From sacadmin Thu Jun 5 19:46:24 2003 Date: Thu, 5 Jun 2003 19:37:35 -0700 (PDT) From: Paul Lodrige Subject: SFastTrack: PSARC 2003/335 EOL of le Ethernet driver [timeout 06/12/2003] To: psarc@sac.eng.sun.com Cc: nspg-interest@sun.com MIME-Version: 1.0 Content-Type: TEXT/plain; charset=us-ascii Content-MD5: G1/9ojGSEoOR1eWSnRkM7g== Content-Length: 2360 Subject: FastTrack: PSARC 2003/335 EOL of le Ethernet driver [timeout 06/12/2003] Template Version: @(#)sac_nextcase 1.47 12/20/02 SMI 1. Introduction 1.1. Project/Component Working Name: EOL of le Ethernet driver PSARC FastTrack, proposed timeout 06/12/2003 Case Number: PSARC/2003/335 1.2. Name of Document Author/Supplier: Author: paul lodrige Sponsor: pfl 1.3 Date of This Document: 05 June, 2003 4. Technical Description See the case directory for more detail URL: http://sac.eng/arc/PSARC/2003/335 6. Resources and Schedule 6.4. Steering Committee requested information 6.4.1. Consolidation C-team Name: ON 6.5. ARC review type: FastTrack timeout 06/12/2003 ******************************************************************************* I'm self-sponsoring this fast-track to remove the le Ethernet driver from the S10 FCS. Timeout is set for 06/12/2003 pfl 1. Introduction 1.1. Project/Component Working Name: EOL of le Ethernet driver 1.2. Name of Document Author/Supplier: Paul Lodrige 1.3. Date of This Document: 06/5/03 1.4. Name of Major Document Customer(s)/Consumer(s): PSARC, SOESC, NSPG 2. Project Summary 2.1. Project Description This project will remove le Ethernet driver support from Solaris 10. 2.2. Risks and Assumptions 3. Business Summary 3.1. Problem Area 3.2. Market/Requester 3.3. Business Justification There are no current systems shipping with the le hardware. The only UltraSparc system which did have le hardware was the early Ultra 1 systems which have already been EOL'ed. The plan is to EOL the le driver in Solaris 10 3.4. Competitive Analysis 3.5. Opportunity Window/Exposure Solaris 10 3.6. How will you know when you are done? Integration into the ON gate 4. Technical Description The packages, source files and source lines exclusively for le are removed from Solaris 10. 5. Reference Documents Document: In End-of-Support Products section of Solaris 9 4/03 Release Notes Supplement for Sun Hardware includes the le driver as Future End-of-Support Products. From sacadmin Fri Jun 6 03:12:50 2003 Date: Fri, 6 Jun 2003 11:05:06 +0100 (BST) From: Andrew Gabriel Subject: Re: SFastTrack: PSARC 2003/335 EOL of le Ethernet driver [timeout 06/12/2003] To: psarc@sac.eng.sun.com, Paul.Lodrige@eng.sun.com Cc: nspg-interest@sun.com MIME-Version: 1.0 Content-Type: TEXT/plain; charset=us-ascii Content-MD5: 4TwywdGoNZxFbbOT26G6cQ== Content-Length: 535 >Subject: FastTrack: PSARC 2003/335 EOL of le Ethernet driver It would be prudent to ensure this was not putback to ON10 prior to Ultrasparc I feature removal (PSARC/2003/014 EOL of 32-bit SPARC Kernels), since the Ultrasparc I feature removal has been deliberately delayed until later in the ON10 development cycle to allow the development labs around Sun time to replace their Ultrasparc I systems. I believe this is currently intended to be September 2003, but whenever it is, this case should not precede it. -- Andrew Gabriel From sacadmin Fri Jun 6 07:58:59 2003 From: Bill Sommerfeld To: Andrew Gabriel cc: psarc@sac.eng.sun.com, Paul.Lodrige@eng.sun.com, nspg-interest@sun.com Subject: Re: SFastTrack: PSARC 2003/335 EOL of le Ethernet driver [timeout 06/12/2003] Date: Fri, 06 Jun 2003 10:56:47 -0400 Content-Length: 735 > It would be prudent to ensure this was not putback to ON10 prior to > Ultrasparc I feature removal (PSARC/2003/014 EOL of 32-bit SPARC Kernels), > since the Ultrasparc I feature removal has been deliberately delayed > until later in the ON10 development cycle to allow the development labs > around Sun time to replace their Ultrasparc I systems. I believe this > is currently intended to be September 2003, but whenever it is, this > case should not precede it. It would also be good to get a sense for the number of stand-alone sbus le cards which this will orphan. I'm willing to believe "small" since the remaining sbus systems are midrange and up (Ex{0,5}00, E10k) which are unlikely to be using that card.. - Bill From sacadmin Fri Jun 6 08:29:30 2003 Date: Fri, 6 Jun 2003 16:27:18 +0100 (BST) From: Andrew Gabriel - Internet Engineering Subject: Re: SFastTrack: PSARC 2003/335 EOL of le Ethernet driver [timeout 06/12/2003] To: psarc@sac.eng.sun.com, Paul.Lodrige@eng.sun.com, nspg-interest@Sun.COM MIME-Version: 1.0 Content-Type: TEXT/plain; charset=us-ascii Content-MD5: BzD8JTKsblCywXaM+bNrXA== Content-Length: 608 > It would also be good to get a sense for the number of stand-alone > sbus le cards which this will orphan. I'm willing to believe "small" > since the remaining sbus systems are midrange and up (Ex{0,5}00, E10k) > which are unlikely to be using that card.. A colleague sitting next to me has just pointed out that we still support the SCSI/Ethernet (SBE/S) and Differential SCSI/Ethernet (DSBE/S) cards in Ultra 2 systems, and these both use the LANCE chipset and hence the le driver. He also said the LANCE sbus cards were never ever officially supported in the Enterprise servers. -- Andrew Gabriel From sacadmin Fri Jun 6 08:40:38 2003 Date: Fri, 6 Jun 2003 08:38:28 -0700 (PDT) From: "Joseph E. Kowalski III" Subject: Re: SFastTrack: PSARC 2003/335 EOL of le Ethernet driver [timeout 06/12/2003] To: psarc@sac.eng.sun.com, Paul.Lodrige@eng.sun.com, Andrew.Gabriel@sun.com Cc: nspg-interest@sun.com MIME-Version: 1.0 Content-Type: TEXT/plain; charset=us-ascii Content-MD5: YWgZsUUzug8o+PlSUXZ9qQ== Content-Length: 908 > From: Andrew Gabriel ... > >Subject: FastTrack: PSARC 2003/335 EOL of le Ethernet driver > > It would be prudent to ensure this was not putback to ON10 prior to > Ultrasparc I feature removal (PSARC/2003/014 EOL of 32-bit SPARC Kernels), > since the Ultrasparc I feature removal has been deliberately delayed > until later in the ON10 development cycle to allow the development labs > around Sun time to replace their Ultrasparc I systems. I believe this > is currently intended to be September 2003, but whenever it is, this > case should not precede it. > > -- > Andrew Gabriel Good observation. Formally, we do this by defining case dependencies. Hence,... This case (PSARC/2003/335) depends upon "PSARC/2003/014 EOL of 32-bit SPARC Kernels" and may not integrate before it. I will assume this is now part of the specification unless Paul asserts otherwise. - jek3 From sacadmin Fri Jun 6 08:52:02 2003 From: Bill Sommerfeld To: "Joseph E. Kowalski III" cc: psarc@sac.eng.sun.com, Paul.Lodrige@eng.sun.com, Andrew.Gabriel@sun.com, nspg-interest@sun.com Subject: Re: SFastTrack: PSARC 2003/335 EOL of le Ethernet driver [timeout 06/12/2003] Date: Fri, 06 Jun 2003 11:49:50 -0400 Content-Length: 526 > Formally, we do this by defining case dependencies. Hence,... > > This case (PSARC/2003/335) depends upon "PSARC/2003/014 EOL of > 32-bit SPARC Kernels" and may not integrate before it. > > I will assume this is now part of the specification unless Paul asserts > otherwise. I'll point out that half of 2003/014 (removal of 32-bit SPARC kernels) has been integrated into on10, but the other half (removal of support for running on UltraSPARC I systems) hasn't for the reasons that Andrew pointed out. - Bill From sacadmin Fri Jun 6 09:14:32 2003 Date: Fri, 6 Jun 2003 09:12:25 -0700 (PDT) From: "Joseph E. Kowalski III" Subject: Re: SFastTrack: PSARC 2003/335 EOL of le Ethernet driver [timeout 06/12/2003] To: Joseph.Kowalski@eng.sun.com, sommerfeld@east.sun.com Cc: psarc@sac.eng.sun.com, Paul.Lodrige@eng.sun.com, Andrew.Gabriel@sun.com, nspg-interest@sun.com MIME-Version: 1.0 Content-Type: TEXT/plain; charset=us-ascii Content-MD5: xTTOd8mFvVFGKhSVQ9CZBw== Content-Length: 548 > I'll point out that half of 2003/014 (removal of 32-bit SPARC kernels) > has been integrated into on10, but the other half (removal of support > for running on UltraSPARC I systems) hasn't for the reasons that > Andrew pointed out. > > - Bill Sigh,... Officially "partial" projects aren't supposed to integrate. The right thing happened here (IMHO, aside from the paperwork) so there is no point in making an issue over this. Just word the dependency appropriately... "until the EOL of UltraSPARC I systems as defined by" ... - jek3 From sacadmin Fri Jun 6 09:22:16 2003 Date: Fri, 6 Jun 2003 10:18:58 -0600 (MDT) From: Brian Botton Subject: Re: SFastTrack: PSARC 2003/335 EOL of le Ethernet driver [timeout 06/12/2003] To: psarc@sac.eng.sun.com, Paul.Lodrige@eng.sun.com MIME-Version: 1.0 Content-Type: TEXT/plain; charset=us-ascii Content-MD5: uolB2/CKRsDqwyWJMpjcTw== Content-Length: 1348 > > It would also be good to get a sense for the number of stand-alone > > sbus le cards which this will orphan. I'm willing to believe "small" > > since the remaining sbus systems are midrange and up (Ex{0,5}00, E10k) > > which are unlikely to be using that card.. > > A colleague sitting next to me has just pointed out that we > still support the SCSI/Ethernet (SBE/S) and Differential > SCSI/Ethernet (DSBE/S) cards in Ultra 2 systems, and these > both use the LANCE chipset and hence the le driver. > > He also said the LANCE sbus cards were never ever officially > supported in the Enterprise servers. I'm facing the same thing with the dbri audio driver. There is an s-bus card that uses this driver, the "bat phone" card. I'm just going to sit back and wait for the s-bus to be EOLed and then remove the driver. While this particular card is supported on the U-2, I doubt it's officially supported on the enterprise servers. However if we've got one customer that is doing so then why make him/her angry by removing the driver? I'm sure as mature as the le driver is that it isn't a maintenance drain. Thanks, Brian ... ___ _________ _][_n_n___i_i _________ | | Brian Botton (____________I.I_______I.|_______| x49533/(719) 265-6108 /ooOOOO OOOOoo oo oooo oo oo Brian.Botton@Sun.COM From sacadmin Fri Jun 6 09:30:19 2003 Date: Fri, 6 Jun 2003 09:27:02 -0700 (PDT) From: Shudong Zhou Subject: Re: SFastTrack: PSARC 2003/335 EOL of le Ethernet driver [timeout 06/12/2003] To: Andrew.Gabriel@sun.com, sommerfeld@east.sun.com Cc: psarc@sac.eng.sun.com, Paul.Lodrige@eng.sun.com, nspg-interest@sun.com MIME-Version: 1.0 Content-Type: TEXT/plain; charset=us-ascii Content-MD5: F9Hyi9OatQsj1XLdnmdA8w== Content-Length: 533 > > It would also be good to get a sense for the number of stand-alone > sbus le cards which this will orphan. I'm willing to believe "small" > since the remaining sbus systems are midrange and up (Ex{0,5}00, E10k) > which are unlikely to be using that card.. > > - Bill We have a number of le cards in our labs, plugged in to old UE3000-6000 machines. They are mostly for testing, not as primary interfaces. I have no trouble with getting rid of them personally. As network interfaces, they are just too slow. Shudong From sacadmin Fri Jun 6 13:14:18 2003 Date: Fri, 6 Jun 2003 13:05:28 -0700 (PDT) From: Paul Lodrige Subject: Re: SFastTrack: PSARC 2003/335 EOL of le Ethernet driver [timeout 06/12/2003] To: psarc@sac.eng.sun.com, Andrew.Gabriel@sun.com, Joseph.Kowalski@eng.sun.com Cc: nspg-interest@sun.com MIME-Version: 1.0 Content-Type: TEXT/plain; charset=us-ascii Content-MD5: FwBlN2xXFUYUbR1cKx2LVA== Content-Length: 334 >Formally, we do this by defining case dependencies. Hence,... > > This case (PSARC/2003/335) depends upon "PSARC/2003/014 EOL of > 32-bit SPARC Kernels" and may not integrate before it. I will add this to the case . pfl > >I will assume this is now part of the specification unless Paul asserts >otherwise. > >- jek3 > From sacadmin Fri Jun 6 13:55:25 2003 Date: Fri, 6 Jun 2003 13:46:34 -0700 (PDT) From: Paul Lodrige Subject: Re: SFastTrack: PSARC 2003/335 EOL of le Ethernet driver [timeout 06/12/2003] To: Andrew.Gabriel@sun.com, sommerfeld@east.sun.com Cc: psarc@sac.eng.sun.com, nspg-interest@sun.com MIME-Version: 1.0 Content-Type: TEXT/plain; charset=us-ascii Content-MD5: 5tUhpByaNvJrUQzaWBYhFg== Content-Length: 1086 >From: Bill Sommerfeld >To: Andrew Gabriel >cc: psarc@sac.eng.sun.com, Paul.Lodrige@eng.sun.com, nspg-interest@sun.com >Subject: Re: SFastTrack: PSARC 2003/335 EOL of le Ethernet driver [timeout 06/12/2003] >Date: Fri, 06 Jun 2003 10:56:47 -0400 > >> It would be prudent to ensure this was not putback to ON10 prior to >> Ultrasparc I feature removal (PSARC/2003/014 EOL of 32-bit SPARC Kernels), >> since the Ultrasparc I feature removal has been deliberately delayed >> until later in the ON10 development cycle to allow the development labs >> around Sun time to replace their Ultrasparc I systems. I believe this >> is currently intended to be September 2003, but whenever it is, this >> case should not precede it. yes - this will be added to the case as a dependency. > >It would also be good to get a sense for the number of stand-alone >sbus le cards which this will orphan. I'm willing to believe "small" >since the remaining sbus systems are midrange and up (Ex{0,5}00, E10k) >which are unlikely to be using that card.. From my group's collective knowledge, there are no "standalone" le cards but this goes back quite a ways so even the old timers aren't completely sure. At least the websites and docs don't exist to indicate we shipped any. We did ship some combo cards with le: a scsi/ethernet sbus card (SBE/S) ( listed on onestop as model 1054A) and a Differential SCSI/Ethernet (DSBE/S) ( listed on onestop as model 1052A). Both are EOL'ed, in 1998 and in 2000 respectively. So hopefully this is not an issue. pfl > > - Bill > > > From sacadmin Fri Jun 6 13:56:38 2003 Date: Fri, 6 Jun 2003 13:47:48 -0700 (PDT) From: Paul Lodrige Subject: Re: SFastTrack: PSARC 2003/335 EOL of le Ethernet driver [timeout 06/12/2003] To: psarc@sac.eng.sun.com, nspg-interest@Sun.COM, Andrew.Gabriel@Sun.COM MIME-Version: 1.0 Content-Type: TEXT/plain; charset=us-ascii Content-MD5: 7aXO3BfJDBwoDpxMCGl4Kg== Content-Length: 1011 >Date: Fri, 6 Jun 2003 16:27:18 +0100 (BST) >From: Andrew Gabriel - Internet Engineering >Subject: Re: SFastTrack: PSARC 2003/335 EOL of le Ethernet driver [timeout 06/12/2003] >To: psarc@sac.eng.sun.com, Paul.Lodrige@eng.sun.com, nspg-interest@Sun.COM >MIME-Version: 1.0 >Content-MD5: BzD8JTKsblCywXaM+bNrXA== > >> It would also be good to get a sense for the number of stand-alone >> sbus le cards which this will orphan. I'm willing to believe "small" >> since the remaining sbus systems are midrange and up (Ex{0,5}00, E10k) >> which are unlikely to be using that card.. > >A colleague sitting next to me has just pointed out that we >still support the SCSI/Ethernet (SBE/S) and Differential >SCSI/Ethernet (DSBE/S) cards in Ultra 2 systems, and these >both use the LANCE chipset and hence the le driver. Both are EOL'ed , in 1998 and in 2000. pfl > >He also said the LANCE sbus cards were never ever officially >supported in the Enterprise servers. > >-- >Andrew Gabriel > From sacadmin Fri Jun 6 14:11:10 2003 Date: Fri, 6 Jun 2003 15:07:52 -0600 (MDT) From: Brian Botton Subject: Re: SFastTrack: PSARC 2003/335 EOL of le Ethernet driver [timeout 06/12/2003] To: psarc@sac.eng.sun.com, nspg-interest@sun.com, Andrew.Gabriel@sun.com, Paul.Lodrige@eng.sun.com MIME-Version: 1.0 Content-Type: TEXT/plain; charset=us-ascii Content-MD5: KEieDUryKfwySdeQX33wTg== Content-Length: 671 > >A colleague sitting next to me has just pointed out that we > >still support the SCSI/Ethernet (SBE/S) and Differential > >SCSI/Ethernet (DSBE/S) cards in Ultra 2 systems, and these > >both use the LANCE chipset and hence the le driver. > > Both are EOL'ed , in 1998 and in 2000. Doesn't that mean we need to support the one card for 5 years? This means it has to be there for S10 since it ships the end of 2004, before the 5 year timeout. Brian ... ___ _________ _][_n_n___i_i _________ | | Brian Botton (____________I.I_______I.|_______| x49533/(719) 265-6108 /ooOOOO OOOOoo oo oooo oo oo Brian.Botton@Sun.COM From sacadmin Fri Jun 6 14:37:45 2003 Date: Fri, 6 Jun 2003 14:28:55 -0700 (PDT) From: Paul Lodrige Subject: Re: SFastTrack: PSARC 2003/335 EOL of le Ethernet driver [timeout 06/12/2003] To: psarc@sac.eng.sun.com, nspg-interest@sun.com, Andrew.Gabriel@sun.com, Brian.Botton@sun.com MIME-Version: 1.0 Content-Type: TEXT/plain; charset=us-ascii Content-MD5: IJXO4DPtAzocTkoEZn36Qg== Content-Length: 1162 >Date: Fri, 6 Jun 2003 15:07:52 -0600 (MDT) >From: Brian Botton >Subject: Re: SFastTrack: PSARC 2003/335 EOL of le Ethernet driver [timeout 06/12/2003] >To: psarc@sac.eng.sun.com, nspg-interest@sun.com, Andrew.Gabriel@sun.com, Paul.Lodrige@eng.sun.com >MIME-Version: 1.0 >Content-MD5: KEieDUryKfwySdeQX33wTg== > > >> >A colleague sitting next to me has just pointed out that we >> >still support the SCSI/Ethernet (SBE/S) and Differential >> >SCSI/Ethernet (DSBE/S) cards in Ultra 2 systems, and these >> >both use the LANCE chipset and hence the le driver. >> >> Both are EOL'ed , in 1998 and in 2000. > > Doesn't that mean we need to support the one card for 5 years? >This means it has to be there for S10 since it ships the end of >2004, before the 5 year timeout. Looks like we stuck with le for another release - it's a shame no one here knew about these cards. I'll withdraw the case. pfl > >Brian > > ... ___ _________ > _][_n_n___i_i _________ | | Brian Botton > (____________I.I_______I.|_______| x49533/(719) 265-6108 > /ooOOOO OOOOoo oo oooo oo oo Brian.Botton@Sun.COM > > From sacadmin Fri Jun 6 15:01:12 2003 Date: Fri, 06 Jun 2003 14:59:03 -0700 (PDT) From: Hui Chen Subject: Re: SFastTrack: PSARC 2003/335 EOL of le Ethernet driver [timeout 06/12/2003] To: psarc@sac.eng.sun.com, nspg-interest@sun.com, Andrew.Gabriel@Sun.COM, Brian.Botton@Sun.COM, Paul.Lodrige@eng.sun.com MIME-version: 1.0 Content-type: TEXT/plain; charset=us-ascii Content-MD5: 2Skp6eLJvedyiC4raLklCw== Content-Length: 1927 Paul, Usually 6 months after the EOL announce of the cards is the last shipment date, no more shipment of the product after that. After 3 years of EOL, the support of the product becomes Time-and-Materials, meaning customer need to pay for the bug to be fixed. So it's Ok to remove le from S10. Please don't withdraw the case. Thanks, Hui > Date: Fri, 06 Jun 2003 14:28:55 -0700 (PDT) > From: Paul Lodrige > Subject: Re: SFastTrack: PSARC 2003/335 EOL of le Ethernet driver [timeout 06/12/2003] > To: psarc@sac.eng.sun.com, nspg-interest@sun.com, Andrew.Gabriel@Sun.COM, Brian.Botton@Sun.COM > MIME-version: 1.0 > Content-MD5: IJXO4DPtAzocTkoEZn36Qg== > > > >Date: Fri, 6 Jun 2003 15:07:52 -0600 (MDT) > >From: Brian Botton > >Subject: Re: SFastTrack: PSARC 2003/335 EOL of le Ethernet driver [timeout > 06/12/2003] > >To: psarc@sac.eng.sun.com, nspg-interest@sun.com, Andrew.Gabriel@sun.com, > Paul.Lodrige@eng.sun.com > >MIME-Version: 1.0 > >Content-MD5: KEieDUryKfwySdeQX33wTg== > > > > > >> >A colleague sitting next to me has just pointed out that we > >> >still support the SCSI/Ethernet (SBE/S) and Differential > >> >SCSI/Ethernet (DSBE/S) cards in Ultra 2 systems, and these > >> >both use the LANCE chipset and hence the le driver. > >> > >> Both are EOL'ed , in 1998 and in 2000. > > > > Doesn't that mean we need to support the one card for 5 years? > >This means it has to be there for S10 since it ships the end of > >2004, before the 5 year timeout. > > Looks like we stuck with le for another release - it's a shame no one here knew > about these cards. > > I'll withdraw the case. > > pfl > > > >Brian > > > > ... ___ _________ > > _][_n_n___i_i _________ | | Brian Botton > > (____________I.I_______I.|_______| x49533/(719) 265-6108 > > /ooOOOO OOOOoo oo oooo oo oo Brian.Botton@Sun.COM > > > > > From sacadmin Fri Jun 6 15:39:23 2003 Date: Fri, 6 Jun 2003 16:36:04 -0600 (MDT) From: Brian Botton Subject: Re: SFastTrack: PSARC 2003/335 EOL of le Ethernet driver [timeout 06/12/2003] To: psarc@sac.eng.sun.com, nspg-interest@sun.com, Andrew.Gabriel@sun.com, Paul.Lodrige@eng.sun.com, hui.chen@sun.com MIME-Version: 1.0 Content-Type: TEXT/plain; charset=us-ascii Content-MD5: dnpJ72YmVS5CkzdsAlyK9w== Content-Length: 1279 > Usually 6 months after the EOL announce of the cards is the last shipment > date, no more shipment of the product after that. After 3 years of EOL, > the support of the product becomes Time-and-Materials, meaning customer > need to pay for the bug to be fixed. So it's Ok to remove le from S10. > Please don't withdraw the case. There's a difference between getting a bug fixed and removing the driver so the device no longer works. However... I've done some research on the EOL rules. Full support has to be provided for one year after the EOL announcement. Once this happens then the EOLed component can be removed. Of course this removal takes an arc case. As I interpret the rules, the 1 year has come and gone WRT the sbus cards. Thus it's okay to proceed with this case. The end result is that it has to be supported on all systems shipped within that 1 year time frame. Since S10 will ship after that 1 year it can be removed from S10. So I stand corrected. Does anyone else have a different interpretation of the EOL rules? Thanks, Brian ... ___ _________ _][_n_n___i_i _________ | | Brian Botton (____________I.I_______I.|_______| x49533/(719) 265-6108 /ooOOOO OOOOoo oo oooo oo oo Brian.Botton@Sun.COM From sacadmin Sat Jun 7 04:12:59 2003 To: Brian Botton cc: psarc@sac.eng.sun.com, nspg-interest@sun.com, Andrew.Gabriel@sun.com, Paul.Lodrige@eng.sun.com, hui.chen@sun.com Subject: Re: SFastTrack: PSARC 2003/335 EOL of le Ethernet driver [timeout 06/12/2003] Date: Sat, 07 Jun 2003 13:10:47 +0200 From: Casper Dik Content-Length: 985 > >> Usually 6 months after the EOL announce of the cards is the last shipment >> date, no more shipment of the product after that. After 3 years of EOL, >> the support of the product becomes Time-and-Materials, meaning customer >> need to pay for the bug to be fixed. So it's Ok to remove le from S10. >> Please don't withdraw the case. > > There's a difference between getting a bug fixed and removing the >driver so the device no longer works. However... > > I've done some research on the EOL rules. Full support has to be provided >for one year after the EOL announcement. Once this happens then the EOLed >component can be removed. Of course this removal takes an arc case. That can't be true; that menas we can stop supporting systems in new OS releases one year after we've stopped shipping them? I thought there was a five year rule for that; consequently, the same fve year rule would apply for all components that could be conceivably shipped in those systems. Casper From sacadmin Sat Jun 7 05:19:17 2003 Date: Sat, 7 Jun 2003 13:11:29 +0100 (BST) From: Andrew Gabriel Subject: Re: SFastTrack: PSARC 2003/335 EOL of le Ethernet driver [timeout 06/12/2003] To: psarc@sac.eng.sun.com, nspg-interest@sun.com, Paul.Lodrige@eng.sun.com, hui.chen@sun.com MIME-Version: 1.0 Content-Type: TEXT/plain; charset=us-ascii Content-MD5: DA/pW3hPACrbFtzL4MfJPQ== Content-Length: 846 >That can't be true; that menas we can stop supporting systems in new >OS releases one year after we've stopped shipping them? >I thought there was a five year rule for that; consequently, the same >fve year rule would apply for all components that could be conceivably >shipped in those systems. My understanding is we could EOF the feature in a future Solaris release prior to the five year rule (although in this case, that looks impossible). In order to EOF these cards in S10, this must be announced in the S9U4 release notice. The last date to do that was 5 May, so it's too late now. I think you will also need Solaris PAC approval to EOF such feature support. However, it would be appropriate to start the process so customers can be warned in the S10 release notices that the feature will be EOF'ed in Solaris 11. -- Andrew Gabriel From sacadmin Sun Jun 8 14:24:37 2003 Date: Sun, 08 Jun 2003 14:22:26 -0700 (PDT) From: Hui Chen Subject: Re: SFastTrack: PSARC 2003/335 EOL of le Ethernet driver [timeout 06/12/2003] To: psarc@sac.eng.sun.com, nspg-interest@sun.com, Paul.Lodrige@eng.sun.com, Andrew.Gabriel@Sun.COM MIME-version: 1.0 Content-type: TEXT/plain; charset=us-ascii Content-MD5: Hycnxfx8ajQVbw4lacFngg== Content-Length: 1651 Let me quote from the very first e-mail which triggered this discussion: "The appropriate statements have been made in the previous releases of Solaris S9 4/03 release per the EOL requirements. See: http://www.sun.com/products-n-solutions/hardware/docs/Software/Solaris_on_Sun_Ha rdware/Solaris_9_403/index.html" Is it OK to proceed this case now? Thanks, Hui > Date: Sat, 07 Jun 2003 13:11:29 +0100 (BST) > From: Andrew Gabriel > Subject: Re: SFastTrack: PSARC 2003/335 EOL of le Ethernet driver [timeout 06/12/2003] > To: psarc@sac.eng.sun.com, nspg-interest@sun.com, Paul.Lodrige@eng.sun.com, Hui.Chen@Sun.COM > MIME-version: 1.0 > Content-MD5: DA/pW3hPACrbFtzL4MfJPQ== > Original-recipient: rfc822;huic@hanwk-mail1.sfbay.sun.com > > >That can't be true; that menas we can stop supporting systems in new > >OS releases one year after we've stopped shipping them? > >I thought there was a five year rule for that; consequently, the same > >fve year rule would apply for all components that could be conceivably > >shipped in those systems. > > My understanding is we could EOF the feature in a future Solaris release > prior to the five year rule (although in this case, that looks impossible). > In order to EOF these cards in S10, this must be announced in the S9U4 > release notice. The last date to do that was 5 May, so it's too late now. > I think you will also need Solaris PAC approval to EOF such feature support. > > However, it would be appropriate to start the process so customers can > be warned in the S10 release notices that the feature will be EOF'ed > in Solaris 11. > > -- > Andrew Gabriel > From sacadmin Mon Jun 9 09:53:39 2003 Date: Mon, 9 Jun 2003 10:50:16 -0600 (MDT) From: Brian Botton Subject: Re: SFastTrack: PSARC 2003/335 EOL of le Ethernet driver [timeout 06/12/2003] To: Casper.Dik@sun.com Cc: psarc@sac.eng.sun.com, nspg-interest@sun.com, Andrew.Gabriel@sun.com, Paul.Lodrige@eng.sun.com, hui.chen@sun.com MIME-Version: 1.0 Content-Type: TEXT/plain; charset=us-ascii Content-MD5: rN9y7wA3P0hksDyRhzzPjw== Content-Length: 1223 > > I've done some research on the EOL rules. Full support has to be provided > >for one year after the EOL announcement. Once this happens then the EOLed > >component can be removed. Of course this removal takes an arc case. > > That can't be true; that menas we can stop supporting systems in new > OS releases one year after we've stopped shipping them? This is what the SAC EOL info says. Yes, we can stop supporting systems on the OS releases after the 1 year has expired. We're doing this with the sun4m platform. It hasn't been 5 years since we anounced the EOL. When I was hired 10 years ago it was to support Online:Backup. It was EOLed 12.5 months before 2.4 shipped. So we never tested it on 2.4 and it wasn't officially supported. > I thought there was a five year rule for that; consequently, the same > fve year rule would apply for all components that could be conceivably > shipped in those systems. The 5 year rule is for the current OS as of the 1 year timeout. Thanks, Brian ... ___ _________ _][_n_n___i_i _________ | | Brian Botton (____________I.I_______I.|_______| x49533/(719) 265-6108 /ooOOOO OOOOoo oo oooo oo oo Brian.Botton@Sun.COM From sacadmin Mon Jun 9 10:02:35 2003 Date: Mon, 9 Jun 2003 09:59:15 -0700 (PDT) From: Shudong Zhou Subject: Re: SFastTrack: PSARC 2003/335 EOL of le Ethernet driver [timeout 06/12/2003] To: psarc@sac.eng.sun.com, nspg-interest@sun.com, Paul.Lodrige@eng.sun.com, hui.chen@sun.com, Andrew.Gabriel@sun.com MIME-Version: 1.0 Content-Type: TEXT/plain; charset=us-ascii Content-MD5: oQ18O5jMcW0TtvuAZsylJQ== Content-Length: 1155 > >That can't be true; that menas we can stop supporting systems in new > >OS releases one year after we've stopped shipping them? > >I thought there was a five year rule for that; consequently, the same > >fve year rule would apply for all components that could be conceivably > >shipped in those systems. > > My understanding is we could EOF the feature in a future Solaris release > prior to the five year rule (although in this case, that looks impossible). > In order to EOF these cards in S10, this must be announced in the S9U4 > release notice. The last date to do that was 5 May, so it's too late now. > I think you will also need Solaris PAC approval to EOF such feature support. > > However, it would be appropriate to start the process so customers can > be warned in the S10 release notices that the feature will be EOF'ed > in Solaris 11. > > -- > Andrew Gabriel > I tend to agree that this is a "software only" EOF. If you use the 5-year rule, it should be applied against when Ultra-1 stopped shipping, not when the latest OS stop supporting the platform. After all, you can still run S8/S9 on Ultra-1's, possibly forever. Shudong From sacadmin Wed Jun 11 17:31:51 2003 Date: Wed, 11 Jun 2003 17:22:37 -0700 (PDT) From: Paul Lodrige Subject: PSARC 2003/335 EOL of le Ethernet driver To: psarc@sun.com Cc: nspg-interest@sun.com MIME-Version: 1.0 Content-Type: TEXT/plain; charset=us-ascii Content-MD5: p/YdnZoTv5RfXs4+2I8yuQ== Content-Length: 3631 RE: PSARC 2003/335 EOL of le Ethernet driver After resolving the timeline issues and incorporating comments from Jordan Brown and Joe Kowalski, I am reopening this fast-track. I have also posted the revised fast-track one pager enclosed to the case materials directory at: /shared/sac/Archives/CaseLog/arc/PSARC/2003/335 or http://sac.sfbay.sun.com/arc/PSARC/2003/335/ Time-out has been reset to 06/18/2003 pfl ******************************************************************************* 1. Introduction 1.1. Project/Component Working Name: EOL of le Ethernet driver 1.2. Name of Document Author/Supplier: Paul Lodrige 1.3. Date of This Document: 06/9/03 1.4. Name of Major Document Customer(s)/Consumer(s): PSARC, NSPG, SOESC, Solaris PAC 2. Project Summary 2.1. Project Description This project is to remove le driver support from Solaris 10. 2.2. Risks and Assumptions Risks: A. If the obsolete devices are being utilized internally, this would force a hardware upgrade prior to running Solaris 10. Assumptions A. This case (PSARC/2003/335) depends upon "PSARC/2003/014 EOL of 32-bit SPARC Kernels" and may not integrate before "PSARC/2003/014 EOL of 32-bit SPARC Kernels" is completed. B. In the Solaris 10 release time frame, the le equipped devices should be entirely irrelevant to our customer base. 3. Business Summary 3.1. Problem Area Potentially, there may be some small number of customers still utilizing Lance (le) equipped obsolete devices when Solaris 10 ships, and those customers would not be able to install Solaris 10 until upgrading their hardware. However, none of these devices are relevant for serious commercial computing in the Solaris 10 release timeframe. It may represent a small problem for home or free-Solaris users with old hardware. 3.2. Market/Requester NSPG 3.3. Business Justification We will be able to focus our energies and financial resources on supporting and enhancing strategically important devices as defined by our internal partners such as VSP and ESP, instead of wasting resources supporting those devices that are or will be obsolete by the time Solaris 10 ships. There are no current systems shipping with the le hardware. The most recent system which did have le hardware onboard was the early Ultra 1s and these systems will not be likely candidates for upgrade to S10. The plan is to EOL the le driver in Solaris 10 3.4. Competitive Analysis Focusing our efforts on the drivers for our current products line will enable us to better compete in our target markets. 3.5. Opportunity Window/Exposure Device support being removed from Solaris 10 must be published as removal candidates in the Solaris 9 Or Solaris 9 Update release notes, providing ample advance warning to customers. This notice was given in the Solaris 9 4/03 Release Notes Solaris 10 is the target shipping release for removal 3.6. How will you know when you are done? Integration into the ON 10 gate 4. Technical Description The packages, source files and source lines, man pages exclusively for le are removed from Solaris 10. 5. Reference Documents Document: In End-of-Support Products section of Solaris 9 4/03 Release Notes Supplement for Sun Hardware includes the le driver as Future End-of-Support Products. see: http://www.sun.com/products-n-solutions/hardware/docs/Software/ Solaris_on_Sun_Hardware/Solaris_9_403/index.html From sacadmin Thu Jun 19 09:18:13 2003 Date: Thu, 19 Jun 2003 09:09:00 -0700 (PDT) From: Paul Lodrige Subject: Re: PSARC 2003/335 EOL of le Ethernet driver To: nspg-interest@sun.com Cc: psarc@sun.com, deepak.bhagat@sun.com MIME-Version: 1.0 Content-Type: TEXT/plain; charset=us-ascii Content-MD5: WNGuUf1lZk/0VtKcoGpwKw== Content-Length: 55 This case was approved in yesterday's PSARC meeting.