer a machine optimized compiler over > assembler (programmers should provide global optimization). > > My approach is pragmatic - use the appropriate tools best suited > for the task at hand. If that means consciously deciding on old > code reuse rather than new development, that may prove faster > and cheaper in the long run. But I agree that old code should > not be used simply because a programmer is not up to speed on > new features. > > Gerhard Postpischil > Bradford, VT
Re: Base registers
I have seen many old IBM modules (in dumps, microfiche, etc.) in which the first few instructions are something like this: MODULE USING *,R15 LRR12,R15 LA R11,4095(,R12) DROP R15 USING MODULE,R12 USING MODULE+4095,R11 This allows 8,191 bytes of local addressability to be established with only two instructions for a total length of 6 bytes of executable code. This kind of code was "state of the art" long ago when each additional byte of storage was vastly more expensive than that same additional byte is today. Back in those days there was no real storage or virtual storage, just "storage", and modules had to be as small as possible. And many modules written way back then are still alive and well inside z/whatever-its-latest-name-is. Yes, the odd offsets look weird, but the weird look does not prevent the DAT's ability to add the base register's contents to the displacement and generate the correct address. When one is writing new code, one is free to be elitist and exploit the latest and greatest instructions that are available on the processors on which the code is expected to run. When one is working with old code, or even new code written by Luddites, denigrating the technology used does not really help in understanding what the module is doing. Bill Fairchild Programmer Rocket Software 408 Chamberlain Park Lane * Franklin, TN 37069-2526 * USA t: +1.617.614.4503 * e: bfairch...@rocketsoftware.com * w: www.rocketsoftware.com -Original Message- From: IBM Mainframe Assembler List [mailto:ASSEMBLER-LIST@LISTSERV.UGA.EDU] On Behalf Of robin Sent: Sunday, June 03, 2012 7:45 PM To: ASSEMBLER-LIST@LISTSERV.UGA.EDU Subject: Re: Base registers From: Chuck Arney Sent: Sunday, 3 June 2012 6:53 AM >I don't think you really want your base register pointing to an odd address. >You need to add 1 more to make it right and that requires another >instruction. There's no need to be scared of an odd value. It is, after all, the assembler that calculates displacements.
Re: Base registers
Bill, Amen, I first wrong BAL on a 360/20, didn't have the 1401 exposure ..man half words were real important Scott ford www.identityforge.com On Jun 4, 2012, at 11:15 AM, Bill Fairchild wrote: > I have seen many old IBM modules (in dumps, microfiche, etc.) in which the > first few instructions are something like this: > MODULE USING *,R15 >LRR12,R15 >LA R11,4095(,R12) >DROP R15 >USING MODULE,R12 >USING MODULE+4095,R11 > This allows 8,191 bytes of local addressability to be established with only > two instructions for a total length of 6 bytes of executable code. This kind > of code was "state of the art" long ago when each additional byte of storage > was vastly more expensive than that same additional byte is today. Back in > those days there was no real storage or virtual storage, just "storage", and > modules had to be as small as possible. And many modules written way back > then are still alive and well inside z/whatever-its-latest-name-is. > Yes, the odd offsets look weird, but the weird look does not prevent the > DAT's ability to add the base register's contents to the displacement and > generate the correct address. > When one is writing new code, one is free to be elitist and exploit the > latest and greatest instructions that are available on the processors on > which the code is expected to run. > When one is working with old code, or even new code written by Luddites, > denigrating the technology used does not really help in understanding what > the module is doing. > > Bill Fairchild > Programmer > Rocket Software > 408 Chamberlain Park Lane * Franklin, TN 37069-2526 * USA > t: +1.617.614.4503 * e: bfairch...@rocketsoftware.com * w: > www.rocketsoftware.com > > -Original Message- > From: IBM Mainframe Assembler List [mailto:ASSEMBLER-LIST@LISTSERV.UGA.EDU] > On Behalf Of robin > Sent: Sunday, June 03, 2012 7:45 PM > To: ASSEMBLER-LIST@LISTSERV.UGA.EDU > Subject: Re: Base registers > > From: Chuck Arney > Sent: Sunday, 3 June 2012 6:53 AM > >> I don't think you really want your base register pointing to an odd address. >> You need to add 1 more to make it right and that requires another >> instruction. > > There's no need to be scared of an odd value. > It is, after all, the assembler that calculates displacements.
Re: Base registers
The code we use here at my job usually has very large modules and used several base registers. I have seen the following technique used: MYCSECT CSECT USING *,R15 B BYID ID DCC'module-name' BASESDCA(MYCSECT) DCA(MYCSECT+4096) DCA(MYCSECT+2*4096) DCA(MYCSECT+3*4096) BYID DS0H LMR9,R12,BASES DROP R15 USING MYCSECT,R9,R10,R11,R12 One CPU instruction to load all base registers. I don't really like dealing with such large programs but I can't do anything about it so I'll support it the way it it. John
Re: Base registers
At 18:44 + on 06/04/2012, Bodoh John Robert wrote about Re: Base registers: The code we use here at my job usually has very large modules and used several base registers. I have seen the following technique used: MYCSECT CSECT USING *,R15 B BYID ID DCC'module-name' BASESDCA(MYCSECT) DCA(MYCSECT+4096) DCA(MYCSECT+2*4096) DCA(MYCSECT+3*4096) BYID DS0H LMR9,R12,BASES DROP R15 USING MYCSECT,R9,R10,R11,R12 One CPU instruction to load all base registers. I don't really like dealing with such large programs but I can't do anything about it so I'll support it the way it it. John By replacing the MYCSECT references with BYID+4 you can pick up some extra free addressability (at the expense of having the offsets non-zero based).
Re: Base registers
On Tue, 5 Jun 2012 00:32:32 -0400, Robert A. Rosenberg wrote: >At 18:44 + on 06/04/2012, Bodoh John Robert wrote about Re: Base registers: > >>MYCSECT CSECT >> USING *,R15 >> B BYID >>ID DCC'module-name' >>BASESDCA(MYCSECT) >> DCA(MYCSECT+4096) >> DCA(MYCSECT+2*4096) >> DCA(MYCSECT+3*4096) >>BYID DS0H >> LMR9,R12,BASES >> DROP R15 >> USING MYCSECT,R9,R10,R11,R12 >> >>One CPU instruction to load all base registers. I don't really like >>dealing with such large programs but I can't do anything about it so >>I'll support it the way it it. >> >>John > >By replacing the MYCSECT references with BYID+4 you can pick up some >extra free addressability (at the expense of having the offsets >non-zero based). Yuck. By adding a few LOCTR instructions, all of the data can be grouped at the beginning of the program. If relative branches are used in the code, the base register is needed only for data and literals. MYCSECT CSECT USING *,R15 J BYID DATA LOCTR ID DCC'module-name' BASESDCA(MYCSECT) DCA(MYCSECT+4096) DCA(MYCSECT+2*4096) DCA(MYCSECT+3*4096) BYID DS0H CODE LOCTR LMR9,R12,BASES DROP R15 USING MYCSECT,R9,R10,R11,R12 ... DATA LOCTR * additional DC statements are coded here LTORG You can probably do with only one base register for the data. -- Tom Marchant
Re: Base registers
> -Original Message- > From: IBM Mainframe Assembler List > [mailto:ASSEMBLER-LIST@LISTSERV.UGA.EDU] On Behalf Of Robert > A. Rosenberg > Sent: Monday, June 04, 2012 11:33 PM > To: ASSEMBLER-LIST@LISTSERV.UGA.EDU > Subject: Re: Base registers > > > By replacing the MYCSECT references with BYID+4 you can pick up some > extra free addressability (at the expense of having the offsets > non-zero based). > Back long ago, before I started writing "pure" or "reentrant" code, I would always start my routines: MYCSECT CSECT USING *,R15 SAVE (14,12),,* BAL R1,AROUND SAVEAREA DC 18A(0) ST R13,4(,R1) ST R1,8(,R13) LR R13,R1 DROP R15 USING SAVEAREA,R13 L R1,4(,R13) L R1,20(,R1) RESTORE ORIGINAL R1 ... I don't do this any more, in most cases. Just to horrify some of you, I now generally code HLASM enabled code, using all the CEExxx macros. === A reason that my manager did not like the above was that if the base register points to the beginning of the CSECT, and you use the SAVE macro properly, you can look at the storage pointed to by the base register to see the "eyecatcher" of the program which abended. The SAVE would usually look something like: SAVE (14,12),,'&SYSECT &SYSCLOCK' And he always used base registers in reverse order: R12, R11, R10, and so on. -- John McKown Systems Engineer IV IT Administrative Services Group HealthMarkets(r) 9151 Boulevard 26 * N. Richland Hills * TX 76010 (817) 255-3225 phone * john.mck...@healthmarkets.com * www.HealthMarkets.com Confidentiality Notice: This e-mail message may contain confidential or proprietary information. If you are not the intended recipient, please contact the sender by reply e-mail and destroy all copies of the original message. HealthMarkets(r) is the brand name for products underwritten and issued by the insurance subsidiaries of HealthMarkets, Inc. -The Chesapeake Life Insurance Company(r), Mid-West National Life Insurance Company of TennesseeSM and The MEGA Life and Health Insurance Company.SM
Re: Base registers
OOPS, the statement: AROUND DS 0H should be before the ST R13,4(,R1) -- John McKown Systems Engineer IV IT Administrative Services Group HealthMarkets(r) 9151 Boulevard 26 * N. Richland Hills * TX 76010 (817) 255-3225 phone * john.mck...@healthmarkets.com * www.HealthMarkets.com Confidentiality Notice: This e-mail message may contain confidential or proprietary information. If you are not the intended recipient, please contact the sender by reply e-mail and destroy all copies of the original message. HealthMarkets(r) is the brand name for products underwritten and issued by the insurance subsidiaries of HealthMarkets, Inc. -The Chesapeake Life Insurance Company(r), Mid-West National Life Insurance Company of TennesseeSM and The MEGA Life and Health Insurance Company.SM > -Original Message- > From: IBM Mainframe Assembler List > [mailto:ASSEMBLER-LIST@LISTSERV.UGA.EDU] On Behalf Of McKown, John > Sent: Tuesday, June 05, 2012 7:08 AM > To: ASSEMBLER-LIST@LISTSERV.UGA.EDU > Subject: Re: Base registers > > > -Original Message- > > From: IBM Mainframe Assembler List > > [mailto:ASSEMBLER-LIST@LISTSERV.UGA.EDU] On Behalf Of Robert > > A. Rosenberg > > Sent: Monday, June 04, 2012 11:33 PM > > To: ASSEMBLER-LIST@LISTSERV.UGA.EDU > > Subject: Re: Base registers > > > > > > > By replacing the MYCSECT references with BYID+4 you can pick up some > > extra free addressability (at the expense of having the offsets > > non-zero based). > > > Back long ago, before I started writing "pure" or "reentrant" > code, I would always start my routines: > > MYCSECT CSECT > USING *,R15 > SAVE (14,12),,* > BAL R1,AROUND > SAVEAREA DC 18A(0) > ST R13,4(,R1) > ST R1,8(,R13) > LR R13,R1 > DROP R15 > USING SAVEAREA,R13 > L R1,4(,R13) > L R1,20(,R1) RESTORE ORIGINAL R1 > ... > > I don't do this any more, in most cases. Just to horrify some > of you, I now generally code HLASM enabled code, using all > the CEExxx macros. > > === > > A reason that my manager did not like the above was that if > the base register points to the beginning of the CSECT, and > you use the SAVE macro properly, you can look at the storage > pointed to by the base register to see the "eyecatcher" of > the program which abended. The SAVE would usually look something like: > > SAVE (14,12),,'&SYSECT &SYSCLOCK' > > And he always used base registers in reverse order: R12, R11, > R10, and so on. > > > -- > John McKown > Systems Engineer IV > IT > > Administrative Services Group > > HealthMarkets(r) > > 9151 Boulevard 26 * N. Richland Hills * TX 76010 > (817) 255-3225 phone * > john.mck...@healthmarkets.com * www.HealthMarkets.com > > Confidentiality Notice: This e-mail message may contain > confidential or proprietary information. If you are not the > intended recipient, please contact the sender by reply e-mail > and destroy all copies of the original message. > HealthMarkets(r) is the brand name for products underwritten > and issued by the insurance subsidiaries of HealthMarkets, > Inc. -The Chesapeake Life Insurance Company(r), Mid-West > National Life Insurance Company of TennesseeSM and The MEGA > Life and Health Insurance Company.SM > >
Re: Base registers
Use R14 instead of R1 to avoid the reloading of R1 (parameter list pointer). We always set the new save are to all X'FF' so it stands out in a DUMP. -Original Message- From: IBM Mainframe Assembler List [mailto:ASSEMBLER-LIST@LISTSERV.UGA.EDU] On Behalf Of McKown, John Sent: Tuesday, June 05, 2012 7:18 AM To: ASSEMBLER-LIST@LISTSERV.UGA.EDU Subject: Re: Base registers OOPS, the statement: AROUND DS 0H should be before the ST R13,4(,R1) -- John McKown Systems Engineer IV IT Administrative Services Group HealthMarkets(r) 9151 Boulevard 26 * N. Richland Hills * TX 76010 (817) 255-3225 phone * john.mck...@healthmarkets.com * www.HealthMarkets.com Confidentiality Notice: This e-mail message may contain confidential or proprietary information. If you are not the intended recipient, please contact the sender by reply e-mail and destroy all copies of the original message. HealthMarkets(r) is the brand name for products underwritten and issued by the insurance subsidiaries of HealthMarkets, Inc. -The Chesapeake Life Insurance Company(r), Mid-West National Life Insurance Company of TennesseeSM and The MEGA Life and Health Insurance Company.SM > -Original Message- > From: IBM Mainframe Assembler List > [mailto:ASSEMBLER-LIST@LISTSERV.UGA.EDU] On Behalf Of McKown, John > Sent: Tuesday, June 05, 2012 7:08 AM > To: ASSEMBLER-LIST@LISTSERV.UGA.EDU > Subject: Re: Base registers > > > -Original Message- > > From: IBM Mainframe Assembler List > > [mailto:ASSEMBLER-LIST@LISTSERV.UGA.EDU] On Behalf Of Robert > > A. Rosenberg > > Sent: Monday, June 04, 2012 11:33 PM > > To: ASSEMBLER-LIST@LISTSERV.UGA.EDU > > Subject: Re: Base registers > > > > > > > By replacing the MYCSECT references with BYID+4 you can pick up some > > extra free addressability (at the expense of having the offsets > > non-zero based). > > > Back long ago, before I started writing "pure" or "reentrant" > code, I would always start my routines: > > MYCSECT CSECT > USING *,R15 > SAVE (14,12),,* > BAL R1,AROUND > SAVEAREA DC 18A(0) > ST R13,4(,R1) > ST R1,8(,R13) > LR R13,R1 > DROP R15 > USING SAVEAREA,R13 > L R1,4(,R13) > L R1,20(,R1) RESTORE ORIGINAL R1 > ... > > I don't do this any more, in most cases. Just to horrify some > of you, I now generally code HLASM enabled code, using all > the CEExxx macros. > > === > > A reason that my manager did not like the above was that if > the base register points to the beginning of the CSECT, and > you use the SAVE macro properly, you can look at the storage > pointed to by the base register to see the "eyecatcher" of > the program which abended. The SAVE would usually look something like: > > SAVE (14,12),,'&SYSECT &SYSCLOCK' > > And he always used base registers in reverse order: R12, R11, > R10, and so on. > > > -- > John McKown > Systems Engineer IV > IT > > Administrative Services Group > > HealthMarkets(r) > > 9151 Boulevard 26 * N. Richland Hills * TX 76010 > (817) 255-3225 phone * > john.mck...@healthmarkets.com * www.HealthMarkets.com > > Confidentiality Notice: This e-mail message may contain > confidential or proprietary information. If you are not the > intended recipient, please contact the sender by reply e-mail > and destroy all copies of the original message. > HealthMarkets(r) is the brand name for products underwritten > and issued by the insurance subsidiaries of HealthMarkets, > Inc. -The Chesapeake Life Insurance Company(r), Mid-West > National Life Insurance Company of TennesseeSM and The MEGA > Life and Health Insurance Company.SM > >
Re: Base registers
At 07:57 -0400 on 06/05/2012, Tom Marchant wrote about Re: Base registers: Yuck. By adding a few LOCTR instructions, all of the data can be grouped at the beginning of the program. If relative branches are used in the code, the base register is needed only for data and literals. MYCSECT CSECT USING *,R15 J BYID DATA LOCTR ID DCC'module-name' BASESDCA(MYCSECT) DCA(MYCSECT+4096) DCA(MYCSECT+2*4096) DCA(MYCSECT+3*4096) BYID DS0H CODE LOCTR LMR9,R12,BASES DROP R15 USING MYCSECT,R9,R10,R11,R12 ... DATA LOCTR * additional DC statements are coded here LTORG You can probably do with only one base register for the data. That BYID needs to be placed after the CODE LOCTR or you will be jumping into your DATA area and a DC or the LTORG.
Re: Base registers
Guys, John and spoke where can you find a good sample of baseless assembler code ? I want to convert some of ours , need a sample to get me started, it would help... Scott ford www.identityforge.com On Jun 5, 2012, at 2:51 PM, "Robert A. Rosenberg" wrote: > At 07:57 -0400 on 06/05/2012, Tom Marchant wrote about Re: Base registers: > >> Yuck. By adding a few LOCTR instructions, all of the data can be grouped at >> the beginning of the program. If relative branches are used in the code, >> the base register is needed only for data and literals. >> >> MYCSECT CSECT >> USING *,R15 >> J BYID >> DATA LOCTR >> ID DCC'module-name' >> BASESDCA(MYCSECT) >> DCA(MYCSECT+4096) >> DCA(MYCSECT+2*4096) >> DCA(MYCSECT+3*4096) >> BYID DS0H >> CODE LOCTR >> LMR9,R12,BASES >> DROP R15 >> USING MYCSECT,R9,R10,R11,R12 >> ... >> DATA LOCTR >> * additional DC statements are coded here >> LTORG >> >> You can probably do with only one base register for the data. > > That BYID needs to be placed after the CODE LOCTR or you will be > jumping into your DATA area and a DC or the LTORG.
Re: Base registers
On Tue, 5 Jun 2012 14:51:48 -0400, Robert A. Rosenberg wrote: >At 07:57 -0400 on 06/05/2012, Tom Marchant wrote about Re: Base registers: > >>By adding a few LOCTR instructions, all of the data can be grouped at >>the beginning of the program. If relative branches are used in the code, >>the base register is needed only for data and literals. >> >>MYCSECT CSECT >> USING *,R15 >> J BYID >>DATA LOCTR >>ID DCC'module-name' >>BASESDCA(MYCSECT) >> DCA(MYCSECT+4096) >> DCA(MYCSECT+2*4096) >> DCA(MYCSECT+3*4096) >>BYID DS0H >>CODE LOCTR >> LMR9,R12,BASES >> DROP R15 >> USING MYCSECT,R9,R10,R11,R12 >>... >>DATA LOCTR >>* additional DC statements are coded here >> LTORG >> >>You can probably do with only one base register for the data. > >That BYID needs to be placed after the CODE LOCTR or you will be >jumping into your DATA area and a DC or the LTORG. Right. Thanks for the correction. The above should be coded as: >>MYCSECT CSECT >> USING *,R15 >> J BYID >>DATA LOCTR >>ID DCC'module-name' >>BASESDCA(MYCSECT) >> DCA(MYCSECT+4096) >> DCA(MYCSECT+2*4096) >> DCA(MYCSECT+3*4096) >>CODE LOCTR >>BYID DS0H >> LMR9,R12,BASES >> DROP R15 >> USING MYCSECT,R9,R10,R11,R12 >>... >>DATA LOCTR >>* additional DC statements are coded here >> LTORG -- Tom Marchant
Re: Base registers
Damn, you had to include the word "good". If you want some baseless code which is LE enabled and is designed to run as a z/OS UNIX command, you can download my UNIX "alpha" code from the CBT. It is FILE864 at http://www.cbttape.org/updates.htm I also attached a non-LE baseless HLASM program source code to this email. It is a do nothing skeleton. I think the listserv will strip it out. -- John McKown Systems Engineer IV IT Administrative Services Group HealthMarkets(r) 9151 Boulevard 26 * N. Richland Hills * TX 76010 (817) 255-3225 phone * john.mck...@healthmarkets.com * www.HealthMarkets.com Confidentiality Notice: This e-mail message may contain confidential or proprietary information. If you are not the intended recipient, please contact the sender by reply e-mail and destroy all copies of the original message. HealthMarkets(r) is the brand name for products underwritten and issued by the insurance subsidiaries of HealthMarkets, Inc. -The Chesapeake Life Insurance Company(r), Mid-West National Life Insurance Company of TennesseeSM and The MEGA Life and Health Insurance Company.SM > -Original Message- > From: IBM Mainframe Assembler List > [mailto:ASSEMBLER-LIST@LISTSERV.UGA.EDU] On Behalf Of Scott Ford > Sent: Tuesday, June 05, 2012 3:00 PM > To: ASSEMBLER-LIST@LISTSERV.UGA.EDU > Subject: Re: Base registers > > Guys, > > John and spoke where can you find a good sample of baseless > assembler code ? > I want to convert some of ours , need a sample to get me > started, it would help... > > Scott ford > www.identityforge.com > > On Jun 5, 2012, at 2:51 PM, "Robert A. Rosenberg" > wrote: > > > At 07:57 -0400 on 06/05/2012, Tom Marchant wrote about Re: > Base registers: > > > >> Yuck. By adding a few LOCTR instructions, all of the data > can be grouped at > >> the beginning of the program. If relative branches are > used in the code, > >> the base register is needed only for data and literals. > >> > >> MYCSECT CSECT > >> USING *,R15 > >> J BYID > >> DATA LOCTR > >> ID DCC'module-name' > >> BASESDCA(MYCSECT) > >> DCA(MYCSECT+4096) > >> DCA(MYCSECT+2*4096) > >> DCA(MYCSECT+3*4096) > >> BYID DS0H > >> CODE LOCTR > >> LMR9,R12,BASES > >> DROP R15 > >> USING MYCSECT,R9,R10,R11,R12 > >> ... > >> DATA LOCTR > >> * additional DC statements are coded here > >> LTORG > >> > >> You can probably do with only one base register for the data. > > > > That BYID needs to be placed after the CODE LOCTR or you will be > > jumping into your DATA area and a DC or the LTORG. > >
Re: Base registers
On Tue, 5 Jun 2012 15:59:36 -0400, Scott Ford wrote:
>where can you find a good sample of baseless assembler code ?
Look for Ed Jaffe's SHARE presentation "Jumpify your code".
"Baseless" is not an accurate description, IMO. You still need
base registers to reference data. You can, however, use very few
base registers for your code simply by using relative branches
("jump"). There are one or two exceptions, depending upon your
hardware. One is that there is no indexed relative branch. It's
hard for me to imagine what such a thing would do anyway. The
other is that EXRL was introduced (IIRC) on the z10. The standard
EXecute instruction requires a base register.
--
Tom Marchant
Re: Base registers
Tom,
Thank you, I have his great presentation and John gave me so code, and who said
we old folks cant change
Scott J Ford
Software Engineer
http://www.identityforge.com
From: Tom Marchant
To: ASSEMBLER-LIST@LISTSERV.UGA.EDU
Sent: Tuesday, June 5, 2012 5:19 PM
Subject: Re: Base registers
On Tue, 5 Jun 2012 15:59:36 -0400, Scott Ford wrote:
>where can you find a good sample of baseless assembler code ?
Look for Ed Jaffe's SHARE presentation "Jumpify your code".
"Baseless" is not an accurate description, IMO. You still need
base registers to reference data. You can, however, use very few
base registers for your code simply by using relative branches
("jump"). There are one or two exceptions, depending upon your
hardware. One is that there is no indexed relative branch. It's
hard for me to imagine what such a thing would do anyway. The
other is that EXRL was introduced (IIRC) on the z10. The standard
EXecute instruction requires a base register.
--
Tom Marchant
Re: Base registers
That should be some codedidnt say my eyes and fingers werent old ...
Scott J Ford
Software Engineer
http://www.identityforge.com
From: Scott Ford
To: ASSEMBLER-LIST@LISTSERV.UGA.EDU
Sent: Tuesday, June 5, 2012 5:23 PM
Subject: Re: Base registers
Tom,
Thank you, I have his great presentation and John gave me so code, and who said
we old folks cant change
Scott J Ford
Software Engineer
http://www.identityforge.com
From: Tom Marchant
To: ASSEMBLER-LIST@LISTSERV.UGA.EDU
Sent: Tuesday, June 5, 2012 5:19 PM
Subject: Re: Base registers
On Tue, 5 Jun 2012 15:59:36 -0400, Scott Ford wrote:
>where can you find a good sample of baseless assembler code ?
Look for Ed Jaffe's SHARE presentation "Jumpify your code".
"Baseless" is not an accurate description, IMO. You still need
base registers to reference data. You can, however, use very few
base registers for your code simply by using relative branches
("jump"). There are one or two exceptions, depending upon your
hardware. One is that there is no indexed relative branch. It's
hard for me to imagine what such a thing would do anyway. The
other is that EXRL was introduced (IIRC) on the z10. The standard
EXecute instruction requires a base register.
--
Tom Marchant
Re: Base registers
Here's one way to do standard EXecute without a base register:
AHI R2,-1 Minus 1 for EX
*!not yetEXRL R2,_EX_MVC_OUTPUT (Move data to output buffer)
LARL R10,_EX_MVC_OUTPUT Move data to
EXR2,0(,R10) output buffer
I believe, for most applications, "baseless" code is the way to go
moving forward when writing new code or refactoring existing code. Of
course, required are bases for save area/local stack and
constants/LTORG. A program manages bases for other data areas, just as
always -- passed-in-by-reference parameters, heap storage, control
blocks, I/O buffers, et cetera. Changing B-s to J-s, BCT to BRCT, et
cetera, is tedious but simple ("change all"). One may encounter macros
(IBM and otherwise) in existing code that, as traditionally used, have
required a base -- there are ways to deal with those.
All in all, I've found this entails a relatively small learning curve
and amount of effort and cures one of the greatest headaches associated
with assembler coding.
Doug Watkins
http://www.compuware.com/
The contents of this e-mail are intended for the named addressee only. It
contains information that may be confidential. Unless you are the named
addressee or an authorized designee, you may not copy or use it, or disclose it
to anyone else. If you received it in error please notify us immediately and
then destroy it.
From: IBM Mainframe Assembler List
[mailto:ASSEMBLER-LIST@LISTSERV.UGA.EDU] On Behalf Of Tom Marchant
Sent: Tuesday, June 05, 2012 5:20 PM
To: ASSEMBLER-LIST@LISTSERV.UGA.EDU
Subject: Re: Base registers
On Tue, 5 Jun 2012 15:59:36 -0400, Scott Ford wrote:
>where can you find a good sample of baseless assembler code ?
Look for Ed Jaffe's SHARE presentation "Jumpify your code".
"Baseless" is not an accurate description, IMO. You still need
base registers to reference data. You can, however, use very few
base registers for your code simply by using relative branches
("jump"). There are one or two exceptions, depending upon your
hardware. One is that there is no indexed relative branch. It's
hard for me to imagine what such a thing would do anyway. The
other is that EXRL was introduced (IIRC) on the z10. The standard
EXecute instruction requires a base register.
--
Tom Marchant
Re: Base registers
On Jun 6, 2012, at 07:28, Watkins, Douglas wrote: > Here's one way to do standard EXecute without a base register: > > AHI R2,-1 Minus 1 for EX > *!not yetEXRL R2,_EX_MVC_OUTPUT (Move data to output buffer) > LARL R10,_EX_MVC_OUTPUT Move data to > EXR2,0(,R10) output buffer > Well, R10 looks like a base register to me. If you're critically register-constrained, this may be a problem. You could likewise do: AHI R2,-1 Minus 1 for EX BALR R10,0 EXR2,EX_MVC_OUTPUT-*(,R10) ... in fewer bytes of code (if I got it right). -- gil
Re: Base registers
From: Watkins, Douglas Sent: Wednesday, 6 June 2012 11:28 PM Here's one way to do standard EXecute without a base register: AHI R2,-1 Minus 1 for EX Still need test for negative. *!not yetEXRL R2,_EX_MVC_OUTPUT (Move data to output buffer) LARL R10,_EX_MVC_OUTPUT Move data to EXR2,0(,R10) output buffer
Re: Base registers
> Still need test for negative. Very true -- unless, as is the case but not shown in the example below, I've already guaranteed R2 holds a value within my "safe range," which includes the maximum, too. http://www.compuware.com/ The contents of this e-mail are intended for the named addressee only. It contains information that may be confidential. Unless you are the named addressee or an authorized designee, you may not copy or use it, or disclose it to anyone else. If you received it in error please notify us immediately and then destroy it. From: IBM Mainframe Assembler List [mailto:ASSEMBLER-LIST@LISTSERV.UGA.EDU] On Behalf Of robin Sent: Wednesday, June 06, 2012 11:19 AM To: ASSEMBLER-LIST@LISTSERV.UGA.EDU Subject: Re: Base registers From: Watkins, Douglas Sent: Wednesday, 6 June 2012 11:28 PM >Here's one way to do standard EXecute without a base register: > AHI R2,-1 Minus 1 for EX Still need test for negative. >*!not yetEXRL R2,_EX_MVC_OUTPUT (Move data to output buffer) > LARL R10,_EX_MVC_OUTPUT Move data to > EXR2,0(,R10) output buffer
Re: Base registers
I had a feeling someone would point out that technicality, so... Let's not consider a temporarily-used-for-one-RX-instruction base register to be the same thing as a program-wide base register ;-) Also, a "baseless" program is less likely to be register-constrained; but, if it were, using "temp" R1 instead of, say R10, works just as well. http://www.compuware.com/ The contents of this e-mail are intended for the named addressee only. It contains information that may be confidential. Unless you are the named addressee or an authorized designee, you may not copy or use it, or disclose it to anyone else. If you received it in error please notify us immediately and then destroy it. From: IBM Mainframe Assembler List [mailto:ASSEMBLER-LIST@LISTSERV.UGA.EDU] On Behalf Of Paul Gilmartin Sent: Wednesday, June 06, 2012 11:02 AM To: ASSEMBLER-LIST@LISTSERV.UGA.EDU Subject: Re: Base registers On Jun 6, 2012, at 07:28, Watkins, Douglas wrote: > Here's one way to do standard EXecute without a base register: > > AHI R2,-1 Minus 1 for EX > *!not yetEXRL R2,_EX_MVC_OUTPUT (Move data to output buffer) > LARL R10,_EX_MVC_OUTPUT Move data to > EXR2,0(,R10) output buffer > Well, R10 looks like a base register to me. If you're critically register-constrained, this may be a problem. You could likewise do: AHI R2,-1 Minus 1 for EX BALR R10,0 EXR2,EX_MVC_OUTPUT-*(,R10) ... in fewer bytes of code (if I got it right). -- gil
Re: Base registers
Douglas, how is this (until EXRL is legalised ;-) EX R2,=S(7*16+&PACK,DWORD,0(R10)) And Yes- I am aware of problems in the literalpool having different usings - (for Robin) and I have verified that R2 is between F and 0 and that R10 is on the beginning of a string) -- Martin Pi_cap_CPU - all you ever need around MWLC/SCRT/CMT in z/VSE more at http://www.picapcpu.de
Re: Base registers
where is the length code for 0(r10) second operand? EX R2,=S(16*(7+16*(length_of_operand2-1)+&PACK,DWORD,0(R10)) IBM Mainframe Assembler List wrote on 06/06/2012 11:46:07 AM: > From: Martin Truebner > To: ASSEMBLER-LIST@LISTSERV.UGA.EDU, > Date: 06/06/2012 11:52 AM > Subject: Re: Base registers > Sent by: IBM Mainframe Assembler List > > Douglas, > > how is this (until EXRL is legalised ;-) > > EX R2,=S(7*16+&PACK,DWORD,0(R10)) > > And Yes- I am aware of problems in the literalpool having > different usings > > - (for Robin) and I have verified that R2 is between F and 0 and > that R10 is on the beginning of a string) > -- > Martin > > Pi_cap_CPU - all you ever need around MWLC/SCRT/CMT in z/VSE > more at http://www.picapcpu.de - The information contained in this communication (including any attachments hereto) is confidential and is intended solely for the personal and confidential use of the individual or entity to whom it is addressed. If the reader of this message is not the intended recipient or an agent responsible for delivering it to the intended recipient, you are hereby notified that you have received this communication in error and that any review, dissemination, copying, or unauthorized use of this information, or the taking of any action in reliance on the contents of this information is strictly prohibited. If you have received this communication in error, please notify us immediately by e-mail, and delete the original message. Thank you
Re: Base registers
Martin, Very clever! Personally, for everyday use, I prefer things a bit more "plain and simple." Though, I might consider creating a macro to incorporate such a technique: EXWOW R2,(PACK,DWORD,0(R10)) EXWOW R2,(MVC,DESTINATION_BUFFER,0(R10)) The latter being only marginally better than: EXR2,=S(&MVC,DESTINATION_BUFFER,0(R10)) Of course, one could simply locate the target instruction in with constant data that is assembled into program object, since that already has a base. I use LARL with EX because I like to keep the target in the vicinity of the code with which it belongs logically. I do like that your technique does keep the essential nature of the target very much within eyeball range. Doug http://www.compuware.com/ The contents of this e-mail are intended for the named addressee only. It contains information that may be confidential. Unless you are the named addressee or an authorized designee, you may not copy or use it, or disclose it to anyone else. If you received it in error please notify us immediately and then destroy it. From: IBM Mainframe Assembler List [mailto:ASSEMBLER-LIST@LISTSERV.UGA.EDU] On Behalf Of Martin Truebner Sent: Wednesday, June 06, 2012 11:46 AM To: ASSEMBLER-LIST@LISTSERV.UGA.EDU Subject: Re: Base registers Douglas, how is this (until EXRL is legalised ;-) EX R2,=S(7*16+&PACK,DWORD,0(R10)) And Yes- I am aware of problems in the literalpool having different usings - (for Robin) and I have verified that R2 is between F and 0 and that R10 is on the beginning of a string) -- Martin Pi_cap_CPU - all you ever need around MWLC/SCRT/CMT in z/VSE more at http://www.picapcpu.de
Re: Base registers
Length of second operand is in R2, ORed with that of the first operand, 7*16, when EX executes. http://www.compuware.com/ The contents of this e-mail are intended for the named addressee only. It contains information that may be confidential. Unless you are the named addressee or an authorized designee, you may not copy or use it, or disclose it to anyone else. If you received it in error please notify us immediately and then destroy it. From: IBM Mainframe Assembler List [mailto:ASSEMBLER-LIST@LISTSERV.UGA.EDU] On Behalf Of Kirk Talman Sent: Wednesday, June 06, 2012 12:13 PM To: ASSEMBLER-LIST@LISTSERV.UGA.EDU Subject: Re: Base registers where is the length code for 0(r10) second operand? EX R2,=S(16*(7+16*(length_of_operand2-1)+&PACK,DWORD,0(R10)) IBM Mainframe Assembler List wrote on 06/06/2012 11:46:07 AM: > From: Martin Truebner > To: ASSEMBLER-LIST@LISTSERV.UGA.EDU, > Date: 06/06/2012 11:52 AM > Subject: Re: Base registers > Sent by: IBM Mainframe Assembler List > > Douglas, > > how is this (until EXRL is legalised ;-) > > EX R2,=S(7*16+&PACK,DWORD,0(R10)) > > And Yes- I am aware of problems in the literalpool having > different usings > > - (for Robin) and I have verified that R2 is between F and 0 and > that R10 is on the beginning of a string) > -- > Martin > > Pi_cap_CPU - all you ever need around MWLC/SCRT/CMT in z/VSE > more at http://www.picapcpu.de - The information contained in this communication (including any attachments hereto) is confidential and is intended solely for the personal and confidential use of the individual or entity to whom it is addressed. If the reader of this message is not the intended recipient or an agent responsible for delivering it to the intended recipient, you are hereby notified that you have received this communication in error and that any review, dissemination, copying, or unauthorized use of this information, or the taking of any action in reliance on the contents of this information is strictly prohibited. If you have received this communication in error, please notify us immediately by e-mail, and delete the original message. Thank you
Re: Base registers
sigh been away from daily use of ASM. brain is starting to rot. which is bad as I am transitioning into doing CICS ASM. IBM Mainframe Assembler List wrote on 06/06/2012 01:15:54 PM: > From: "Watkins, Douglas" > To: ASSEMBLER-LIST@LISTSERV.UGA.EDU, > Date: 06/06/2012 01:18 PM > Subject: Re: Base registers > Sent by: IBM Mainframe Assembler List > > Length of second operand is in R2, ORed with that of the first operand, > 7*16, when EX executes. > > > > http://www.compuware.com/ > The contents of this e-mail are intended for the named addressee > only. It contains information that may be confidential. Unless you > are the named addressee or an authorized designee, you may not copy > or use it, or disclose it to anyone else. If you received it in > error please notify us immediately and then destroy it. > > From: IBM Mainframe Assembler List > [mailto:ASSEMBLER-LIST@LISTSERV.UGA.EDU] On Behalf Of Kirk Talman > Sent: Wednesday, June 06, 2012 12:13 PM > To: ASSEMBLER-LIST@LISTSERV.UGA.EDU > Subject: Re: Base registers > > where is the length code for 0(r10) second operand? > >EX R2,=S(16*(7+16*(length_of_operand2-1)+&PACK,DWORD,0(R10)) > > IBM Mainframe Assembler List wrote on > 06/06/2012 11:46:07 AM: > > > From: Martin Truebner > > To: ASSEMBLER-LIST@LISTSERV.UGA.EDU, > > Date: 06/06/2012 11:52 AM > > Subject: Re: Base registers > > Sent by: IBM Mainframe Assembler List > > > > > Douglas, > > > > how is this (until EXRL is legalised ;-) > > > > EX R2,=S(7*16+&PACK,DWORD,0(R10)) > > > > And Yes- I am aware of problems in the literalpool having > > different usings > > > > - (for Robin) and I have verified that R2 is between F and 0 and > > that R10 is on the beginning of a string) > > -- > > Martin > > > > Pi_cap_CPU - all you ever need around MWLC/SCRT/CMT in z/VSE > > more at http://www.picapcpu.de > > > - > The information contained in this communication (including any > attachments hereto) is confidential and is intended solely for the > personal and confidential use of the individual or entity to whom > it is addressed. If the reader of this message is not the intended > recipient or an agent responsible for delivering it to the intended > recipient, you are hereby notified that you have received this > communication in error and that any review, dissemination, copying, > or unauthorized use of this information, or the taking of any > action in reliance on the contents of this information is strictly > prohibited. If you have received this communication in error, > please notify us immediately by e-mail, and delete the original > message. Thank you
Re: Base registers
I suggest you learn how to use LOCTR. It's a much simpler method than using
=S() and certainly more readable. I'm not recommending in the way LOCTR was
used with DATA / CODE example although that is certainly one way.
My programs have the following in the program start macro (after the csect is
generated). &CSECT contains the CSECT name since &SYSECT and &SYSLOC are not set
when the CSECT is generated in the macro that generated the CSECT.
#EXLOCTR ,Execute instructions inserted with this location
counter
DS 0H Instructions must be halfword aligned
&CSECT LOCTR , Back to program
If you use the DATA / CODE method, then you will want this just before the CODE
LOCTR. I'm not suggesting you use this method but just in case you do.
Here is a sample call to my #EX macro below
#EX R2,'MVC SOURCE(0),DEST'
and it generates (length is checked where length was calculated)
BCTR R2,0Length relative to 0
EX R2,#EX_xxx Execute the instruction
LAR2,1(R2) Restore length (does not modify CC)
#EX LOCTR ,
#EX_xxx MVC SOURCE(0),DEST
MYPGM LOCTR , Back to program
Hope this helps, Jon.
.**
.* Desc: Execute instruction
.*
.* Copyright: 2010-2012 Jon Perryman
.*
.* Function:
.* Makes the length relative to 0 and executes the
.* specified instruction using the specified length.
.* The register is restored back to a length relative
.* to 1.
.*
.* Change log:
.* 2/12/2010 JPPCreated
.**
MACRO ,
&LABEL #EX &RELATIVE=1 Specify 0 if reg already rel=0
AIF ('&LABEL' EQ '').NOLABEL
&LABEL EQU *
.NOLABEL ANOP *
AIF ('&RELATIVE' EQ '0').LEN_OK1
BCTR &SYSLIST(1),0 Relative to 0
.LEN_OK1 ANOP ,
EX&SYSLIST(1),#EX_&SYSNDX
AIF ('&RELATIVE' EQ '0').LEN_OK2
LA&SYSLIST(1),1(&SYSLIST(1)) Relative to 1
.LEN_OK2 ANOP ,
&WORKSETC DEQUOTE('&SYSLIST(2)')
&WORKN SETA INDEX('&WORK',' ')
&WORK2 SETC '&WORK'(&WORKN+1,99)
&WORK2 SETC DCVAL('&WORK2')
&WORKSETC '&WORK'(1,&WORKN-1)
#EX LOCTR , Group #EX
#EX_&SYSNDX &WORK &WORK2
&SYSLOC LOCTR , Restore location counter
MEXIT ,
MEND ,
---
On Behalf Of Martin Truebner
Sent: Wednesday, June 06, 2012 11:46 AM
Subject: Re: Base registers
how is this (until EXRL is legalised ;-)
EX R2,=S(7*16+&PACK,DWORD,0(R10))
And Yes- I am aware of problems in the literalpool having
different usings
Re: Base registers
Jon, >> I suggest you learn how to use LOCTR. NO COMMENT >> It's a much simpler method than using =S() and certainly more readable. Simpler- maybe! more readable- judge it here: My suggestion: a SETC as intro and EX R2,=S(&PKA,QWORD,0(R10)) . or (more readable): a CODE LOCTR and DEFS LOCTR as intro and EX R2,PKA_02 DEFS LOCTR PKA_02 PKA QWORD,0(0,R10) CODE LOCTR plus the extra of inventing unique names .. I like the expression "in eyeball range" ;-) -- Martin Pi_cap_CPU - all you ever need around MWLC/SCRT/CMT in z/VSE more at http://www.picapcpu.de
Re: Base registers
Let's not get in a spitting match. anyone can write bad code as an example of why someone else's style is bad. I am pretty sure I could write some pretty unintelligible code if I tried using any of the suggested methods. That doesn't mean a style is bad, it probably means I am not using it as intended. -Original Message- From: IBM Mainframe Assembler List [mailto:ASSEMBLER-LIST@LISTSERV.UGA.EDU] On Behalf Of Martin Truebner Sent: Wednesday, June 06, 2012 11:58 AM To: ASSEMBLER-LIST@LISTSERV.UGA.EDU Subject: Re: Base registers Jon, >> I suggest you learn how to use LOCTR. NO COMMENT >> It's a much simpler method than using =S() and certainly more readable. Simpler- maybe! more readable- judge it here: My suggestion: a SETC as intro and EX R2,=S(&PKA,QWORD,0(R10)) . or (more readable): a CODE LOCTR and DEFS LOCTR as intro and EX R2,PKA_02 DEFS LOCTR PKA_02 PKA QWORD,0(0,R10) CODE LOCTR plus the extra of inventing unique names .. I like the expression "in eyeball range" ;-) -- Martin Pi_cap_CPU - all you ever need around MWLC/SCRT/CMT in z/VSE more at http://www.picapcpu.de
Re: Base registers
On Wed, 6 Jun 2012 20:57:47 +0200 Martin Truebner wrote: :>>> I suggest you learn how to use LOCTR. :>NO COMMENT It is good advice. :>>> It's a much simpler method than using :>=S() and certainly more readable. :>Simpler- maybe! more readable- judge it here: :> :>My suggestion: :>a SETC as intro :>and EX R2,=S(&PKA,QWORD,0(R10)) :>. :>or (more readable): :>a CODE LOCTR and DEFS LOCTR as intro :>and :>EX R2,PKA_02 :>DEFS LOCTR :>PKA_02 PKA QWORD,0(0,R10) :>CODE LOCTR I have a @DC macro as well as an @EX macro to allow the constants to be placed in the source file near the code that uses it and the assembler neatly packages all the constant data (including the EXed instructions) together. Isn't @EXR2,PKA,QSWORD,0(0,R10) clearer? -- Binyamin Dissen http://www.dissensoftware.com Director, Dissen Software, Bar & Grill - Israel Should you use the mailblocks package and expect a response from me, you should preauthorize the dissensoftware.com domain. I very rarely bother responding to challenge/response systems, especially those from irresponsible companies.
Re: Base registers
On 6/2/2012 5:38 PM, Tony Thigpen wrote: > but there is none to be made for doing so in > writing even a new single RSECT. How about this reason. We have several customers running our software that are on pre-MP3000 machines that don't even support relative instructions. They still pay us for support and that includes software upgrades. Other vendors may not care about existing customers, but we do. Almost all of these customers also can't upgrade their VSE. They fell into the ESL trap with IBM many years ago and now can't get their budget dollars back to get current because of the IBM monthly charges. As faithful customers for many years, we do not want to kick them while they are down just so we can do 'fancier' code. There are two sides to this. You need to care about all of your customers, not just those with no money. Dragging premier customers, with the latest hardware and software, down to the level of the laggards is not fair to them. Customers that continue to invest heavily in the platform must be rewarded for doing so. IBM and ISV software should exploit the latest hardware and software features in ways that provide these leading edge customers with the very best performance and feature set possible lest they find an alternative (seemingly cheaper) way to host their applications. It's not just 'fancier' code. The productivity gains accrued by allowing programmers to leverage new facilities are immense; new features can be implemented in far less time giving customers more of they want for their maintenance dollars; tremendous run-time performance savings can be realized by leveraging new hardware and software features; today's memory rich (especially) 64-bit environments allow programs to do things only dreamed of twenty years ago; and, so on... The ESL problem is a unique problem for VSE that does not exist for MVS customers. Nevertheless, the reality is that very, Very, VERY few customers running severely back-level operating system releases are interested in installing the latest and greatest release of 'Product X' -- even on VSE. Often, such systems exist because the customer is in their 15th year of their 3-year plan to migrate off the mainframe. -- Edward E Jaffe Phoenix Software International, Inc 831 Parkview Drive North El Segundo, CA 90245 310-338-0400 x318 edja...@phoenixsoftware.com http://www.phoenixsoftware.com/
Re: Base registers
On 6/5/2012 2:19 PM, Tom Marchant wrote: On Tue, 5 Jun 2012 15:59:36 -0400, Scott Ford wrote: where can you find a good sample of baseless assembler code ? Look for Ed Jaffe's SHARE presentation "Jumpify your code". The original "jumpify your code" presentation was from 2007. When I updated it in 2011 for zEnterprise, I retitled it "jumpify your programs". I *knew* I should have kept the original title! >:o -- Edward E Jaffe Phoenix Software International, Inc 831 Parkview Drive North El Segundo, CA 90245 310-338-0400 x318 edja...@phoenixsoftware.com http://www.phoenixsoftware.com/
Re: Base registers
ISVs that have many VSE customers confront a special problem. If they judge it morally and/or economically appropriate to continue to support the ESL-trapped among these VSE customers they should do so by providing two code paths. To do this is a trivial and inexpensive, even in open code, as Tony Thigpen knows very well. I think that Edward Jaffe has this morning made the case for the use of the new instructions about as well as it can be made; and I will not therefore revisit it except to note that support for ARCH ()-driven multiple code paths, which can be packaged up into reusable macros readily (and could even be introduced into the structured-programming macro package) is the appropriate resolution of any need to support very different hardware groups concurrently; a much better one than the systematic use of LCD technology that Tony, if I have understood his views correctly, is advocating. John Gilmore, Ashland, MA 01721 - USA
Re: Base registers
John, Can't comment on VSE , we don't have any VSE customers, our customers are z/OS only. I not sure what ESL is, I assume it's a pricing plan from IBM for VSE, haven't worked VSE in yrs Scott ford www.identityforge.com On Jun 16, 2012, at 9:59 AM, John Gilmore wrote: > ISVs that have many VSE customers confront a special problem. If they > judge it morally and/or economically appropriate to continue to > support the ESL-trapped among these VSE customers they should do so by > providing two code paths. To do this is a trivial and inexpensive, > even in open code, as Tony Thigpen knows very well. > > I think that Edward Jaffe has this morning made the case for the use > of the new instructions about as well as it can be made; and I will > not therefore revisit it except to note that support for ARCH > ()-driven multiple code paths, which can be packaged up into > reusable macros readily (and could even be introduced into the > structured-programming macro package) is the appropriate resolution > of any need to support very different hardware groups concurrently; a > much better one than the systematic use of LCD technology that Tony, > if I have understood his views correctly, is advocating. > > John Gilmore, Ashland, MA 01721 - USA
Re: Base registers
'Jumpify your code' suggests to me that we should perhaps replace 'baseless', which in ordinary English means 'unfounded', with the acronym JBC, for 'Jump-based code'. John Gilmore, Ashland, MA 01721 - USA
Re: Base registers
Ed, I saw your 2007 presentation and have a copy. I am always looking for good examples to aid my education and understanding. Regards, Scott ford www.identityforge.com On Jun 15, 2012, at 8:36 PM, Edward Jaffe wrote: > On 6/5/2012 2:19 PM, Tom Marchant wrote: >> On Tue, 5 Jun 2012 15:59:36 -0400, Scott Ford wrote: >> >>> where can you find a good sample of baseless assembler code ? >> Look for Ed Jaffe's SHARE presentation "Jumpify your code". > > The original "jumpify your code" presentation was from 2007. When I updated it > in 2011 for zEnterprise, I retitled it "jumpify your programs". > > I *knew* I should have kept the original title! >:o > > -- > Edward E Jaffe > Phoenix Software International, Inc > 831 Parkview Drive North > El Segundo, CA 90245 > 310-338-0400 x318 > edja...@phoenixsoftware.com > http://www.phoenixsoftware.com/
Re: Base registers
On Jun 16, 2012, at 14:10, John Gilmore wrote: > 'Jumpify your code' suggests to me that we should perhaps replace > 'baseless', which in ordinary English means 'unfounded', with the > acronym JBC, for 'Jump-based code'. > No need. There are many instances where English accepts specialized meanings for words in special contexts which don't directly contradict each other, and of which some may even become pejorative. The first example that comes to my mind is "vulgar", as in "The vulgar meaning of 'baseless' is 'unfounded'." -- gil
Re: Base registers
Words can of course have different specialized meanings in different contexts, but there is ordinarily an evolutionary path between these meanings. Physicians, for example, talk of "senile changes", meaning those associated with aging, in a way that is entirely devoid of pejorative intent. Or again, Chaucer and his contemporaries used the word "lewd" to mean lay, not in holy orders; but there is a path between this meaning and the modern one: the clergy did not often make what we call lewd gestures in public. I myself find 'baseless' very unsatisfactory, in part because it is not at all transparent. Thus, while I have no emotional investment in the term "jump-based", I do believe a new one is needed; 'baseless' can scarcely avoid connotations of dispensability when in fact it is the base registers that are largely dispensable. We need to look forward to a time when the use of base registers, multiple ones in particular, and the arbitrary segmentation of code into 4096-byte pieces will be perceived as a quaint, historically interesting but obsolete practices; and a new contrasting term will be helpful in changing the current "vulgar" mind set. (Mr Gilmartin's use of vulgar, which evolved from the Latin phrase "mobile vulgus", is open to criticism; but that is a subject for another time and place.) Alternative suggestions? John Gilmore, Ashland, MA 01721 - USA
Re: Base registers
On 6/17/2012 9:05 AM, John Gilmore wrote: Words can of course have different specialized meanings in different contexts, but there is ordinarily an evolutionary path between these meanings. My favorite along these lines is "stench", that did not always mean unpleasant. I vividly remember a streetcar ride where almost every passenger carried freshly cut lilacs, to the point where the smell was overpowering, though not unpleasant. Alternative suggestions? I used "unbased". Gerhard Postpischil Bradford, VT
Re: Base registers
Unbased is better by a wide margin than baseless. I should still, however, prefer a non-negative form. John Gilmore, Ashland, MA 01721 - USA On 6/17/12, Gerhard Postpischil wrote: > On 6/17/2012 9:05 AM, John Gilmore wrote: >> Words can of course have different specialized meanings in different >> contexts, but there is ordinarily an evolutionary path between these >> meanings. > > My favorite along these lines is "stench", that did not always > mean unpleasant. I vividly remember a streetcar ride where > almost every passenger carried freshly cut lilacs, to the point > where the smell was overpowering, though not unpleasant. > >> Alternative suggestions? > > I used "unbased". > > Gerhard Postpischil > Bradford, VT >
Re: Base registers
Why don't we "jump" to the underlying notion of the "jump" instructions, or more accurately "branch relative" instructions, which is relative addressing: "relative address oriented programming". I'll admit that it's not concise, but I'm optimistic we won't have a religious war about the resulting acronym. -- Regards, Gord Tomlin Action Software International (a division of Mazda Computer Corporation) Tel: (905) 470-7113, Fax: (905) 470-6507 On 2012-06-17 09:05, John Gilmore wrote: Words can of course have different specialized meanings in different contexts, but there is ordinarily an evolutionary path between these meanings. Physicians, for example, talk of "senile changes", meaning those associated with aging, in a way that is entirely devoid of pejorative intent. Or again, Chaucer and his contemporaries used the word "lewd" to mean lay, not in holy orders; but there is a path between this meaning and the modern one: the clergy did not often make what we call lewd gestures in public. I myself find 'baseless' very unsatisfactory, in part because it is not at all transparent. Thus, while I have no emotional investment in the term "jump-based", I do believe a new one is needed; 'baseless' can scarcely avoid connotations of dispensability when in fact it is the base registers that are largely dispensable. We need to look forward to a time when the use of base registers, multiple ones in particular, and the arbitrary segmentation of code into 4096-byte pieces will be perceived as a quaint, historically interesting but obsolete practices; and a new contrasting term will be helpful in changing the current "vulgar" mind set. (Mr Gilmartin's use of vulgar, which evolved from the Latin phrase "mobile vulgus", is open to criticism; but that is a subject for another time and place.) Alternative suggestions? John Gilmore, Ashland, MA 01721 - USA
Re: Base registers
On 6/17/2012 3:13 PM, Gord Tomlin wrote: Why don't we "jump" to the underlying notion of the "jump" instructions, or more accurately "branch relative" instructions, which is relative addressing: "relative address oriented programming". I'll admit that it's not concise, but I'm optimistic we won't have a religious war about the resulting acronym. You mean "Religious Argument Ontological Parsing"? :-) I was going to suggest 'free base' as a 'positive' way of saying one is relatively free from using base registers, but that term also has unfortunate conotations. -- Regards, Gord Tomlin Action Software International (a division of Mazda Computer Corporation) Tel: (905) 470-7113, Fax: (905) 470-6507 On 2012-06-17 09:05, John Gilmore wrote: Words can of course have different specialized meanings in different contexts, but there is ordinarily an evolutionary path between these meanings. Physicians, for example, talk of "senile changes", meaning those associated with aging, in a way that is entirely devoid of pejorative intent. Or again, Chaucer and his contemporaries used the word "lewd" to mean lay, not in holy orders; but there is a path between this meaning and the modern one: the clergy did not often make what we call lewd gestures in public. I myself find 'baseless' very unsatisfactory, in part because it is not at all transparent. Thus, while I have no emotional investment in the term "jump-based", I do believe a new one is needed; 'baseless' can scarcely avoid connotations of dispensability when in fact it is the base registers that are largely dispensable. We need to look forward to a time when the use of base registers, multiple ones in particular, and the arbitrary segmentation of code into 4096-byte pieces will be perceived as a quaint, historically interesting but obsolete practices; and a new contrasting term will be helpful in changing the current "vulgar" mind set. (Mr Gilmartin's use of vulgar, which evolved from the Latin phrase "mobile vulgus", is open to criticism; but that is a subject for another time and place.) Alternative suggestions? John Gilmore, Ashland, MA 01721 - USA -- Kind regards, -Steve Comstock The Trainer's Friend, Inc. 303-355-2752 http://www.trainersfriend.com * To get a good Return on your Investment, first make an investment! + Training your people is an excellent investment * Try our tool for calculating your Return On Investment for training dollars at http://www.trainersfriend.com/ROI/roi.html
Re: Base registers
The problem is that all addresses are relative to something. Some are relative to a base register, some are relative to the current instruction. Even absolute addresses are relative to zero. ;-) How about base-free code? That sounds like a positive thing. (I avoided the temptation to to suggest "freebase" lest that be considered to have a negative connotation.) === > Date: Sun, 17 Jun 2012 17:13:30 -0400 > From: gt.ibm.li...@actionsoftware.com > Subject: Re: Base registers > To: ASSEMBLER-LIST@LISTSERV.UGA.EDU > > Why don't we "jump" to the underlying notion of the "jump" instructions, > or more accurately "branch relative" instructions, which is relative > addressing: "relative address oriented programming". > > I'll admit that it's not concise, but I'm optimistic we won't have a > religious war about the resulting acronym. > > -- > > Regards, Gord Tomlin > Action Software International > (a division of Mazda Computer Corporation) > Tel: (905) 470-7113, Fax: (905) 470-6507 > > > On 2012-06-17 09:05, John Gilmore wrote: > > Words can of course have different specialized meanings in different > > contexts, but there is ordinarily an evolutionary path between these > > meanings. > > > > Physicians, for example, talk of "senile changes", meaning those > > associated with aging, in a way that is entirely devoid of pejorative > > intent. Or again, Chaucer and his contemporaries used the word "lewd" > > to mean lay, not in holy orders; but there is a path between this > > meaning and the modern one: the clergy did not often make what we call > > lewd gestures in public. > > > > I myself find 'baseless' very unsatisfactory, in part because it is > > not at all transparent. Thus, while I have no emotional investment in > > the term "jump-based", I do believe a new one is needed; 'baseless' > > can scarcely avoid connotations of dispensability when in fact it is > > the base registers that are largely dispensable. > > > > We need to look forward to a time when the use of base registers, > > multiple ones in particular, and the arbitrary segmentation of code > > into 4096-byte pieces will be perceived as a quaint, historically > > interesting but obsolete practices; and a new contrasting term will be > > helpful in changing the current "vulgar" mind set. (Mr Gilmartin's > > use of vulgar, which evolved from the Latin phrase "mobile vulgus", is > > open to criticism; but that is a subject for another time and place.) > > > > Alternative suggestions? > > > > John Gilmore, Ashland, MA 01721 - USA > > > >
Re: Base registers
Sorry Steve. I didn't see your post until after I posted mine. === > Date: Sun, 17 Jun 2012 17:35:36 -0400 > From: jayare...@hotmail.com > Subject: Re: Base registers > To: ASSEMBLER-LIST@LISTSERV.UGA.EDU > > The problem is that all addresses are relative to something. > Some are relative to a base register, some are relative > to the current instruction. Even absolute addresses are > relative to zero. ;-) > > How about base-free code? That sounds like a positive thing. > > (I avoided the temptation to to suggest "freebase" lest that > be considered to have a negative connotation.) > > === > > > Date: Sun, 17 Jun 2012 17:13:30 -0400 > > From: gt.ibm.li...@actionsoftware.com > > Subject: Re: Base registers > > To: ASSEMBLER-LIST@LISTSERV.UGA.EDU > > > > Why don't we "jump" to the underlying notion of the "jump" instructions, > > or more accurately "branch relative" instructions, which is relative > > addressing: "relative address oriented programming". > > > > I'll admit that it's not concise, but I'm optimistic we won't have a > > religious war about the resulting acronym. > > > > -- > > > > Regards, Gord Tomlin > > Action Software International > > (a division of Mazda Computer Corporation) > > Tel: (905) 470-7113, Fax: (905) 470-6507 > > > > > > On 2012-06-17 09:05, John Gilmore wrote: > > > Words can of course have different specialized meanings in different > > > contexts, but there is ordinarily an evolutionary path between these > > > meanings. > > > > > > Physicians, for example, talk of "senile changes", meaning those > > > associated with aging, in a way that is entirely devoid of pejorative > > > intent. Or again, Chaucer and his contemporaries used the word "lewd" > > > to mean lay, not in holy orders; but there is a path between this > > > meaning and the modern one: the clergy did not often make what we call > > > lewd gestures in public. > > > > > > I myself find 'baseless' very unsatisfactory, in part because it is > > > not at all transparent. Thus, while I have no emotional investment in > > > the term "jump-based", I do believe a new one is needed; 'baseless' > > > can scarcely avoid connotations of dispensability when in fact it is > > > the base registers that are largely dispensable. > > > > > > We need to look forward to a time when the use of base registers, > > > multiple ones in particular, and the arbitrary segmentation of code > > > into 4096-byte pieces will be perceived as a quaint, historically > > > interesting but obsolete practices; and a new contrasting term will be > > > helpful in changing the current "vulgar" mind set. (Mr Gilmartin's > > > use of vulgar, which evolved from the Latin phrase "mobile vulgus", is > > > open to criticism; but that is a subject for another time and place.) > > > > > > Alternative suggestions? > > > > > > John Gilmore, Ashland, MA 01721 - USA > > > > > > >
Re: Base registers
Just my two cents: 1. There are architectures which have always had relative addressing, and use that term. 2. Some day, relative braches will be the norm and not the exception. As time goes on, any term with "base" as the root (e.g. unbased, baseless, etc.) is likely to need increasingly superfluous explanations to novices. Thus, I suggest "relative branch" and/or "jump" should suffice. For the exclusive case (non-base register branching) perhaps "relative-only branching" and/or "jump-only" program[ming]. On Sun, Jun 17, 2012 at 5:38 PM, J R wrote: > Sorry Steve. I didn't see your post until after I posted mine. > > === > > > Date: Sun, 17 Jun 2012 17:35:36 -0400 > > From: jayare...@hotmail.com > > Subject: Re: Base registers > > To: ASSEMBLER-LIST@LISTSERV.UGA.EDU > > > > The problem is that all addresses are relative to something. > > Some are relative to a base register, some are relative > > to the current instruction. Even absolute addresses are > > relative to zero. ;-) > > > > How about base-free code? That sounds like a positive thing. > > > > (I avoided the temptation to to suggest "freebase" lest that > > be considered to have a negative connotation.) > > > > === > > > > > Date: Sun, 17 Jun 2012 17:13:30 -0400 > > > From: gt.ibm.li...@actionsoftware.com > > > Subject: Re: Base registers > > > To: ASSEMBLER-LIST@LISTSERV.UGA.EDU > > > > > > Why don't we "jump" to the underlying notion of the "jump" > instructions, > > > or more accurately "branch relative" instructions, which is relative > > > addressing: "relative address oriented programming". > > > > > > I'll admit that it's not concise, but I'm optimistic we won't have a > > > religious war about the resulting acronym. > > > > > > -- > > > > > > Regards, Gord Tomlin > > > Action Software International > > > (a division of Mazda Computer Corporation) > > > Tel: (905) 470-7113, Fax: (905) 470-6507 > > > > > > > > > On 2012-06-17 09:05, John Gilmore wrote: > > > > Words can of course have different specialized meanings in different > > > > contexts, but there is ordinarily an evolutionary path between these > > > > meanings. > > > > > > > > Physicians, for example, talk of "senile changes", meaning those > > > > associated with aging, in a way that is entirely devoid of pejorative > > > > intent. Or again, Chaucer and his contemporaries used the word > "lewd" > > > > to mean lay, not in holy orders; but there is a path between this > > > > meaning and the modern one: the clergy did not often make what we > call > > > > lewd gestures in public. > > > > > > > > I myself find 'baseless' very unsatisfactory, in part because it is > > > > not at all transparent. Thus, while I have no emotional investment > in > > > > the term "jump-based", I do believe a new one is needed; 'baseless' > > > > can scarcely avoid connotations of dispensability when in fact it is > > > > the base registers that are largely dispensable. > > > > > > > > We need to look forward to a time when the use of base registers, > > > > multiple ones in particular, and the arbitrary segmentation of code > > > > into 4096-byte pieces will be perceived as a quaint, historically > > > > interesting but obsolete practices; and a new contrasting term will > be > > > > helpful in changing the current "vulgar" mind set. (Mr Gilmartin's > > > > use of vulgar, which evolved from the Latin phrase "mobile vulgus", > is > > > > open to criticism; but that is a subject for another time and place.) > > > > > > > > Alternative suggestions? > > > > > > > > John Gilmore, Ashland, MA 01721 - USA > > > > > > > > > > > > -- OREXXMan
Re: Base registers
From: "Watkins, Douglas" Sent: Wednesday, 6 June 2012 11:28 PM Here's one way to do standard EXecute without a base register: AHI R2,-1 Minus 1 for EX BTW, BCTR 2,0 will do a better job.
Re: Base registers
On Mon, 2012-06-18 at 11:24 +1000, Robin Vowels wrote: > From: "Watkins, Douglas" > Sent: Wednesday, 6 June 2012 11:28 PM > > > > Here's one way to do standard EXecute without a base register: > > > > AHI R2,-1 Minus 1 for EX > > BTW, > BCTR 2,0 will do a better job. I agree. It is two bytes shorter. Possibly faster. Does not disturb the condition code. And is a generally understood standard notation. Might even be considered like the DECrement instruction in other assemblers. As an aside, I was actually (foolishly) considering making macros called DEC (as in DEC Reg,Value => AHI Reg,0-&Value with default of 1), INC (INC Reg,Value => AHI Reg,Value, default of 1), PRED Reg (DEC Reg,1), SUCC Reg (INC Reg). Then realized it would likely be more confusing than helpful. -- John McKown Maranatha! <><
Re: Base registers
On 6/17/2012 10:44 PM, John McKown wrote: As an aside, I was actually (foolishly) considering making macros called DEC (as in DEC Reg,Value => AHI Reg,0-&Value with default of 1), INC (INC Reg,Value => AHI Reg,Value, default of 1), PRED Reg (DEC Reg,1), SUCC Reg (INC Reg). Then realized it would likely be more confusing than helpful. Not sure what's confusing about it; I've had an INC for ages (no DEC, instead put INC=-1 on INC - that may be confusing?). It changes either a register or storage, and had a version for halfwords named INCH Gerhard Postpischil Bradford, VT
Re: Base registers
On Mon, Jun 18, 2012 at 6:52 AM, Gerhard Postpischil wrote: > Not sure what's confusing about it; I've had an INC for ages (no > DEC, instead put INC=-1 on INC - that may be confusing?). It > changes either a register or storage, and had a version for > halfwords named INCH Yep! It's a ver useful way to abstract from the actual implementation. It helps distinguish the BCTR x,0 from the ones that implement a loop. Since I have *no* branch instructions or labels in my new code other than generated by structured programming macros, going to basefree code will only take changes to the macros. A BCTR would stand out more than I want. Rob
Re: Base registers
On Mon, 18 Jun 2012 11:24:14 +1000 Robin Vowels wrote: :>From: "Watkins, Douglas" :>Sent: Wednesday, 6 June 2012 11:28 PM :>> Here's one way to do standard EXecute without a base register: > :>> AHI R2,-1 Minus 1 for EX :>BTW, :>BCTR 2,0 will do a better job. Unless you place a JM after the AHI -- Binyamin Dissen http://www.dissensoftware.com Director, Dissen Software, Bar & Grill - Israel Should you use the mailblocks package and expect a response from me, you should preauthorize the dissensoftware.com domain. I very rarely bother responding to challenge/response systems, especially those from irresponsible companies.
Re: Base registers
> All your bases are belong to us ? For great justice. Fred! - ATTENTION: The information in this electronic mail message is private and confidential, and only intended for the addressee. Should you receive this message by mistake, you are hereby notified that any disclosure, reproduction, distribution or use of this message is strictly prohibited. Please inform the sender by reply transmission and delete the message without copying or opening it. Messages and attachments are scanned for all viruses known. If this message contains password-protected attachments, the files have NOT been scanned for viruses by the ING mail domain. Always scan attachments before opening them. -
Re: Base registers
On Mon, Jun 18, 2012 at 10:36 AM, Thomas Berg wrote: > All your bases are belong to us ? LOL. Just when I was convinced we had cross-talk with the IBM-MAIN list ;-) May the source be with you...
Re: Base registers
On 6/18/2012 4:05 AM, Fred van der Windt wrote: All your bases are belong to us ? For great justice. Fred! Says the Dutchman replying to the Swede! I love this list! -- Kind regards, -Steve Comstock The Trainer's Friend, Inc. 303-355-2752 http://www.trainersfriend.com * To get a good Return on your Investment, first make an investment! + Training your people is an excellent investment * Try our tool for calculating your Return On Investment for training dollars at http://www.trainersfriend.com/ROI/roi.html
Re: Base registers
Steve Comstock wrote: >I was going to suggest 'free base' as a 'positive' way >of saying one is relatively free from using base registers, >but that term also has unfortunate conotations. Indeed..."debased" would be my choice! :-) ...phsiii
Re: Base registers
> -Original Message- > From: IBM Mainframe Assembler List > [mailto:ASSEMBLER-LIST@LISTSERV.UGA.EDU] On Behalf Of Rob van der Heij > Sent: Monday, June 18, 2012 6:23 AM > To: ASSEMBLER-LIST@LISTSERV.UGA.EDU > Subject: Re: Base registers > > On Mon, Jun 18, 2012 at 10:36 AM, Thomas Berg > wrote: > > > All your bases are belong to us ? > > LOL. Just when I was convinced we had cross-talk with the > IBM-MAIN list ;-) > > May the source be with you... Which is guaranteed if you use only GPL'd software! "Just say NO to OCO!" still lives. Confirmed Penguinista, John McKown Systems Engineer IV IT Administrative Services Group HealthMarkets(r) 9151 Boulevard 26 * N. Richland Hills * TX 76010 (817) 255-3225 phone * john.mck...@healthmarkets.com * www.HealthMarkets.com Confidentiality Notice: This e-mail message may contain confidential or proprietary information. If you are not the intended recipient, please contact the sender by reply e-mail and destroy all copies of the original message. HealthMarkets(r) is the brand name for products underwritten and issued by the insurance subsidiaries of HealthMarkets, Inc. -The Chesapeake Life Insurance Company(r), Mid-West National Life Insurance Company of TennesseeSM and The MEGA Life and Health Insurance Company.SM
Re: Base registers
On Sun, 17 Jun 2012 17:01:05 -0400, John Gilmore wrote: >Unbased is better by a wide margin than baseless. > >I should still, however, prefer a non-negative form. I don't have an entirely satisfactory suggestion, but I sometimes refer to it as "Relative code". IMO, Unbased suffers from the same problem as Baseless, that it is inaccurate. Base registers are still needed for data. Relative addressing removes the requirement for most code. -- Tom Marchant
Re: Base registers
My quibble with the terms being bandied about is that neither "unbased" nor "baseless" is factually correct for a large percentage of modules that use relative branches. They have a "base" to their static data (and, yes, sometimes that "base" is not persistent and is created only when needed). They just do not tend to need a base for their code (aside from when they use macros that have not been "jumpified"). If you think of the addressablity to code as being via "codereg" they have been de-codereg'd (not a pleasing term, of course). I think everyone these days understands what someone means when they use the non-word "jumpify". Is more than that needed? Creating more terminology just for the sake of doing so seems unwieldy (and perhaps it could be said "IBM-like"). Peter Relson z/OS Core Technology Design
Re: Base registers
OK, I can live with "jumpify". I would just like something that is "generally acceptable" to the community. Only minus is that it doesn't say anything about the use of Relative and Immediate instructions. Of which I am also a big fan. -- John McKown Systems Engineer IV IT Administrative Services Group HealthMarkets(r) 9151 Boulevard 26 * N. Richland Hills * TX 76010 (817) 255-3225 phone * john.mck...@healthmarkets.com * www.HealthMarkets.com Confidentiality Notice: This e-mail message may contain confidential or proprietary information. If you are not the intended recipient, please contact the sender by reply e-mail and destroy all copies of the original message. HealthMarkets(r) is the brand name for products underwritten and issued by the insurance subsidiaries of HealthMarkets, Inc. -The Chesapeake Life Insurance Company(r), Mid-West National Life Insurance Company of TennesseeSM and The MEGA Life and Health Insurance Company.SM
Re: Base registers
To make it IBM-like, you'd have to call it the "Jumpify Facility". :-) -- From: "Peter Relson" Sent: Tuesday, June 19, 2012 9:21 AM To: Subject: Re: Base registers My quibble with the terms being bandied about is that neither "unbased" nor "baseless" is factually correct for a large percentage of modules that use relative branches. They have a "base" to their static data (and, yes, sometimes that "base" is not persistent and is created only when needed). They just do not tend to need a base for their code (aside from when they use macros that have not been "jumpified"). If you think of the addressablity to code as being via "codereg" they have been de-codereg'd (not a pleasing term, of course). I think everyone these days understands what someone means when they use the non-word "jumpify". Is more than that needed? Creating more terminology just for the sake of doing so seems unwieldy (and perhaps it could be said "IBM-like"). Peter Relson z/OS Core Technology Design
Re: Base registers
"Jumpify" is acceptable as a nonce word, in, say, EJ's title. It is nevertheless barbarous English and should not be adopted for routine use. "JRI"? --jg On 6/19/12, McKown, John wrote: > OK, I can live with "jumpify". I would just like something that is > "generally acceptable" to the community. Only minus is that it doesn't say > anything about the use of Relative and Immediate instructions. Of which I am > also a big fan. > > -- > John McKown > Systems Engineer IV > IT > > Administrative Services Group > > HealthMarkets(r) > > 9151 Boulevard 26 * N. Richland Hills * TX 76010 > (817) 255-3225 phone * > john.mck...@healthmarkets.com * www.HealthMarkets.com > > Confidentiality Notice: This e-mail message may contain confidential or > proprietary information. If you are not the intended recipient, please > contact the sender by reply e-mail and destroy all copies of the original > message. HealthMarkets(r) is the brand name for products underwritten and > issued by the insurance subsidiaries of HealthMarkets, Inc. -The Chesapeake > Life Insurance Company(r), Mid-West National Life Insurance Company of > TennesseeSM and The MEGA Life and Health Insurance Company.SM >
