Date: Thu, 23 Jun 2022 09:44:28 -0600 From: Warner Losh <imp@bsdimp.com> To: Eugene Grosbein <eugen@grosbein.net> Cc: Steve Kargl <sgk@troutmask.apl.washington.edu>, Stefan Blachmann <sblachmann@gmail.com>, Emmanuel Vadot <manu@bidouilliste.com>, Ed Maste <emaste@freebsd.org>, Poul-Henning Kamp <phk@phk.freebsd.dk>, FreeBSD Hackers <freebsd-hackers@freebsd.org> Subject: Re: Reasons for keeping sc(4) and libvgl ? Message-ID: <CANCZdfpZR6H1s=zLTOdCrt429MHy0fdVh5m9HUFMmhnJH6nF3g@mail.gmail.com> In-Reply-To: <8d6ae56e-9265-6b2e-c966-1c51f00f6c88@grosbein.net> References: <20211126160454.3eb827365a02103169ab9adc@bidouilliste.com> <20220621201924.e9b96876c947140ac1f3b7a4@bidouilliste.com> <3d09c86a-9840-f8bf-4725-8098d958a01d@plan-b.pwste.edu.pl> <202206211901.25LJ1uBd067376@critter.freebsd.dk> <CAPyFy2Ca83X042jc5QE-g=eHAfnukHScrTSyaLRi4UxeTBasJQ@mail.gmail.com> <20220622044923.6e2fac81c1e8205872d9de11@bidouilliste.com> <CACc-My0q-Ods_O-TDru=tEwjSOaUJZZd=ZTzD46nY1gjGYO_VA@mail.gmail.com> <CANCZdfoRCB5RuszKWTQazRuseRnapVqMTANnQTR0b61AHw54aQ@mail.gmail.com> <YrKhlZnlF%2ByxZd9X@troutmask.apl.washington.edu> <CANCZdfraihrmoHsu85aFWUT3zwTO%2B-x_Jbcdk0f%2Bxrr7kMxckg@mail.gmail.com> <6b7997d6-f8ed-c4e4-91eb-da9b20eb0a14@grosbein.net> <CANCZdfq1mGynidmm%2BPniYdmb5Tu-fh6N6C8zm8oKwB5mnoj5yA@mail.gmail.com> <8d6ae56e-9265-6b2e-c966-1c51f00f6c88@grosbein.net>
next in thread | previous in thread | raw e-mail | index | archive | help
[-- Attachment #1 --] On Thu, Jun 23, 2022 at 9:37 AM Eugene Grosbein <eugen@grosbein.net> wrote: > 23.06.2022 22:26, Warner Losh wrote: > > > Giant is being removed entirely, and with it all straggler drivers that > aren't converted by the removal date. > > There's no fixed date for this, at the present time, but I'm about to > commit changes that make it impossible > > for new code to reference Giant. Having Giant, at all, causes slow downs > elsewhere in the system, which > > is why we're pushing to remove it entirely. > > Why is it better to lose working code then to keep it "slow"? > Supporting Giant, at all, means creating extra taskqueues, processes, etc. It means extra checks in all the code paths since Giant is so 'special'. To do this just to support an obsolete console seems to many to be an unwise tradeoff once everything else is in order. Especially since there have been years for people that care about the problems to arrange solutions. We are still some time away from everything else eliminating Giant, so there's still time to get things fixed. However, the increasingly obscure nature of the problems and/or their diminished relevancy to the project means that absent code showing up (either from the hobbyist community or from funded work), the problems will remain because the limited resources of those working on the project aren't ample enough for them to be solved. If they are important to you, and nobody else is working on them, now is your chance. Ideally, there'd be enough time and people to solve all the problems, but there is not. Warner [-- Attachment #2 --] <div dir="ltr"><div dir="ltr"><br></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Thu, Jun 23, 2022 at 9:37 AM Eugene Grosbein <<a href="mailto:eugen@grosbein.net">eugen@grosbein.net</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">23.06.2022 22:26, Warner Losh wrote:<br> <br> > Giant is being removed entirely, and with it all straggler drivers that aren't converted by the removal date.<br> > There's no fixed date for this, at the present time, but I'm about to commit changes that make it impossible<br> > for new code to reference Giant. Having Giant, at all, causes slow downs elsewhere in the system, which<br> > is why we're pushing to remove it entirely.<br> <br> Why is it better to lose working code then to keep it "slow"?<br></blockquote><div><br></div><div>Supporting Giant, at all, means creating extra taskqueues, processes, etc. It means extra checks in</div><div>all the code paths since Giant is so 'special'. To do this just to support an obsolete console seems to</div><div>many to be an unwise tradeoff once everything else is in order. Especially since there have been years</div><div>for people that care about the problems to arrange solutions.</div><div><br></div><div>We are still some time away from everything else eliminating Giant, so there's still time to get things</div><div>fixed. However, the increasingly obscure nature of the problems and/or their diminished relevancy</div><div>to the project means that absent code showing up (either from the hobbyist community or from</div><div>funded work), the problems will remain because the limited resources of those working on the</div><div>project aren't ample enough for them to be solved. If they are important to you, and nobody</div><div>else is working on them, now is your chance. Ideally, there'd be enough time and people to</div><div>solve all the problems, but there is not.</div><div><br></div><div>Warner</div></div></div>
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CANCZdfpZR6H1s=zLTOdCrt429MHy0fdVh5m9HUFMmhnJH6nF3g>