[gentoo-user] Re: Computer system email no not working since gmail change.
On 2024-09-27, Dale wrote: >> You do know about the msmtp man page, right? > > Well, if copying someone's known working config file doesn't work right, > I'm not sure a man page is going to help much. The working configuration I provided didn't have an alias (or aliases) command in it. > That said, I changed the alias line and then got a error about the > account info. What error? > I commented it out too. What was "it"? > When I did that, I got this with your test command. [...] > Now things may be working better. So, I restarted smartd and guess > what, I got emails from smart about my hard drives. I have nine hard > drives. I got a email for each one. Can I get a YEPPIE Great!
[gentoo-user] Re: Computer system email no not working since gmail change.
On 2024-09-27, Dale wrote: > OK. I added that line to the config file. Then it gives me this error. > > > root@Gentoo-1 / # echo foo | msmtp -v bogus > msmtp: /etc/msmtprc: line 5: unknown command alias > root@Gentoo-1 / # Sorry, my bad. It's aliases not alias: $ man msmtp | grep alias Note that hdrs is accepted as an alias for headers to be compatible --aliases=[file] Set or unset an aliases file. See the aliases command. aliases [file] Replace local recipients with addresses in the aliases file. The aliases cal address is not found in the aliases file. If no default alias is found, Note that alias expansion only affects the mail envelope. The To and Cc head‐ An empty argument to the aliases command disables the replacement of local # Example aliases file You do know about the msmtp man page, right?
[gentoo-user] Re: Computer system email no not working since gmail change.
On 2024-09-27, Dale wrote: >>> Sep 26 19:04:26 Gentoo-1 smartd[18737]: Executing test of to root ... >>> Sep 26 19:04:36 Gentoo-1 msmtp[18815]: host=smtp.gmail.com tls=on >>> auth=off from=rdalek1967gmail.com recipients=root errormsg='the >>> server sent an empty reply' exitcode=EX_PROTOCOL >> According to that message, the msmtp auth option is off. It needs to >> be on. > This is what I copied from yours, with obvious bits changed. I'll put > "noneya" in those, so you know that isn't the real info. > > syslog LOG_MAIL > > account default > maildomain gmail.org > syslog on > from rdalek1967gmail.com > host smtp.gmail.com > port 465 > tls on > tls_certcheck off > tls_starttls off > auth on > user rdalek1...@gmail.com > password "noneya" # That is the 16 character thing with spaces in it. That's not the config that msmtp is using. See in the log where it says "auth=off"? It should say "auth=on". And I think it's also using the wrong port number (or you should turn tls_startls off). Here's what my log looks like: Sep 26 15:57:47 aleph msmtp[21363]: host=smtp.gmail.com tls=on auth=on user=grant.b.edwa...@gmail.com from=grant.b.edwa...@gmail.com recipients=@.com mailsize=317 smtpstatus=250 smtpmsg='250 2.0.0 OK 1727384267 e9e14a558f8ab-3a344d605d5sm1326875ab.7 - gsmtp' exitcode=EX_OK > I'm by no means a expert on this but I see 'auth on' in there. I see > what you talking about in the error to tho. As I mentioned in other > reply, I think something else is amiss somewhere. If that config works > with gmail for you, it should work here. msmtp is not using the configuration you showed above. > On the root thing, I have a alias set up in some file that tells it that > root is my gmail address. It worked before but maybe not now. File is > here: /etc/mail/aliases It has this info about root being my gmail > address. Msmtp doesn't read an alias file by default. You need to add an alias command to the msmtp config: alias /etc/mail/aliases Then run msmtp from the command line like this so you can see all the settings and the messages exchanged. Pay particular attention to where it's reading the configuration from, port number, and tls_starttls. Also look to see where it's reading aliases from. $ echo foo | msmtp -v bogus msmtp: recipient address bogus not accepted by the server msmtp: server message: 553-5.1.3 The recipient address is not a valid RFC 5321 address. For msmtp: server message: 553-5.1.3 more information, go to msmtp: server message: 553-5.1.3 https://support.google.com/a/answer/3221692 and review RFC 5321 msmtp: server message: 553 5.1.3 specifications. 8926c6da1cb9f-4df9f55sm483180173.166 - gsmtp msmtp: could not send mail (account default from /etc/msmtprc) loaded system configuration file /etc/msmtprc ignoring user configuration file /home/grante/.msmtprc: No such file or directory falling back to default account *** using account default from /etc/msmtprc host = smtp.gmail.com *** port = 465 source ip = (not set) proxy host = (not set) proxy port = 0 socket = (not set) timeout = off protocol = smtp domain = localhost *** auth = choose user = grant.b.edwa...@gmail.com password = * passwordeval = (not set) ntlmdomain = (not set) *** tls = on *** tls_starttls = off tls_trust_file = system tls_crl_file = (not set) tls_fingerprint = (not set) tls_key_file = (not set) tls_cert_file = (not set) tls_certcheck = off tls_min_dh_prime_bits = (not set) tls_priorities = (not set) tls_host_override = (not set) auto_from = off maildomain = gmail.org from = grant.b.edwa...@gmail.com from_full_name = (not set) allow_from_override = on set_from_header = auto set_date_header = auto remove_bcc_headers = on undisclosed_recipients = off dsn_notify = (not set) dsn_return = (not set) logfile = (not set) logfile_time_format = (not set) syslog = LOG_MAIL *** aliases = (not set) reading recipients from the command line TLS session parameters: (TLS1.3)-(ECDHE-X25519)-(ECDSA-SECP256R1-SHA256)-(AES-256-GCM) TLS certificate information: Subject: CN=smtp.gmail.com Issuer: C=US,O=Google Trust Services,CN=WR2 Validity: Activation time: Mon 26 Aug 2024 02:12:09 AM CDT Expiration time: Mon 18 Nov 2024 01:12:08 AM CST Fingerprints: SHA256: 01:AF:90:6E:FC:06:5C:B5:5D:B9:55:AB:27:07:B0:E7:8C:4F:EA:46:70:67:86:A9:E0:F1:BB:F7:5A:2E:1B:64 SHA1 (deprecated): F2:B7:9C:3C:4C:FD:57:31:37:BB:8D:F6:DD:F7:FB:A2:D7:09:B2:BD <-- 220 smtp.gmail.com ESMTP 8926c6da1cb9f-4df9f55sm483180173.166 - gsmtp --> EHLO localhost <-- 250-smtp.gmail.com at your service, [24.152.157.105] <-- 250-SIZE 35882577 <-- 250-8BITMIME <-- 250-AUTH LOGIN PLAIN XOAUTH2 PLAIN-CLIENTTOKE
[gentoo-user] Re: Computer system email no not working since gmail change.
On 2024-09-27, Dale wrote: > It says port 465 but it is using Oauth2 if that matters. It doesn't. > I'll admit, the last time I got this working, I followed a guide and > it just worked. Once it worked, I left it alone. I was scared that > if I touched it, it would stop working. LOL > > I changed the config to port 465 and it still failed with this. > > > Sep 26 19:04:26 Gentoo-1 smartd[18737]: Executing test of to root ... > Sep 26 19:04:36 Gentoo-1 msmtp[18815]: host=smtp.gmail.com tls=on > auth=off from=rdalek1967gmail.com recipients=root errormsg='the > server sent an empty reply' exitcode=EX_PROTOCOL According to that message, the msmtp auth option is off. It needs to be on. Look at the working configuraiton I posted. Use that configuration with your email address and app password. It also looks like smartd is trying to send mail to the recipient "root". That's not a valid destination when sending via most smtp servers. You need to be sending to a recipient that looks like "user@domain". What port number are you trying to use? Please re-read what I wrote about port numbers and starttls. If you're using port 465, you need to turn tls_starttls OFF. If you're using port 587, tls_starttls needs to be ON. That's the default, but I recommend turning it on explicitly. My advice: don't use smartd to try to get msmtp working. Use something easier to work with. The usual way to do it is using something like mailx. The Arch wiki page is an excellent resource: https://wiki.archlinux.org/title/Msmtp It shows how to test msmtp using mailx. It also addresses the error you're seeing explicitly: 8.2 Server sent empty reply If you get a "server sent empty reply" error, this probably means the mail server does not support STARTTLS over port 587, but requires TLS over port 465. Note that some systems will ignore a message without a Message-Id: header, so the example from the Arch page that cats a message from a file to msmtp might not work reliably.
[gentoo-user] Re: Computer system email no not working since gmail change.
On 2024-09-26, Dale wrote: > I removed ssmtp and installed msmtp. I think I got the config set up > but it is different so I may not have it right. It doesn't work tho. > From messages. > > > Sep 26 10:03:33 Gentoo-1 smartd[27728]: Executing test of to root ... > Sep 26 10:05:40 Gentoo-1 msmtp[30861]: host=smtp.gmail.com tls=on > auth=off from=rdalek1967gmail.com recipients=root errormsg='cannot > connect to smtp.gmail.com, port 587: Connection timed out' > exitcode=EX_TEMPFAIL It looks like your network is broken. Try this: $ telnet smtp.gmail.com 587 Trying 209.85.145.109... Connected to smtp.gmail.com. Escape character is '^]'. 220 smtp.gmail.com ESMTP 8926c6da1cb9f-4d60978sm58959173.69 - gsmtp Note: Port 587 is for plaintext connection and then shifting into TLS mode with the starttls command. When configuring msmtp: port 587 tls on tls-starttls on or port 465 tls on tls-starttls off Port 465 starts using TLS immediately $ openssl s_client -connect smtp.gmail.com:465 CONNECTED(0003) depth=2 C = US, O = Google Trust Services LLC, CN = GTS Root R1 verify return:1 depth=1 C = US, O = Google Trust Services, CN = WR2 verify return:1 depth=0 CN = smtp.gmail.com verify return:1 --- Certificate chain 0 s:CN = smtp.gmail.com i:C = US, O = Google Trust Services, CN = WR2 a:PKEY: id-ecPublicKey, 256 (bit); sigalg: RSA-SHA256 v:NotBefore: Aug 26 07:12:09 2024 GMT; NotAfter: Nov 18 07:12:08 2024 GMT [...] --- read R BLOCK 220 smtp.gmail.com ESMTP e9e14a558f8ab-3a344d9250esm496675ab.41 - gsmtp
[gentoo-user] Re: Computer system email no not working since gmail change.
On 2024-09-26, Dale wrote: > root@Gentoo-1 / # telnet smtp.gmail.com 587 > Trying 142.251.116.108... > Trying 2607:f8b0:4023:1000::6c... > telnet: Unable to connect to remote host: Network is unreachable > root@Gentoo-1 / # > > Can't connect. Well, that explains a lot. It can't reach anything to > log into. It looks like it is trying both IPv4 and v6. So, I used > ping. It works there. > > root@Gentoo-1 / # ping smtp.gmail.com > PING smtp.gmail.com (142.250.115.108) 56(84) bytes of data. > 64 bytes from rq-in-f108.1e100.net (142.250.115.108): icmp_seq=1 ttl=53 Those commands are using two different IP addresses. It's not uncommon that DNS resolvers for heavily used services rotate through a pool of addresses, but for testing purposes you should pick a single IP address. >>> I am connected through a VPN but Seamonkey works fine. I can check and >>> send email there, >> Do you mean can send email via Gmail's SMTP server using Seamonkey? >> >> How are the SMTP server settings configured in Seamonkey? >> >> Is it using Oauth2 or an app password? > > I checked, It is using Oauth2. I had to change it a good while back > but I think SMART could still send emails for a while after that. Is it using the same port number as msmtp? >> I just did a quick test, and sending via smtp.gmail.com using an app >> password worked fine from mutt. I don't have msmtp set up at the >> moment. > > That is true but why buy one if you can't run it? LOL This is yet > another reason I want to switch from Gmail. They nothing but nosy > anyway. I think it is common knowledge that they scan all emails and > use the info for various things, including ads, which I block by the > way. There's nothing wrong with Gmail. The app password feature works exactly as documented, and is definitely the right way to provide authentication for "dumb" programs that don't know how to do OAUTH2.
[gentoo-user] Re: Computer system email no not working since gmail change.
On 2024-09-26, Grant Edwards wrote: > I just did a quick test, and sending via smtp.gmail.com using an app > password worked fine from mutt. I don't have msmtp set up at the > moment. Ijust set up msmtp and it works too. Below is the msmtp config, * If you want, replace "account gmail" with "account default", then you don't have to supply the "-a gmail" option to msmtp. * replace (in two places) usern...@gmail.com with your email address. * replace with your app password. ---8<- syslog LOG_MAIL account gmail maildomain gmail.org syslog on from gmailusern...@gmail.com host smtp.gmail.com port 465 tls on tls_certcheck off tls_starttls off auth on user gmailusern...@gmail.com password " " ---8<-
[gentoo-user] Re: Computer system email no not working since gmail change.
On 2024-09-26, Dale wrote: > Grant Edwards wrote: > >> It looks like your network is broken. Try this: >> >> $ telnet smtp.gmail.com 587 >> Trying 209.85.145.109... >> Connected to smtp.gmail.com. >> Escape character is '^]'. >> 220 smtp.gmail.com ESMTP 8926c6da1cb9f-4d60978sm58959173.69 - gsmtp >> >> Note: Port 587 is for plaintext connection and then shifting into TLS >> mode with the starttls command. When configuring msmtp: >> >> port 587 >> tls on >> tls-starttls on >> or >> port 465 >> tls on >> tls-starttls off >> >> Port 465 starts using TLS immediately >> >> [...] > Which package do I need for that telnet? I see a few packages with that > name. Any of them should work. I use net-misc/netkit-telnetd. It will also install a telnet daemon, but it won't enbale it > I am connected through a VPN but Seamonkey works fine. I can check and > send email there, Do you mean can send email via Gmail's SMTP server using Seamonkey? How are the SMTP server settings configured in Seamonkey? Is it using Oauth2 or an app password? > hence this thread. I figure I got something set up wrong after a > Gmail change. Sort of stupid to make something so secure you can't > use it. It's like buying a computer and saying the only way to > secure it is to keep it turned off. :/ Well, that last part is true, as long as you keep it in a locked room. I just did a quick test, and sending via smtp.gmail.com using an app password worked fine from mutt. I don't have msmtp set up at the moment.
[gentoo-user] Re: Computer system email no not working since gmail change.
On 2024-09-26, Grant Edwards wrote: > On 2024-09-26, Dale wrote: > >> Sep 26 10:03:33 Gentoo-1 smartd[27728]: Executing test of to root ... >> Sep 26 10:05:40 Gentoo-1 msmtp[30861]: host=smtp.gmail.com tls=on >> auth=off from=rdalek1967gmail.com recipients=root errormsg='cannot >> connect to smtp.gmail.com, port 587: Connection timed out' >> exitcode=EX_TEMPFAIL > > It looks like your network is broken. Try this: > > $ telnet smtp.gmail.com 587 > Trying 209.85.145.109... > Connected to smtp.gmail.com. > Escape character is '^]'. > 220 smtp.gmail.com ESMTP 8926c6da1cb9f-4d60978sm58959173.69 - gsmtp BTW, some ISPs block connections to SMTP ports on hosts other than their own mail servers. [Which is a particularly annoying type of "broken".]
[gentoo-user] Re: Computer system email no not working since gmail change.
On 2024-09-26, cov...@ccs.covici.com wrote: > > I was looking at fastmail but which ever you go with, you should setup > your own domain and that way if you don't like your provider you can > change without changing your email address. I"ve decided to do that several times over the past couple decades. Each time I started out trying to pick a domain (that was avaialable) which I wanted to live with for the rest of my life. I could never get past that step. -- Grant
[gentoo-user] Re: Computer system email no not working since gmail change.
On 2024-09-26, Dale wrote: > >> It used to be the case you could set up an 'App Password' without having to >> provide them with your phone number and other 'none-of-their-business' >> personal information, but for some years now they have been asking for more >> information to allow you to complete setting up 'App Password' for your >> Google >> account(s) and device(s). >> > > I also tried the 16 character password method too. It still didn't work. For what value of "didn't work"? I don't use it regularly, but I set it up several years ago and it was working fine the last time I tried it (a couple months ago). -- Grant
[gentoo-user] Re: Why did Synaptics touchpad stop working?
On 2024-09-09, Jack wrote: > On 9/8/24 10:20 PM, Grant Edwards wrote: > >> This morning when I booted my Thinkpad T580, the Synaptics touchpad >> buttons didn't work at all, and the "pointer" function just barely >> worked: the response was slow and jerky with a noticeable delay. >> >> In order to get it working again, I had to enable some rmi4 stuff in >> my kernel config: >>[...] >> Apparently, it used to work as a PS/2 mouse, but then it "just quit", >> and I hand to enable rmi4 SMBus support. >>[...] >> As I said, I've got it working again, but I'm baffled what caused it >> to stop working. > > Any chance some internal connection has come loose? How old is the > laptop? I'd probably suspect hardware before software, but I wouldn't > put money on anything. It's a bit less than 5 years old and has been very lightly used. It only leaves the house maybe a half-dozen times a year. I'd probably put my money down on the spot marked "Microsoft's Fault", but I've got no real theory as to why that would be unless a recent Windows update somehow changed a BIOS setting for Synaptics PS/2 compatiblity. [I never got to the point of messing with BIOS settings.] -- Grant
[gentoo-user] Why did Synaptics touchpad stop working?
This morning when I booted my Thinkpad T580, the Synaptics touchpad buttons didn't work at all, and the "pointer" function just barely worked: the response was slow and jerky with a noticeable delay. In order to get it working again, I had to enable some rmi4 stuff in my kernel config: < # CONFIG_RMI4_CORE is not set --- > CONFIG_RMI4_CORE=y > # CONFIG_RMI4_I2C is not set > CONFIG_RMI4_SMB=m > CONFIG_RMI4_F03=y > CONFIG_RMI4_F03_SERIO=y > CONFIG_RMI4_2D_SENSOR=y > CONFIG_RMI4_F11=y > CONFIG_RMI4_F12=y > # CONFIG_RMI4_F30 is not set > # CONFIG_RMI4_F34 is not set > # CONFIG_RMI4_F3A is not set > # CONFIG_RMI4_F55 is not set Apparently, it used to work as a PS/2 mouse, but then it "just quit", and I hand to enable rmi4 SMBus support. The Gentoo Wiki page didn't mention the rmi4 options and instead said to enable some I2C HID stuff that didn't seem to be relevent for my hardware. There was a dmsg message that pointed me towards the rmi4 options. What's really bugging me is why did it quit working? Nothing relevent seems to have been updated recently: 1725658908: >>> emerge (1 of 6) dev-libs/libffi-3.4.6 to / 1725658926: >>> emerge (2 of 6) dev-libs/gobject-introspection-1.78.1-r1 to / 1725658987: >>> emerge (3 of 6) media-libs/graphene-1.10.8-r1 to / 1725659006: >>> emerge (4 of 6) gui-libs/gtk-4.12.5-r2 to / 1725659219: >>> emerge (5 of 6) app-crypt/gcr-4.2.1 to / 1725659258: >>> emerge (6 of 6) app-crypt/pinentry-1.3.0-r3 to / 1725762247: >>> emerge (1 of 7) sys-firmware/intel-microcode-20240813_p20240815 to / 1725762261: >>> emerge (2 of 7) dev-python/jaraco-context-6.0.1 to / 1725762273: >>> emerge (3 of 7) dev-python/setuptools-73.0.1 to / 1725762292: >>> emerge (4 of 7) dev-python/idna-3.8 to / 1725762305: >>> emerge (5 of 7) dev-python/truststore-0.9.2 to / 1725762317: >>> emerge (6 of 7) gui-libs/gtk-4.14.4-r1 to / 1725762557: >>> emerge (7 of 7) www-client/google-chrome-128.0.6613.119 to / I did boot into Windows yesterday at one point, but I do that fairly regularly (a few times a month) and it has never caused any problems in the past. As I said, I've got it working again, but I'm baffled what caused it to stop working. -- Grant
[gentoo-user] Re: Package compile failures with "internal compiler error: Segmentation fault".
On 2024-09-04, Dale wrote: > At one point, I looked for a set of four sticks of the memory. I > couldn't find any. They only come in sets of two. I read somewhere > that the mobo expects each pair to be matched. Yep, that's definitely how it was supposed to work. I fully expected my two (identically spec'ed) sets of two work. All the documentation I could find said it should. It just didn't. :/ -- Grant
[gentoo-user] Re: Package compile failures with "internal compiler error: Segmentation fault".
On 2024-09-04, Frank Steinmetzger wrote: > Am Wed, Sep 04, 2024 at 07:09:43PM - schrieb Grant Edwards: >> […] >> I plugged them in alongside the recently purchased pair. Wouldn't >> work. Either pair of SIMMs worked fine by themselves, but the only way >> I could get both pairs to work together was to drop the clock speed >> down to about a third the speed they were supposed to support. > > Indeed that was my first thought when Dale mentioned getting another > pair. I don’t know if it’s true for all Ryzen chips, but if you use > four sticks, they may not work at the maximum speed advertised by > AMD (not counting in overlcocking). If you kept the settings to Auto > you shouldn’t get problems, but RAM may work slower then. Yea, I thought auto should work, but it didn't. I had to manually lower the RAM clock speed to get all four to work at the same time. The BIOS screens were a bit mind-boggling (very high on graphics, dazzle, and flash -- very low on usability). So it's possible I didn't really have auto mode correctly enabled. > OTOH, since you don’t do hard-core gaming or scientific > number-crunching, it is unlikely you will notice a difference in > your every-day computing. In my case I compared an "emerge" that took several minutes, and it took significantly longer with the lower RAM clock speed. I decided I was better off with fewer GB of faster RAM.
[gentoo-user] Re: Package compile failures with "internal compiler error: Segmentation fault".
On 2024-09-04, Dale wrote: > I ordered another set of memory sticks. I figure I will have to send > them both back which means no memory at all. I wasn't planning to go to > 128GBs yet but guess I am now. [...] Good luck. The last time I had one fail, I needed the machine for work and couldn't wait for the replacement to ship. So, I went to either MicroCenter or Best Buy and picked up another pair of SIMMs with the exact same specs (different brand, of course). A couple weeks later, my replacemnts arrived. "Yippe!" I say to myself, "twice as much RAM!" I plugged them in alongside the recently purchased pair. Wouldn't work. Either pair of SIMMs worked fine by themselves, but the only way I could get both pairs to work together was to drop the clock speed down to about a third the speed they were supposed to support. I read and re-read the motherboard specs and manual. I spent hours tweaking different memory settings in the BIOS, but no joy. Now I've got a backup pair of SIMMs sitting on the shelf. -- Grant
[gentoo-user] Re: Package compile failures with "internal compiler error: Segmentation fault".
On 2024-09-04, Dale wrote: > I forgot to ask, is there anything else that bad memory could affect? > I'm doing the emerge -e world to make sure no programs were affected but > what about other stuff? Could this affect hard drive data for example? Unfortunately, yes. I have had some failing RAM that resulted in some files getting corrupted. -- Grant
[gentoo-user] Re: Package compile failures with "internal compiler error: Segmentation fault".
On 2024-09-04, Frank Steinmetzger wrote: > Am Wed, Sep 04, 2024 at 05:48:29AM -0500 schrieb Dale: > >> I wonder how much fun getting this memory replaced is going to be. o_O > > I once had a bad stick of Crucial Ballistix DDR3. I think it also started > with GCC segfaults. So I took a picture of the failing memtest, e-mailed > that to crucial and they sent me instructions what to do. Yep, I got free replacement from Crucial once years ago also. It was pretty easy, but it took a week or two. -- Grant
[gentoo-user] Re: Package compile failures with "internal compiler error: Segmentation fault".
On 2024-09-03, Dale wrote: > I was trying to re-emerge some packages. The ones I was working on > failed with "internal compiler error: Segmentation fault" or similar > being the common reason for failing. In my experience, that usually means failing RAM. I'd try running memtest86 for a day or two. -- Grant
[gentoo-user] Re: emerge keeps installing then uninstalling qtbase, qttools, ...
On 2024-09-03, Grant Edwards wrote: > On 2024-09-03, Matthew Brooks wrote: > >> It might be worth seeing what a full update of world, with the >> --emptytree flag says (though without actually doing the >> rebuild). Sometimes including that will notice inconsistencies that >> a regular emerge doesn't spot. > > I don't see anything. It still wants to install those 4 "new" packages. I uninstalled ksnip and then did a --depclean (which uninstalled 8-10 packages). Now portage no longer wants to repeatedly install and uninstall those four packages. :/
[gentoo-user] Re: emerge keeps installing then uninstalling qtbase, qttools, ...
On 2024-09-03, Matthew Brooks wrote: > It might be worth seeing what a full update of world, with the > --emptytree flag says (though without actually doing the > rebuild). Sometimes including that will notice inconsistencies that > a regular emerge doesn't spot. I don't see anything. It still wants to install those 4 "new" packages.
[gentoo-user] Re: emerge keeps installing then uninstalling qtbase, qttools, ...
On 2024-09-03, Grant Edwards wrote: > On 2024-09-03, Matthew Brooks wrote: > >> While I'm not familiar enough with those packages to know for >> certain, it sounds like they're probably *build* dependencies for >> something, [...] > > I don't think so. Nothing else is getting built. > > I can alternately do -auvND and --dpeclean all day, and only those > four get installed and then uninstalled. Just to clarify, that's not doing any --sync operations at all. I'm simply alternating --depclean and -auvND world.
[gentoo-user] Re: emerge keeps installing then uninstalling qtbase, qttools, ...
On 2024-09-03, Matthew Brooks wrote: > It might be worth seeing what a full update of world, with the > --emptytree flag says (though without actually doing the > rebuild). Sometimes including that will notice inconsistencies that > a regular emerge doesn't spot. Thanks, I'll try that next time that machine is up. I'm not sure, but I think this may have started when I installed ksnip from the guru overlay: https://gpo.zugaina.org/x11-misc/ksnip It has Qt 5 deps (and there might not be any other Qt apps on that machine). But, it's Qt 6 stuff that keeps getting installed/uninstalled. So that doesn't really make sense either. -- Grant
[gentoo-user] Re: emerge keeps installing then uninstalling qtbase, qttools, ...
On 2024-09-03, Matthew Brooks wrote: > While I'm not familiar enough with those packages to know for > certain, it sounds like they're probably *build* dependencies for > something, but not actual *runtime* dependencies, and so depclean > prunes them, and then whenever the package that needs them gets > built again they get pulled in again. I don't think so. Nothing else is getting built. I can alternately do -auvND and --dpeclean all day, and only those four get installed and then uninstalled.
[gentoo-user] Re: emerge keeps installing then uninstalling qtbase, qttools, ...
On 2024-09-03, Grant Edwards wrote: > For the past 4 days or so, every time I do a sync and then > 'emerge -auvND world', portage installs the following: > > qttools > qtbase > qttranslations > xcb-util-cursor > > Afterwards, when I do 'emerge --depclean', it uninstalls them. > > Any ideas why? It's getting a little annoying. # emerge -auvNDt world * IMPORTANT: 30 news items need reading for repository 'gentoo'. * Use eselect news read to view new items. These are the packages that would be merged, in reverse order: Calculating dependencies... done! Dependency resolution took 26.51 s (backtrack: 0/20). [nomerge ] dev-qt/qttools-6.7.2:6/6.7.2::gentoo USE="assistant linguist qdbus widgets (zstd) -clang -designer -distancefieldgenerator -gles2-only -opengl -pixeltool -qdoc -qml -qtattributionsscanner -qtdiag -qtplugininfo -vulkan" LLVM_SLOT="18 -15 -16 -17" [nomerge ] dev-qt/qtbase-6.7.2-r4:6/6.7.2::gentoo USE="X concurrent dbus gtk gui libinput network nls sql sqlite ssl udev widgets xml (zstd) -accessibility -brotli -cups -eglfs -evdev -gles2-only -gssapi -icu -journald -libproxy -mysql -oci8 -odbc -opengl -postgres -renderdoc -sctp -syslog -test -tslib -vulkan -wayland" [ebuild N ] dev-qt/qttranslations-6.7.2:6/6.7.2::gentoo 0 KiB [ebuild N ]dev-qt/qttools-6.7.2:6/6.7.2::gentoo USE="assistant linguist qdbus widgets (zstd) -clang -designer -distancefieldgenerator -gles2-only -opengl -pixeltool -qdoc -qml -qtattributionsscanner -qtdiag -qtplugininfo -vulkan" LLVM_SLOT="18 -15 -16 -17" 0 KiB [ebuild N ] dev-qt/qtbase-6.7.2-r4:6/6.7.2::gentoo USE="X concurrent dbus gtk gui libinput network nls sql sqlite ssl udev widgets xml (zstd) -accessibility -brotli -cups -eglfs -evdev -gles2-only -gssapi -icu -journald -libproxy -mysql -oci8 -odbc -opengl -postgres -renderdoc -sctp -syslog -test -tslib -vulkan -wayland" 0 KiB [ebuild N ] x11-libs/xcb-util-cursor-0.1.5::gentoo ABI_X86="(64) -32 (-x32)" 0 KiB Total: 4 packages (4 new), Size of downloads: 0 KiB
[gentoo-user] emerge keeps installing then uninstalling qtbase, qttools, ...
For the past 4 days or so, every time I do a sync and then 'emerge -auvND world', portage installs the following: qttools qtbase qttranslations xcb-util-cursor Afterwards, when I do 'emerge --depclean', it uninstalls them. Any ideas why? It's getting a little annoying. -- Grant
[gentoo-user] Re: Do I need firmware for an integrated graphics unit?
On 2024-08-23, Alan Mackenzie wrote: >>> That's a separate graphics card, isn't it? I'm trying to use the >>> integrated graphics processor on my Ryzen 7900. > >> No, it's integrated into the Ryzen 5 3400G. > > Sorry, I didn't recognise the chip number. Is it a laptop chip rather > than a desktop one? No it's a desktop Zen+ (Picasso) socket AM4 processor with a Radeon RX Vega-11 GPU. It's about 5 years old. https://www.techpowerup.com/cpu-specs/ryzen-5-3400g.c2204 > Anyhow, in the last couple of days, I've managed to get my new system to > boot, and I'm currently busy trying to bring X and XFCE4 up. So I'm > getting there, slowly. Hang in there...
[gentoo-user] Re: Do I need firmware for an integrated graphics unit?
On 2024-08-21, Wol wrote: > On 21/08/2024 14:49, Michael wrote: >>> That would involve me learning how to make and handle a modular kernel, >>> something I'd really rather not have to do. > >> Well, there's nothing to it really. Just configure your kernel with the >> drivers needed by your graphics card, but set them as modules. Then boot >> with >> it and check dmesg. The kernel will load the modules and try to fetch the >> requisite firmware. > > Just don't forget to "make modules" and "make modules_install". Then you > need to make sure it's included in any initramfs, so no it's not quite > as simple as "but set them as modules" if you don't have any other modules. Ah, I forgot about initramfs [I never use one]. Do the amdgpu firmware files really need to part of the initramfs or can the modules wait until the real /usr is mounted to load the firmware?
[gentoo-user] Re: Do I need firmware for an integrated graphics unit?
On 2024-08-21, Michael wrote: > Alternatively, as Wol mentioned, you can set up your kernel graphics drivers > as modules (temporarily) and inspect dmesg to find out what firmware is being > loaded. Then use this information to add the firmware file names to be built > in the kernel and also configure to be built-in any kernel graphics drivers. That's by far the simplest option.
[gentoo-user] Re: Do I need firmware for an integrated graphics unit?
On 2024-08-21, Alan Mackenzie wrote: > Hello, Wol. > > On Wed, Aug 21, 2024 at 11:32:18 +0100, Wol wrote: >> On 20/08/2024 22:16, Peter Böhm wrote: >> > Hello Alan, > >> >> Anyhow, I'm up to the stage of configuring the kernel, and I'm stuck at >> >> the bit where I need to specify the firmware to be incorporated into the >> >> kernel for the integrated graphics processor. > >> > Yes, you surely will need some firmware files for your GPU of your AMD >> > CPU. The >> > easiest way to find out what you need is: > >> The other thing I'll throw in is do you actually want to load the >> firmware into the kernel? > > Yes, I do. Why? It's a lot simpler letting it get loaded at boot time. -- Grant
[gentoo-user] Re: Do I need firmware for an integrated graphics unit?
On 2024-08-21, Alan Mackenzie wrote: > Hello, Grant. > > On Wed, Aug 21, 2024 at 00:30:25 -0000, Grant Edwards wrote: >> On 2024-08-20, Alan Mackenzie wrote: > >> > I've just treated myself to a new machine based on a Ryzen 9 7900 >> > processor. I chose the second newest generation so as not to get caught >> > out with not quite debugged systems like I did the last time round. > >> > Anyhow, I'm up to the stage of configuring the kernel, and I'm stuck at >> > the bit where I need to specify the firmware to be incorporated into the >> > kernel for the integrated graphics processor. > >> I'm running a Ryzen 5 3400G with Radeon Vega graphics: > > That's a separate graphics card, isn't it? I'm trying to use the > integrated graphics processor on my Ryzen 7900. No, it's integrated into the Ryzen 5 3400G.
[gentoo-user] Re: Do I need firmware for an integrated graphics unit?
On 2024-08-20, Alan Mackenzie wrote: > I've just treated myself to a new machine based on a Ryzen 9 7900 > processor. I chose the second newest generation so as not to get caught > out with not quite debugged systems like I did the last time round. > > Anyhow, I'm up to the stage of configuring the kernel, and I'm stuck at > the bit where I need to specify the firmware to be incorporated into the > kernel for the integrated graphics processor. I'm running a Ryzen 5 3400G with Radeon Vega graphics: $ lspci | grep -i vga 2a:00.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI] Picasso/Raven 2 [Radeon Vega Series / Radeon Vega Mobile Series] (rev c8) $ dmesg | grep -i firmware [0.091814] Spectre V2 : Enabling Speculation Barrier for firmware calls [0.244487] ACPI: [Firmware Bug]: BIOS _OSI(Linux) query ignored [0.261256] acpi PNP0A08:00: [Firmware Info]: MMCONFIG for domain [bus 00-3f] only partially covers this bridge [3.021472] Loading firmware: amdgpu/picasso_gpu_info.bin [3.045180] Loading firmware: amdgpu/picasso_sdma.bin [3.046559] Loading firmware: amdgpu/picasso_asd.bin [3.047121] Loading firmware: amdgpu/picasso_ta.bin [3.047408] Loading firmware: amdgpu/raven_dmcu.bin [3.048027] Loading firmware: amdgpu/picasso_pfp.bin [3.048163] Loading firmware: amdgpu/picasso_me.bin [3.048380] Loading firmware: amdgpu/picasso_ce.bin [3.048491] Loading firmware: amdgpu/picasso_rlc_am4.bin [3.048624] Loading firmware: amdgpu/picasso_mec.bin [3.048891] Loading firmware: amdgpu/picasso_mec2.bin [3.049687] Loading firmware: amdgpu/picasso_vcn.bin [3.050007] [drm] Found VCN firmware Version ENC: 1.15 DEC: 3 VEP: 0 Revision: 0 [3.050015] amdgpu :2a:00.0: amdgpu: Will use PSP to load VCN firmware [5.407436] Loading firmware: rtl_nic/rtl8168h-2.fw $ zcat /proc/config.gz | grep -i firmware CONFIG_PREVENT_FIRMWARE_BUILD=y # Firmware loader CONFIG_EXTRA_FIRMWARE="" # end of Firmware loader # Firmware Drivers CONFIG_FIRMWARE_MEMMAP=y # CONFIG_GOOGLE_FIRMWARE is not set # Tegra firmware driver # end of Tegra firmware driver # end of Firmware Drivers # CONFIG_DRM_LOAD_EDID_FIRMWARE is not set # CONFIG_FIRMWARE_EDID is not set # CONFIG_TEST_FIRMWARE is not set CONFIG_GENTOO_PRINT_FIRMWARE_INFO=y $ equery files linux-firmware | egrep 'picasso_|raven_' /lib/firmware/amdgpu/picasso_asd.bin /lib/firmware/amdgpu/picasso_ce.bin /lib/firmware/amdgpu/picasso_gpu_info.bin /lib/firmware/amdgpu/picasso_me.bin /lib/firmware/amdgpu/picasso_mec.bin /lib/firmware/amdgpu/picasso_mec2.bin /lib/firmware/amdgpu/picasso_pfp.bin /lib/firmware/amdgpu/picasso_rlc.bin /lib/firmware/amdgpu/picasso_rlc_am4.bin /lib/firmware/amdgpu/picasso_sdma.bin /lib/firmware/amdgpu/picasso_ta.bin /lib/firmware/amdgpu/picasso_vcn.bin /lib/firmware/amdgpu/raven_asd.bin /lib/firmware/amdgpu/raven_ce.bin /lib/firmware/amdgpu/raven_dmcu.bin /lib/firmware/amdgpu/raven_gpu_info.bin /lib/firmware/amdgpu/raven_kicker_rlc.bin /lib/firmware/amdgpu/raven_me.bin /lib/firmware/amdgpu/raven_mec.bin /lib/firmware/amdgpu/raven_mec2.bin /lib/firmware/amdgpu/raven_pfp.bin /lib/firmware/amdgpu/raven_rlc.bin /lib/firmware/amdgpu/raven_sdma.bin /lib/firmware/amdgpu/raven_ta.bin /lib/firmware/amdgpu/raven_vcn.bin
[gentoo-user] Re: 17 new packages because pip wants to write poetry?
On 2024-07-29, Eli Schwartz wrote: >> [...] I (for one) would appreciate some sort of notice when such an >> unbundling happens so that I don't waste time trying to track down >> why emerge suddenly wants to install a bunch of new packages. I >> can't really come up with a good mechanism for that other than news >> items. > > Well, it is there in the `git log` of the package. And at > https://packages.gentoo.org/packages/dev-python/pip/changelog > > Commits on 2024-06-23 > > dev-python/pip: Unbundle dependencies Right, but that's only useful after you track down the trigger for the new packages. What would be nice is avoiding that "tracking down" effort. [I know, I should just relax, hit 'Y', and trust that emerge and the devs know what they're doing.] >> OK, now I'm genuinely curious: what does pip need to do with markdown >> and RTF? My first guess was that its man pages aren't written in nroff >> any more and somehow markdown was being used. [I already had at least >> one markdown implementation installed, but pip apparently demands a >> different one.] But there is no man page for pip. There are a bunch of >> documents in /usr/share/doc/pip-, but they're all reStructured Text. >> >> Oh well, I guess I should be thankful it didn't force LaTeX and pandoc >> installs. > > Python software loves to depend on python software and hates to depend > on anything that isn't written in python, so I think you're pretty safe > there. :) Good point. :) > As for why it needs to format markdown -- build dependencies of > python software often do, since they want to render the darned stuff > into the https://pypi.org display page for that software Oh. Is that display page (in html?) written into a log somewhere or shown during the build? > and the one and only way to do so is to render it into the metadata > files which get installed at runtime -- and which are also directly > displayed on https://pypi.org > > I don't understand it myself, either. You certainly understand more of it than I do. > In this case it's actually a bit worse because pip internally uses, and > usually bundles a code copy, of https://pypi.org/project/rich/ > > "rich" can do a lot of things, including take markdown and print it with > fancy formatting and colors to your terminal emulator. pip isn't using > most of the features of rich, but it is *using* rich at all, and > therefore markdown ends up as a recursive dependency. > ... > > aside: there are pip manpages, funny you should mention that. When installed on Gentoo using dev-python/pip? > I could totally add another bdepend on sphinx for this! But I would have > to package some things first. :( No thanks, sphinx would pull in 10 more packages. :) If I need pip documentation, I can google for it or look at the rst.bz2 files install in /usr/share/doc... Thanks for tolerating my whinging. -- Grant
[gentoo-user] Re: 17 new packages because pip wants to write poetry?
On 2024-07-29, Eli Schwartz wrote: > >> It turns out dev-python/poetry has nothing to do with poetry, so my AI >> paranoia was unjustified (this time), but one wonders what devs are >> thinking when the decide they add dozens of new dependencies like >> that. Why does pip suddenly need to format (or produce?) both markdown >> and RTF when it's been able to get along fine without them for so many >> years? > > For many years, pip has contained bundled libraries. These libraries > recently got unbundled, and now you're installing a system copy. Yep, I figured that was probably the answer. I (for one) would appreciate some sort of notice when such an unbundling happens so that I don't waste time trying to track down why emerge suddenly wants to install a bunch of new packages. I can't really come up with a good mechanism for that other than news items. > pip has always "needed to format (or produce?) both markdown and RTF", OK, now I'm genuinely curious: what does pip need to do with markdown and RTF? My first guess was that its man pages aren't written in nroff any more and somehow markdown was being used. [I already had at least one markdown implementation installed, but pip apparently demands a different one.] But there is no man page for pip. There are a bunch of documents in /usr/share/doc/pip-, but they're all reStructured Text. Oh well, I guess I should be thankful it didn't force LaTeX and pandoc installs. > but it also "needs" to use bundled libraries to do it without people > noticing that it does it. Some of those packages are only bdeps, > and you can feel free to e.g. delete poetry via > emerge -c --with-bdeps=n once you're done updating. I usually just leave bdeps installed. Otherwise that would remove 130 other packages as well (some of which take a looong time to build). -- Grant
[gentoo-user] 17 new packages because pip wants to write poetry?
This morning a routine emerge -auvND wanted to install 17 new packages for no apparent reason. Adding a 't' to the emerge options seems to point to pip, which now wants to install a whole shed-load of new packages — among them dev-python/poetry and a bunch of markdown and rich-text libraries. Oh great, pip has incorporated AI, become self-aware, and wants to write angst-ridden poetry about the futility of "life". It turns out dev-python/poetry has nothing to do with poetry, so my AI paranoia was unjustified (this time), but one wonders what devs are thinking when the decide they add dozens of new dependencies like that. Why does pip suddenly need to format (or produce?) both markdown and RTF when it's been able to get along fine without them for so many years? -- Grant
[gentoo-user] Re: Can't get the GUI to stay up for more than a minute or so before crashing
On 2024-06-28, Dale wrote: > Before I ran out of steam this morning, I tried the nouveau drivers > again. I never can remember how to spell that. :/ I unmerged the > nvidia drivers to do this. I used the in tree nouveau drivers tho. For > some reason, even tho I removed the nvidia package and rebooted, it > still showed it was loading the nvidia drivers which shouldn't even > exist. No matter what I did, it loaded the nvidia drivers. I could see > it with lsmod and lspci -k. It's like I can't get rid of the nvidia > drivers now. There are two parts to the nvidia drivers: the kernel module (which is what you're seeing with lspci -k and lsmod) and the user-space Xorg driver (which is presumably what you unmerged). -- Grant
[gentoo-user] Re: Eliminating unwanted linux-firmware blobs.
On 2024-06-28, Eli Schwartz wrote:
> On 6/28/24 6:31 PM, Peter Humphrey wrote:
>> On Friday, 28 June 2024 20:32:11 BST Tsukasa Mcp_Reznor wrote:
>>
>>> Remove the date.so it becomes
>>> /etc/portage/savedconfig/sys-kernel/linux-firmware then it applies to all
>>> of them and not the specified version.
>>>
>>> Hope that helps
>>
>> It certainly does. I wish I'd known that years ago: it would have saved me
>> an
>> awful lot of renaming.
>
> The elog message states:
>
> * Your configuration for sys-kernel/linux-firmware-20240610-r1 has been
> saved in
> * "/etc/portage/savedconfig/sys-kernel/linux-firmware-20240610-r1" for
> your editing pleasure.
> * You can edit these files by hand and remerge this package with
> * USE=savedconfig to customise the configuration.
> * You can rename this file/directory to one of the following for
> * its configuration to apply to multiple versions:
> * ${PORTAGE_CONFIGROOT}/etc/portage/savedconfig/
> * [${CTARGET}|${CHOST}|""]/${CATEGORY}/[${PF}|${P}|${PN}]
>
> I admit that this is a bit hard to analyze...
Just a bit. :)
Sure would be nice if it was mentioned on whe Wiki page. Mayber later
today...
--
Grant
[gentoo-user] Re: Eliminating unwanted linux-firmware blobs.
On 2024-06-28, Tsukasa Mcp_Reznor wrote: > Remove the date.so it becomes > /etc/portage/savedconfig/sys-kernel/linux-firmware > then it applies to all of them and not the specified version. Yes, that's the clue I was missing. -- Grant
[gentoo-user] Eliminating unwanted linux-firmware blobs.
Is there any graceful way to handle the elimination of unwanted linux-firmware blobs when doing an update? I believe I understand the process as outlined at https://wiki.gentoo.org/wiki/Linux_firmware: 1. install/upgrade sys-kernel/linux-firmware 2. edit /etc/portage/savedconfig/sys-kernel/linux-firmware-ddmm 3. re-emerge sys-kernel/linux-firmware with the savedcofnig USE flag I've tried that a few times, but it's rather annoying to have to do that every time linux-firmware gets updated. AFAICT, the list of three or four blobs that I actually need on a specific machine never changes. It seems like there ought to be a way to configure that required firmware list and have the emerge -u "just work", but I can't find it. Have I missed something? Yes, I know... Disk space is cheap. Premature optimization ... etc. It still annoys me. -- Grant
[gentoo-user] Re: Config file updates and using diff.
On 2024-06-27, Dale wrote: > Howdy, > > I just finished a large update on my main rig. I have a lot of config > files to update and some have new entries that are needed but I don't > want to lose the ones I've already set. Usually, I just pick the new > one and have a saved copy of the old config to put back my settings. > This is a lot of config files tho. I been using dispatch-conf for this > but I've never figured out how to use the diff feature and tell it which > parts I want to keep and what I want to update. > > First, I've looked on the wiki and can't find a howto on using the diff > tool. Is there a guide on there I can't find? Second, is there a > better way to do this, a very easy way if you will? Like maybe a GUI > thing where I use the mouse to select. It would be nice if it is > something I can easily remember how to do given how rare it is I need to > do a diff of the files. I used the "meld" utility for doing visual side-by side diffs and resolving merges. At one point, I had figured out how to get etc-udpate to invoke meld for me, but I've lost that setting somehow. I usually just use etc-udpate, and then individually invoke meld with the two file paths shown by etc-udpate. I merge the relevent bits of the new default config file into my existing one using meld, then I tell etc-udpate to "discard the new file" (or whatever that option is).
[gentoo-user] Re: Can't get the GUI to stay up for more than a minute or so before crashing
On 2024-06-24, Dale wrote: > Michael wrote: >> On Monday, 24 June 2024 20:47:15 BST Dale wrote: >> >>> Have you seen this before? >> No, because I've never used dracut. > > I just had a thought. I have /usr on the root partition now. Do I even > need a init thingy? Same question as always: does your kernel have enough built-in drivers/modules to mount the root fileystem on /? If yes, then you don't need an initrd. If no, then you do need an initrd. I don't think where /usr is matters, does it? -- Grant
[gentoo-user] Re: Arrow and edit keys?
On 2024-06-15, Michael wrote: > On Saturday, 15 June 2024 19:20:26 BST Alan Grimes wrote: >> A number of my softwarez requires the use of the arrow keys and can't >> use the numpad in edit mode to work around it. So who do I need to kill >> to get arrow keys to work in x11 again? > > I don't understand what is the "edit mode" you refer to. There > ought to be a NumLock key to enable/disable the numeric pad. Also, > your DE would have some config GUI for enabling/disabling the > numlock. I don't read Mr. Grimes postings, but if the question is how to control the numlock state. The Xorg config file option XkbOptions can be used to turn numlock on/off. There's also the numlockx command-line utility to do the same. Most of the options described here can be modified to turn numlock off instead of on: https://wiki.archlinux.org/title/Activating_numlock_on_bootup
[gentoo-user] Re: python 3.12 update
On 2024-06-05, Wols Lists wrote: > On 05/06/2024 13:12, Eli Schwartz wrote: >> Which I think is fine, if people want that, but not everyone does, so >> delaying the update altogether might be preferable to those people. > > Ie people like me who don't give a monkeys about python, and consider it > a necessary evil. I care quite a bit about Python and do a fair amount of developemnt in Python. Howver, I don't (at this point) give a monkey's about 3.12 vs 3.11, so I'm just going to delay the 3.12 upgrade on all my machines except for the one where I got stuck after step 1 of the now-famouse "3 step upgrade process". :) That one machine seems to be working fine (though it's still missing a few packages that I removed because the build failed with 3.12). So I'm going to leave it where it is (possibly reinstalling the missing packages as I run into the need for them). At some point, emerge -auvND will complain because some installed package doesn't support 3.11 any longer — and then I'll upgrade to 3.12. -- Grant
[gentoo-user] Re: python 3.12 update
On 2024-06-05, Arve Barsnes wrote: > On Wed, 5 Jun 2024 at 20:05, Grant Edwards wrote: >> What I found misleading (and tripped over) was the implication that >> the three step migration process outlined in the news item had a >> reasonable likelyhood of working for a large percentage of users. >> >> If the new items had warned that anybody using one of >> packages that won't work with 3.12 are going to have to stop after >> step 1 until those packages have been brought "up to date" so that >> they can build with 3.12. Had I known that, I wouldn't have tried the >> three step migration and would have simply postponed the upgrade. > > Maybe I've been using gentoo too long, but I never assumed these > three steps were a 'single upgrade' recipe, I assumed each step were > to be undertaken as the portage tree allowed. > > Reading the news item again, I see that no such thing is implied, > but perhaps it should have been? Otherwise, what is the point of > these steps as opposed to the automatic upgrade portage initially > attempts? I wondered about that too. But, I decided that whoever wrote the news item probably knew what they were talking about, and chose the "Safer upgrade procedure". As you said, there's no mention of waiting between steps for various packages to catch up: First, enable both Python 3.11 and Python 3.12, and then run the upgrade commands: */* PYTHON_TARGETS: -* python3_11 python3_12 */* PYTHON_SINGLE_TARGET: -* python3_11 Then switch PYTHON_SINGLE_TARGET and run the second batch of upgrades: */* PYTHON_TARGETS: -* python3_11 python3_12 */* PYTHON_SINGLE_TARGET: -* python3_12 Finally, switch to the final version and upgrade: */* PYTHON_TARGETS: -* python3_12 */* PYTHON_SINGLE_TARGET: -* python3_12 -- Grant
[gentoo-user] Re: python 3.12 update
On 2024-06-05, byte.size...@simplelogin.com wrote: > 2) Was anything really 'broken'? Most certainly no, going by the above > definition and the fact that the news item provided for a very clear > pathway to maintain compatibility that was essentially a two-line solution. I think that build failures caused by the 3.11 ⇒ 3.12 count as breakage. There was some, but not a lot of that. What I found misleading (and tripped over) was the implication that the three step migration process outlined in the news item had a reasonable likelyhood of working for a large percentage of users. If the new items had warned that anybody using one of packages that won't work with 3.12 are going to have to stop after step 1 until those packages have been brought "up to date" so that they can build with 3.12. Had I known that, I wouldn't have tried the three step migration and would have simply postponed the upgrade. -- Grant
[gentoo-user] Re: python 3.12 update
On 2024-06-05, Eli Schwartz wrote: > On 6/4/24 11:04 PM, Grant Edwards wrote: >> On 2024-06-04, Eli Schwartz wrote: >>> If a package claimed to support python 3.12 and nonetheless failed to >>> build with it, that's a bug in the package -- can you provide more details? >> >> IIRC, the first one was pycxx. The build faild during the config >> step. I uninstalled it (which meant I also had to uninstall pysvn, >> which I think means that meld is going to stop working). > > Thanks for the report. This was migrated to python 3.12 in > https://bugs.gentoo.org/929486 and indeed it builds fine for me. > > However, it is very obvious, *to me*, why it failed for you (even > without build logs). The package doesn't have a build dependency on the > dev-python/setuptools package, and python 3.12 removed a standard > library module (distutils) which is henceforth provided by this package. > And pycxx doesn't use pep517 mode either. > > It's a trivial fix. You would have seen a build error that looked > something like > > ModuleNotFoundError: No module named 'distutils' Ah, thanks for taking care of that. I would have put more time into figuring out that (and similar upgrade issues) and filing bug reports, but I'm right in the middle of migrating 400 projects from SVN to Gerrit/Git. It's going about as well as could be hoped, but... [A smarter person would not have allowed a major update like that to proceed in the middle of a big fire-drill.]
[gentoo-user] Re: python 3.12 update
On 2024-06-04, Eli Schwartz wrote: > On 6/4/24 4:58 PM, Grant Edwards wrote: >> On 2024-06-04, Eli Schwartz wrote: >> >>> Note that it's not a build failure -- it is an upgrade calculation >>> failure. It fails before upgrading any packages since it knows it can't >>> resolve the dependencies. >> >> I had plenty of both. > > > If a package claimed to support python 3.12 and nonetheless failed to > build with it, that's a bug in the package -- can you provide more details? IIRC, the first one was pycxx. The build faild during the config step. I uninstalled it (which meant I also had to uninstall pysvn, which I think means that meld is going to stop working). I started the emerge auvND again and there's a different build failure with a different package -- so I uninstall that. Start the emerge again, a third package fails to build. I honestly don't remember the 2nd and 3rd packages. At that point, I added these lines to package.use: */* PYTHON_TARGETS: -* python3_11 python3_12 */* PYTHON_SINGLE_TARGET: -* python3_11 Then the emerge ran through to the end with no problems. I'm going to disable 3.12 on all my other machines and wait until some package/app I need will no longer work with 3.11. -- Grant
[gentoo-user] Re: python 3.12 update
On 2024-06-04, Eli Schwartz wrote: > Note that it's not a build failure -- it is an upgrade calculation > failure. It fails before upgrading any packages since it knows it can't > resolve the dependencies. I had plenty of both. -- Grant
[gentoo-user] Re: python 3.12 update
On 2024-06-04, Dr Rainer Woitok wrote: > > Did nobody of ye all ever read news item 48, dated 2024-05-09? It laid > out a three-step approach which surely caused at least some packages to > be built twice or even three times, but it JUST WORKED (tm), at least > here. It only required creating and later updating a file in "/etc/por- > tage/package.use/" and then running > >emerge --depclean ; emerge -UD @world > > for each of the three steps. I don't see how (even in theory) steps 2 and 3 can work when you have packages installed that won't build with 3.12. I appear to have more than a few packages that won't build with 3.12 (some projects delcare that they don't support 3.12. Some don't declare that they don't support 3.12, the builds just fail when 3.12 is selected). -- Grant
[gentoo-user] Re: preparing /dev/sda1 for gentoo install x86 handbook
On 2024-06-01, Wol wrote: > I've got news for you, there are quite a few weirdos on the list, Hey! I resemble that remark. [Hmm. That's not as funny in print.]
[gentoo-user] Re: Getting started with a web server
On 2024-06-01, George Kettleborough wrote: > If you only want to build a static site (ie. just HTML, CSS, JS etc; no > server-side scripting) then you don't need to install and configure > something like Apache to test it out. You could just open the files you're > working on straight from the disk. Or if you want to test with server you > can use one of the super simple servers designed for testing. There is one > built in to Python. Simply run "python -m http.server" in your project > directory. For old-school dynamic content, busybox has a trivial web server that also supports cgi executables. CGI helpers written in something like Python is actually not a bad option for many dymanic things (e.g. generating SVG data plots using matplotlib). If you want to do PHP... Don't. Just don't. If you're looking at a hosting service that only supports PHP as a backend language, then pick another hosting service. As has already been mentioned, the big-time frameworks (Flask, Django, TurboGears, etc) all have their own test/developement server schemes. If you want generic server that's simpler than apache, lighttpd is a good option. -- Grant
[gentoo-user] Re: PCIe version 2, 3 etc and how to know which a card is.
On 2024-05-28, Dale wrote:
> Grant Edwards wrote:
>> On 2024-05-21, Dale wrote:
>>
>>>> Here's my udev rules file that defines my network interface names
>>>> for the machine I'm on at the moment:
>>>>
>>>> --/etc/udev/rules.d/70-my-persistent-net.rules---
>>>> SUBSYSTEM=="net", ACTION=="add", ATTR{address}=="2c:f0:5d:6f:10:af",
>>>> NAME="net0"
>>>> SUBSYSTEM=="net", ACTION=="add", ATTR{address}=="00:1b:21:b1:d1:e9",
>>>> NAME="net1"
>>>> -
> Got a little busy with my garden. Found my first zucchini yesterday.
> Ready to pick in a few days. Found some small tomatoes too. Anyway.
> Did manage to create this rule tho. This look reasonable? I'm thinking
> it should be named something else tho. It could clash with the usual
> name.
>
> # PCI device 0x11ab:0x4363 (Intel e1000e)
> #SUBSYSTEM=="net", ACTION=="add", DRIVERS=="?*",
> ATTR{address}=="68:05:ca:42:17:39",ATTR{dev_id}=="0x0", ATTR{type}=="1",
> KERNEL=="eth*", NAME="enp3s0"
Did my examples (with the MAC addresses and device names changed) not
work?
> I got the ATTR address from ifconfig. I'm not real sure on the other
> ATTR variables tho.
I don't use the other other ATTRs, ACTION, DRIVERS, or KERNEL and I
don't know why you added them, so I can't comment.
--
Grant
[gentoo-user] Re: Off Topic - UPnP servers
On 2024-05-24, Mark Knecht wrote: > The unit showed up today and was a breeze to set up and get running > at a basic level. The device requires an app on my phone. That sets of an alarm for me. > The app is available for Android and Apple but not available for the > Amazon Fire tablet. Good luck... I avoid products like that. There have been too many "smart" things in the past that required an app -- then the app stopped working two or three years later. The purchaser of the thing now has a useless lump, and has to start shopping for a replacement. -- Grant
[gentoo-user] Re: Off Topic - UPnP servers
On 2024-05-24, Mark Knecht wrote: > I'm a Plex user for video and have also ripped my CD > collection. Plex plays audio fine to TVs that have a Plex app but > apparently sometimes doesn't work well (as of yet untested by me) to > network streaming players. I never got the Plex app for Roku to work in a usable manner (and I think it eventually got discontinued?). The Plex app in Kodi has always worked fine for me (though I haven't used it for probably about a year). Plex also worked with other DLNA clients I've tried (Kodi, VLC). > While I don't know if the above will be a problem I've purchased a > network streaming player and will test it out over the weekend when it > arrives but if Plex doesn't work, or doesn't work well, I'd be interested to hear what player you got and how it works with Plex. -- Grant
[gentoo-user] Re: PCIe version 2, 3 etc and how to know which a card is.
On 2024-05-21, Dale wrote:
>> Here's my udev rules file that defines my network interface names
>> for the machine I'm on at the moment:
>>
>> --/etc/udev/rules.d/70-my-persistent-net.rules---
>> SUBSYSTEM=="net", ACTION=="add", ATTR{address}=="2c:f0:5d:6f:10:af",
>> NAME="net0"
>> SUBSYSTEM=="net", ACTION=="add", ATTR{address}=="00:1b:21:b1:d1:e9",
>> NAME="net1"
>> -
>
> Examples do help a lot. I do use the enp* naming scheme. My
> understanding, that is the "new" way.
The suffix for those enp* names comes from the PCI bus subsystem based
on bus number, slot number, etc. [Yes, slot number apparently does
change based on what PCIe cards are present. No, that doesn't make
sense to me either]
> Based on your config, I would need to change the NAME= to enp* names
> and that would correct that.
I suppose you could, but I would not use enp* names. Those could
conflict with the autogenerated names.
> Where you have ATTR, is that a quote or did you edit to remove a
> number, MAC address, IP or something?
What I posted is exactly what's in the file
(without the --- delimiters).
Here's more documentation:
https://wiki.gentoo.org/wiki/Udev
https://wiki.archlinux.org/title/udev
https://wiki.archlinux.org/title/Network_configuration#Change_interface_name
[The arch Wiki is always a good fallback if the Gentoo manual/Wiki
don't have what you're looking for.]
> If it is one of those, where do I find that info? I checked
> ifconfig and didn't see a MAC address. I also checked lspci -v.
> I'm not sure where you get the needed info from. BTW, right now,
> I'm on my main rig.
The only thing you need to change from my example would be the mac
address(es) (e.g. 2c:f0:5d:6f:10:af) and the names (e.g. net0).
> I have the package net-misc/networkmanager installed. Most likely
> pulled in by something else. Could I use it to configure this?
Possibly, I don't use networkmanager and don't know how it works on
Gentoo. I use the default Gentoo netifrc scheme
https://wiki.gentoo.org/wiki/Netifrc.
> I also have KDE installed on the NAS box, it is also a backup rig in
> case my main rig dies. It may have a GUI that I could use. I'm not
> opposed to the command line way tho. Biggest thing, copy and paste
> would be nice.
I don't know much of anything about KDE.
--
Grant
[gentoo-user] Re: PCIe version 2, 3 etc and how to know which a card is.
On 2024-05-21, Dale wrote:
>> If you want consisent network device names (even when you change
>> hardware), you need to either
>>
>> 1. create udev rules that assign device names based on MAC addresses.
>>
>> 2. use a network configuration subsystem that assigns device names
>> and configurations based on MAC addresses.
>
> Do you, or someone else, know of a good howto on how to use MAC
> addresses like that? Given this thing is usually remotely accessed, I
> really need it to be consistent with or without the card. Maybe you
> have a bookmarked link saved somewhere. I'm on openrc to. I'll google
> around but you, or someone else here, may have a really good and simple
> howto link.
The udev way is probably the most universal. Some distros will create
udev rules automagically so that network interface names persist over
hardware changes, but Gentoo doesn't. Here's my udev rules file that
defines my network interface names for the machine I'm on at the moment:
--/etc/udev/rules.d/70-my-persistent-net.rules---
SUBSYSTEM=="net", ACTION=="add", ATTR{address}=="2c:f0:5d:6f:10:af", NAME="net0"
SUBSYSTEM=="net", ACTION=="add", ATTR{address}=="00:1b:21:b1:d1:e9", NAME="net1"
-
I used to use "ethN" instead of "netN", but those names are used
internally by the kernel during startup, and people were warned not to
use them in udev rules because of certain race conditions that might
happen. I never ran into problems using "ethN" names, but eventually
decided not to push my luck.
The network configuration route depends on what network configuration
(and possibly init) system you use. I know how to do it that way on
Ubunutu (which is systemd based) using netplan...
--
Grant
[gentoo-user] Re: PCIe version 2, 3 etc and how to know which a card is.
On 2024-05-21, Dale wrote: > So they both show up. When I try to start the network, it says: > > ERROR: Interface enp3s0 does not exist. Are you sure the network interface name hasn't changed? What does "ifconfig -a" or "ip addr" show? After booting up, what does "dmesg | grep enp" show? > Ensure that you have loaded the correct kernel module for your hardware. > > > I find that odd since it obviously sees the card. It's in the list > above after all. So, it sees the card but can't see it. 0_o Odd. Identifying the presense of a PCI card and creating the device by which is is accessed are two different things. > I tried different slots for the SATA card and they all do the same > thing. Wouldn't each slot have a different interrupt? No. If cards are using legacy interrupts (most do) there are four interrupts (A,B,C,D) that are shared by all cards (just like there always were). Newer cards and motherboards can use something called MSI or MSI-X interrupts that aren't shared, but in my experience the use of those isn't very widespread. > It was at this point, I checked your suggestion. I looked and noticed > that the network card was now at slot 4 not slot 3 like it used to be. > So, I created a new link to slot 4. The network came up. So, > basically, it changed names as you suggested. I thought the purpose of > the enp* names was that they are consistent. They are consistent through reboots. They are not consistent if you change hardware. > Adding or removing cards wouldn't change the names of cards, like > network cards. Yes, it can. > It seems, in this case at least, the names can change. Any way to > make adding the card not change this?? I tend to not have a monitor > or keyboard connected to this rig. If you want consisent network device names (even when you change hardware), you need to either 1. create udev rules that assign device names based on MAC addresses. 2. use a network configuration subsystem that assigns device names and configurations based on MAC addresses. -- Grant
[gentoo-user] Re: PCIe version 2, 3 etc and how to know which a card is.
On 2024-05-20, Dale wrote: A 3.0 card is supposed to work fine in a 2.0 slot. > You, or anyone, have any idea why that card would kill my network? > I suspect the card itself is fine. It did see the drive. I just > need the internet to work since it may be used in a NAS rig. Is it causing the network interface to not show up at all in lspci? Is it causing the network device name to change? Or is the network interface still detected, still named the same, and just doesn't send/receive packets? It could be some sort of interrupt sharing problem. Even with PCI express, cards still sometimes have to share interrupts. Intel/IBM made that bad decision 45 years ago, and we're still suffering because of it. If that the problem, sometimes you can avoid it by physically rearranging the cards. The later PCI hosts/boards finally came up with a way to avoid it, but a lot of cards still don't support that. -- Grant
[gentoo-user] Re: Grub, gpt partitions and BIOS, not uefi thing.
On 2024-05-15, Wols Lists wrote: > On 15/05/2024 11:40, Peter Humphrey wrote: >> I think whoever named grub had delusions of grandeur. 🙂 Anyway, I never let >> it near my systems. > > I liked lilo. And then it disappeared :-( > > Grub isn't that bad - it's just that insists on trying to do everything > itself - and if you've got at all a strange setup it makes a complete > hash of it. Grub2 is a bit overblown, but it's quite usable as long as you stick to a manually generated grub.cfg file and stay away from the auto-magical disk-probing configuration script world-domination scheme. -- Grant
[gentoo-user] Re: Encrypted drives, password generation and management howto, guide.
On 2024-05-15, Michael wrote: > On Wednesday, 15 May 2024 15:37:22 BST Grant Edwards wrote: >> On 2024-05-15, Michael wrote: > >> > The Clipboard may be stored in RAM or cache of any applications >> > which use this method. >> >> AFAICT, the clipboard contents is stored in the X server. When you >> cut/copy something, the application sends that something to the X >> server where it's stored. When that application exits, the clipboard >> contents are still there in the X server, and can still be requested >> by other applications who want to do a "paste". > > What you write makes sense. I got curious, and did some more Googling. It looks like the clipboard contents only survive application exit if the application explicitly tells the server it wants the clipboard contents to persist. But, AFIACT, that's what all apps do. > I am not sure what happens in Wayland, where application windows are > supposed to be isolated. I try not to think about Wayland and dread the day when I'm forced to switch. :) It's taken me 40 years to figure out X (most-sort-of)... > I recall in earlier days the Primary selection would not work > between windows, which was rather frustrating. I think at present > the Plasma desktop clipboard application acts as a mediator, > probably engaging Xwayland - but I am not sure. > > There are quite a few settings in Plasma's clipboard application to > configure interoperability between Primary & Clipboard selection and > can be set to save the Primary selection in the Clipboard section > and its history if so desired. > > With my current settings I can middle click to paste a Primary > selection into Konsole, but Shift+Insert which works with Xterm & > friends does not work with Konsole. There probably should have been a section on cutbuffers, selections, and clipboards in the X11 section of the Unix Hater's Handbook. which I highly recommend, BTW: https://web.mit.edu/~simsong/www/ugh.pdf https://en.wikipedia.org/wiki/The_UNIX-HATERS_Handbook
[gentoo-user] Re: Graphics configuration for a Ryzen 7 7700X chip and water cooling.
On 2024-05-15, Alan Mackenzie wrote: > But in the doc on wiki.gentoo.org, I can't find any mention of inbuilt > graphics; all references are to graphics _cards_. Does Gentoo support > my intended processor's graphics, Technically, no. Gentoo doesn't. However, the Linux kernel, Xorg, and Mesa do. You'll need "recent" versions of those. According to this article: https://www.phoronix.com/review/amd-ryzen7-7700x You'll need kernel 5.18 and Mesa 22 plus recent firmware. That article was almost 2 years old, so I'd be surprised if all those are not stable in Gentoo by now. -- Grant
[gentoo-user] Re: Encrypted drives, password generation and management howto, guide.
On 2024-05-15, Michael wrote: > As far as I know the Primary selection is not stored anywhere - > other than within the application's memory space where the range of > characters have been selected. The xserver will call for this when > you middle click to paste it on another application's window. Right. When you highlight some text, the application asserts ownership of the primary selection, but no contents of the selection are transferred to the X server. So, the X server knows who owns the selection, but it doesn't actually store the contents anywhere. If you middle-click on a window, the X server will make a call to the owner of selection to get the selection contents and then provide that contents to the active window. When process (X client) that owns the selection exits, the selection becomes "empty" (unavailable). > The Clipboard may be stored in RAM or cache of any applications > which use this method. AFAICT, the clipboard contents is stored in the X server. When you cut/copy something, the application sends that something to the X server where it's stored. When that application exits, the clipboard contents are still there in the X server, and can still be requested by other applications who want to do a "paste". With the usual behavior, the selection and clipboard sort of overlap: When you highlight something the application asserts ownership of the primary selection, but nothing is transferred to the X server. If you then do a "copy", the application will send that highlighted text to the clipbard. If you haven't selected anything else, now you can either middle-click or paste, and you'll get the same thing. If you exit the app, then middle-click will produce nothing because there is no selection owner. But, paste will still get the "copied" data from the X server. That said, something doesn't have to be selected (in the X11 sense) to be copied into the clipboard -- but that's how most applications work first you select (in the X11 sense) something then you copy it to the clipboard. -- Grant
[gentoo-user] Re: Encrypted drives, password generation and management howto, guide.
On 2024-05-15, Dale wrote: >> Or just select some empty space in an application, to overwrite your >> previous >> selection. > > Well, since it works, something is acting as a clipboard. It's part of the X server. Same for the two selections. > It doesn't seem to be xclip in my case. Anyway, that's what I been > doing is highlighting something else and that makes it paste the new > highlighted info instead of previous info. I have no idea if those > entries are stored somewhere or when gone, they gone. I'm hoping > they are gone.
[gentoo-user] Re: Encrypted drives, password generation and management howto, guide.
On 2024-05-15, Dale wrote: > I thought that too. I highlighted some text in a Konsole and then > looked in the KDE clipboard, what I highlighted was not there. > > It wasn't there after I pasted it either. It goes to a clipboard > somewhere but it appears it only remembers one entry then forgets > when you highlight something else. You're conflating to different but related things. In X, the selection and the clipboard are two different "places". When you click-drag to highlight text, that goes into the selection. In X, there are actually two different selections: the primary and the secondary. By default highlighted text goes into the primary selection. Middle-clicking shoves the contents of the primary selection into stdin for whatever window is selected. When you do "cut" or "copy" something, it goes into the clipboard. When you "paste" it comes from the clipboard. xclip can access (read or write) all three (primary selection, secondary selection, and clipboard). https://unix.stackexchange.com/questions/139191/whats-the-difference-between-primary-selection-and-clipboard-buffer https://superuser.com/questions/90257/what-is-the-difference-between-the-x-clipboards https://www.reddit.com/r/linux/comments/mgr0v/til_x11_has_three_clipboards/ -- Grant
[gentoo-user] Re: Grub, gpt partitions and BIOS, not uefi thing.
On 2024-05-01, Dale wrote: > Grant Edwards wrote: > >> The partition type code for 'swap' is wrong -- it should be >> 8200. According to the gdisk help info Linux /home is supposed to be >> 8302, but I've always used the same generic "Linux filesystem" type >> for both /home and root. >> >> Is the 'boot' partition for future possible UEFI use, for Linux /boot, >> or both? [I've never used a separate partition for Linux /boot, I >> just use a /boot directory on the root FS.] > > I noticed the other day that some new ones was added. I always leave it > as 8300 and it works. It even works for swap. I dunno. If you have an entry in /etc/fstab for swap, it might not matter if the partition type is set to 'Linux swap' or not. I always set the swap parition type to 'Linux swap', and then it doesn't seem to matter if there's a swap entry in the fstab or not. > The /boot is where kernels and init thingys go. Keep in mind, this is > on a old rig that has no idea what UEFI is. When I build my new rig > later, I'll do a install from scratch anyway. Also, it will go on a SSD. OK, so 'boot' is for the Linux /boot directory. I was just curious since I had never used one. > I mostly want to post so that a person can see the layout. Really, the > first one is what a person wanting to use GPT on a old BIOS system needs > to see. After that, they can do partitions anyway they want. Right.
[gentoo-user] Re: Grub, gpt partitions and BIOS, not uefi thing.
On 2024-05-01, Dale wrote: > OK. One last update in case someone googles and runs up on this > thread. I'm using gdisk to display this, because I think it will do > better in email. If I use cgdisk, it is wider and will wrap more. > This is what the partition table looks like for GPT, old BIOS and no > uefi thingy. Just a straight forward and simple old school setup. > Once the first one is done, the rest can be anything. > > > Number Start (sector) End (sector) Size Code Name > 1 2048 10239 4.0 MiB EF02 BIOS-boot > 2 10240 4204543 2.0 GiB 8300 boot > 3 4204544 12593151 4.0 GiB 8300 swap > 4 12593152 327165951 150.0 GiB 8300 root > 5 327165952 625141759 142.1 GiB 8300 home The partition type code for 'swap' is wrong -- it should be 8200. According to the gdisk help info Linux /home is supposed to be 8302, but I've always used the same generic "Linux filesystem" type for both /home and root. Is the 'boot' partition for future possible UEFI use, for Linux /boot, or both? [I've never used a separate partition for Linux /boot, I just use a /boot directory on the root FS.] -- Grant
[gentoo-user] Re: Grub, gpt partitions and BIOS, not uefi thing.
On 2024-04-28, Grant Edwards wrote: > With DOS disk lables, Grub uses empty space between the boot sector > and the first partition as a location to store it's core image file. > That empty space does not exist when using GPT disk label. When using > a GPT disk label, Grub requires that you need to create a "BIOS Boot" > or "Grub Boot" partition so that Grub has somwhere to store it's core > image[1]. And it bears repeating that the bios/grub boot partition only needs to be 1 or 2MB in size, is _not_ formatted with a filesystem, and is _not_ the same as either 1) The "boot" directory where the kernel images and grubs other files are installed within a Linux filesystem. [Which you still need when booting in Legacy/BIOS mode.] or 2) The UEFI partition that's formated with a FAT filesystem and used in UEFI boot mode [which you don't need when booting in Legacy/BIOS mode.]
[gentoo-user] Re: Grub, gpt partitions and BIOS, not uefi thing.
On 2024-04-27, Michael wrote: > On Saturday, 27 April 2024 17:53:25 BST Dale wrote: >> Howdy, >> >> I'm installing Gentoo on another old box. To be consistent I like >> to use cgdisk, GPT I think it is called, to partition all my >> drives, regardless of size. > > GPT is the partition table structure, which is more advanced than > the old DOS partition table structure. > >> Thing is, Grub works differently with GPT than it does with the old >> DOS or whatever it is called, like fdisk does in the old days. > > GRUB works the same, but the disk/partition table structure is different. No, grub doesn't work the with GPT disk labels as it did with DOS disk labels. With DOS disk lables, Grub uses empty space between the boot sector and the first partition as a location to store it's core image file. That empty space does not exist when using GPT disk label. When using a GPT disk label, Grub requires that you need to create a "BIOS Boot" or "Grub Boot" partition so that Grub has somwhere to store it's core image[1]. https://wiki.gentoo.org/wiki/GRUB#BIOS_with_GPT https://wiki.archlinux.org/title/GRUB#BIOS_systems [1] There is an alternative installation method where Grub will record the disk block numbers occupied by the core image files as they reside in the normal filesystem. That's extra work to maintain and might not be reliable for some filesystem types, so it's not recommended.
[gentoo-user] Re: NAS box and switching from Phenom II X6 1090T to FX-6300
On 2024-04-17, Frank Steinmetzger wrote: >> If you don't play games, then definitely get integrated graphics. >> Even if the CPU costs a tiny bit more, it will give you a free empty >> 16x PCIe slot at whatever speed the CPU supports (v5 in this case - >> which is as good as you can get right now). > > Not to mention a cut in power draw. And one fewer fan to listen to. [I was also pretty annoyed with NVidia when they stopped offering fanless Quadro boards. I'm a big fan of fanless]
[gentoo-user] Re: Slightly corrupted file systems when resuming from hibernation
On 2024-04-17, Dale wrote: > I still use Nvidia and use nvidia drivers. I to run into problems > on occasion with drivers and kernels. When you switched from > Nvidia, what did you switch too? Do you still use drivers you > install or kernel drivers? All in-tree kernel drivers for integrated GPUs: * Intel UHD Graphics 620 * Intel HD Graphics 4000 * Intel Xeon E3-1200 * AMD Picasso Radeon Vega After I had to recycle my second perfectly functional NVidia card simply because NVidia stopped driver support, I got fed up. I tried the open-source nvidia drivers for those cards, but could never get multiple screens to work. > How well does the video system work? In other words, plenty fast > enough for what you do. They're all fast enough for what I do (no heavy gaming, but I do play with an RC flight simulator). All will drive at least two digital monitors. The last machine that had an NVidia card removed is also the oldest of the machines (Gigabyte Z77X-UD5H Intel i5-3570K w/ HD 4000 graphics), and it's happily driving three monitors (1 HDMI, 1 DVI, 1 DP). When running the flight-sim, the newest of them (the AMD/Radeon) is noticeably smoother and runs at higher frame rates than the older Intel GPUs. I didn't really have any complaints about the older ones, but I don't expect a real gamer would have been satisfied with the Intel ones. > I don't do any sort of heavy gaming. Since I have a nice game on my > cell phone now, I play it almost all the time. I can't recall > playing a game of solitaire on my computer in a long while. My > biggest thing, two video ports, one for monitor and one for TV. > Most TV videos aren't very high def but some are 1080P. That's all > my TV can handle. They all seem to handle HD video playback just fine. How many and what type of monitors can be driven is very much dependent on the motherboard. -- Grant
[gentoo-user] Re: Slightly corrupted file systems when resuming from hibernation
On 2024-04-17, Dr Rainer Woitok wrote: > Grant, > > On Wednesday, 2024-04-17 14:11:21 -, you wrote: > >> ... >> If what you want is access to all upstream longeterm kernel versions, >> then you should be using sys-kernel/vanilla-sources. > > I was not aware of this package. Excatly what could come in handy, if > everything else fails. Thank you for that pointer :-) Just be aware that gentoo-sources contains an "extra" set of gentoo-specific patches. So version x.y.z of gentoo-sources isn't identical to version x.y.z of vanilla-sources. https://dev.gentoo.org/~mpagano/genpatches/ -- Grant
[gentoo-user] Re: Slightly corrupted file systems when resuming from hibernation
On 2024-04-17, Dr Rainer Woitok wrote: > Grant, > > On Tuesday, 2024-04-16 19:26:25 -, you wrote: > >> ... >> That means that all gentoo-sources stable kernels are "longterm" >> kernel versions on kernel.org. It does not mean that all "longterm" >> kernel versions from kernel.org are available as "stable" in >> gentoo-sources. >> >> It is a statement that "gentoo-sources stable" is a subset of >> "kernel.org longterm". > > This sort of deteriorates into a debate about words rather than meanings I'm sorry to have caused "deterioration" by trying to explain the statement you said a) you didn't understand and b) was contridicted by the contents of the gentoo portage tree. The statement was not contridicted by what you pointed out. > without explaining HOW LONG such a series of related kernels are > main- tained and provided. That was not the subject of the statement you claimed was wrong which I then tried to explain. The gentoo-sources versions of upstream "longterm" kernels are maintained and provided for as long as the volunteers who do the work maintain and provide them. > After all, "longterm" or "LTS" suggest that these lines of > developement are less short-lived than others. They are. Upstream longterm kernel versions get updates for several years longer than versions that are not longterm. > To give an ex- ample: the oldest "longterm" kernels listed on > "kernel.org" are 4.19.*, 5.4.* and 5.10.*. Of these only 5.10.* is > still available from Gentoo. You should certainly demand that all of the money you paid for gentoo-sources be refunded. It takes work to maintain gentoo-sources ebuilds. I'm sure if you volunteered to maintain ebuilds for the older longterm kernels, the help would be happily accepted. Free clue: It's _hard_work_ to support old verions of things (especially kernels). They usually won't build with recent tools and won't run on recent hardware or with recent versions of other software. You often have to keep around entire virtual machines that have old tools and utilities. If what you want is access to all upstream longeterm kernel versions, then you should be using sys-kernel/vanilla-sources. > So what time span are we talking about when we say "LTS Gentoo > kernel"? We don't say that. "LTS" and "longterm" are not Gentoo designations, they are kernel.org designations. Gentoo has "stable" and "testing". Only upstream "longterm" kernel versions get marked as "stable" in gentoo-sources. They are then supported for as long as somebody supports them. > Roughly four, three or two years? And why is the support provided > by Gentoo significantly shorter than that by "kernel.org"? Because you're not helping with the support? -- Grant
[gentoo-user] Re: Slightly corrupted file systems when resuming from hibernation
On 2024-04-17, Michael wrote: >> > But, to get back to the beginning of this discussion: if there is a >> > risk that my aging hardware possibly can less and less cope with >> > newer and newer kernels, should I put something like >> > >> >>=sys-kernel/gentoo-sources-6.7.0 >> > >> > into file "package.mask" to stay with "longterm" 6.6.* kernels? >> >> Yes: if you want to avoid getting upgraded to 6.8 when it gets >> kernel.org "longterm" status and gentoo-sources "stable" status, then >> a statement like that in in package.mask will keep you on >> gentoo-sources 6.6 kernels (which are "longterm" on kernel.org). > I am not sure the assumption "... aging hardware possibly can less and less > cope with newer and newer kernels" is correct. Good point. My "yes" was in response to a question of the form "if X is true, should I do Y". I did not attempt to address the likelyhood that X was actually ture, only whether Y was appropriate if X was true. > As already mentioned newer kernels have both security and bug fixes. > As long as you stick with stable gentoo-sources you'll have these in > your system. Later kernels also come with additional kernel drivers > for new(er) hardware. You may not need/want these drivers if you do > not run the latest hardware. Using 'make oldconfig' allows you to > exclude such new drivers, but include new security options and/or > functionality as desired. > > It can happen for new code to introduce some software regression. The usual worries with running newer kernels on older hardware are: 1) Performance degradation when upgrading kernels on older hardware. On one embedded project I work with we're still running a 2.6 kernel because network throughput drops by 25-30% when we upgrade to 3.x kernels. There's nothing "wrong" with those 3.x kernels, they're just bigger and significantly slower. [Even when built with a config that's as identical to the 2.6 kernels as possible.] 2) Lack of support for old hardware when running a newer kernels. I used to run into this when running nvidia-drivers. Gentoo-sources would mark a new kernel stable, but my video board would not be supported by nvidia-drivers versions that were supported for that new stable kernel. I would mask newer kernels until and run older "longterm" kernels as long as I could. I would evenually be forced to buy a new video card. After going through that cycle a couple times, I swore off NVidia video cards and life's been much eaiser since.
[gentoo-user] Re: Slightly corrupted file systems when resuming from hibernation
On 2024-04-16, Dr Rainer Woitok wrote: > Arve, > > On Tuesday, 2024-04-16 15:53:48 +0200, you wrote: > >> ... >> Only LTS kernels get stabilised, so this information is readily available. > > I'm sure I don't understand this: According to "https://www.kernel.org/"; > kernel 6.6.27 is "longterm", but according to "eix" the most recent > 6.6.* kernels are 6.6.22 and 6.6.23 which both are non-stable (well, I > ran my last "sync" immediately before the profile upgrade, so this might > not be current). I'm still using stable kernel 6.6.13 as my backup ker- > nel, but this kernel is no longer provided by Gentoo. So, what precise- > ly does LTS or "longterm" mean? That means that all gentoo-sources stable kernels are "longterm" kernel versions on kernel.org. It does not mean that all "longterm" kernel versions from kernel.org are available as "stable" in gentoo-sources. It is a statement that "gentoo-sources stable" is a subset of "kernel.org longterm". It is not a statement that the two sets are identical. In other words: "ONLY LTS kernels get stabilized." is a different statement from "ALL LTS kernels get stabilized." The former is true. The latter is not. > But, to get back to the beginning of this discussion: if there is a > risk that my aging hardware possibly can less and less cope with > newer and newer kernels, should I put something like > >>=sys-kernel/gentoo-sources-6.7.0 > > into file "package.mask" to stay with "longterm" 6.6.* kernels? Yes: if you want to avoid getting upgraded to 6.8 when it gets kernel.org "longterm" status and gentoo-sources "stable" status, then a statement like that in in package.mask will keep you on gentoo-sources 6.6 kernels (which are "longterm" on kernel.org). Again: not all longterm 6.6.x kernel versions get marked as "stable" for gentoo-sources. If you have not enabled the testing keyword for gentoo-sources, then you'll only get the 6.6.x kernel versions that the gentoo-sources maintainers have declared as "stable". -- Grant
[gentoo-user] Re: Slightly corrupted file systems when resuming from hibernation
On 2024-04-16, Dale wrote: > I've never understood what is supported long term either. I use > gentoo-sources. I've never figured out just how to pick a kernel that > is supposed to be stable for the larger version. In other words, only > security and bug fixes, no new hardware. Right now, 6.8.5 is the > highest version in the tree here but there are more versions of it to > come. So, I tend to go back to 6.7.X and pick the highest version of > that. The first two digits used to mean something but I think that > changed a long time ago. Any gentoo-soruces ebuild that's "stable" is an upstream LTS kernel. The 6.8 version of gentoo-sources are all "testing". They're "stable" on kernel.org, but theyre _not_ LTS. I think I read that 6.8 is expected to become the next LTS, but I don't really pay attention. > I try to avoid the absolute latest because my video drivers tend to lag > behind a little. They won't emerge for anything very new sometimes. > That's why I go back a little as described above. Thing is, I have no > idea if that is the right way or if it really even matters if I pick > 6.8.1 over 6.7.12 or vice versa. Neither are stable in Gentoo. Neither are longterm on kernel.org. 6.8 is stable on kernel.org. 6.7 is EOL on kernel.org. I would only choose 6.7 as a last resort. I would only choose 6.8 if the latest longterm (6.6) won't work. > I wish they were clearly marked somehow myself. Something in the name > that shows it is stable. Given I rarely have problems with kernels, > maybe none of this matters. Thing is, I plan to build a new rig soon. > Might help then. Maybe. Look at https://packages.gentoo.org/packages/sys-kernel/gentoo-sources The ones in green are the kernel.org "longterm" supported kernel versions that are stable in Gentoo. Here you can see which ones are lonterm, stable, mainline, and EOL upstream: https://kernel.org/ I would never run something that's not longterm unless there's a specific reason you have to choose something else. If you have to pick something that's not longterm, go with "stable" and not EOL if you can.
[gentoo-user] Re: Slightly corrupted file systems when resuming from hibernation
On 2024-04-16, Arve Barsnes wrote: > On Tue, 16 Apr 2024 at 15:29, Dr Rainer Woitok > wrote: >> > My understanding is the gentoo-sources kernels are aligned with the LTS >> > upstream releases. >> >> Right, they use the same version numbers. But you can't see from just >> looking at the available "gentoo-sources" which one is LTS and which one >> is not. You have to consult "https://www.kernel.org/"; to get this in- >> formation. > > Only LTS kernels get stabilised, so this information is readily available. "Stablized" as in the corresponding gentoo-sources ebuild is marked as stable. [Not to be confused with Linux "stable" kernels -- not all of which end up with LTS status.] Getnoo-sources also includes "stable" but not "LTS" Linux kernels, but the gentoo-sources ebuild for those is always "testing". IOW, if you install gentoo-sources, and don't keyword it to allow "testing" ebuilds, then you won't get anything other than LTS kernel sources.
[gentoo-user] Re: How to synchronise between 2 locations
On 2024-03-27, Mark Knecht wrote: > On Wed, Mar 27, 2024 at 11:59 AM J. Roeleveld wrote: >> I am looking for a way to synchronise a filesystem between 2 >> servers. Changes can occur on both sides which means I need to >> have it synchronise in both directions. > > How synchronized? For instance, does it need to handle identicals where > a file is on both sides but has been moved? Does it need to handle the case where the same file is modified independently on both sides?
[gentoo-user] Re: New profiles 23.0
On 2024-03-26, Walter Dnes wrote: > On Tue, Mar 26, 2024 at 04:21:23PM +, Michael wrote >> On Tuesday, 26 March 2024 15:21:32 GMT Walter Dnes wrote: >> > I assume my system is already "merged-usr". Current profile... >> > >> > [12] default/linux/amd64/17.1/no-multilib (exp) * >> > [...] > > Thanks for the clarification. So my system is considered > "split-usr", regardless of everything being on one partition. Yes. "split user" means that /bin and /usr/bin are two different directories. Likewise for /lib and /usr/lib, and so on... It doesn't matter that /foo and /usr/foo directories are in the same filesystem. > I got the news item when I ran "emerge --sync". My understanding is > that step 1 in the news item says "Please also update your system > fully and depclean before proceeding" so I should update world > first. Yes. And depclean. > Since I'm currently on profile > "default/linux/amd64/17.1/no-multilib" then I should migrate to > profile "default/linux/amd64/23.0/split-usr/no-multilib" as per the > news item. Yes -- If you're using OpenRC. I assume you are not using systemd since your old profile wasn't a systemd profile [In theory you could be running systemd on a non-systemd profile, but it's not common.] If you are running systemd, you should first update to the "merged-usr" flavor of your current profile. There's a detailed table at https://wiki.gentoo.org/wiki/Project:Toolchain/23.0_update_table It shows exactly what new profile corresponds to what old profile. > Migrating from there to "default/linux/amd64/23.0/no-multilib" is a > separate process as per the Gentoo wiki. Is my understanding > correct? Yes: https://wiki.gentoo.org/wiki/Merge-usr
[gentoo-user] Re: New profiles 23.0
On 2024-03-26, Walter Dnes wrote: > I'm AMD64 stable OpenRC. I got tired of dicking around resizing > partitions years ago, so I have all data and binaries in one honking > big partition. Also separate partitions for UEFI and swap. I assume > my system is already "merged-usr". Current profile... > > [12] default/linux/amd64/17.1/no-multilib (exp) * No, it is not merged user. If it were, the profile would say above would say "/merged-usr" at the end. > I just ran "emerge --sync" and got the profile news item. So do I > update world and then update profile? emerge -pv has 3 interesting > lines... Follow the instructions in the new item.
[gentoo-user] Re: Stage-3 and profile 23.x
On 2024-03-25, Michael wrote: > On Monday, 25 March 2024 21:48:24 GMT Peter Humphrey wrote: >> On Monday, 25 March 2024 16:52:19 GMT Michael wrote: >> >>> The default OpenRC installation now assumes a merged-usr fs structure - >>> therefore make sure you select the appropriate profile in a new >>> installation. >> >> I was wondering about that. Now that we have 23.0 in place, are we meant to >> change to merged-usr? Should I run the eponymous script? > > You can, if you want to. I've installed sys-apps/merge-usr and ran it on my > OpenRC system, after I completed the migration to profile 23.0. It didn't > take any time at all: > > https://wiki.gentoo.org/wiki/Merge-usr I'm in the process of switching two machines fom 17.1 to 23.0/split-usr [emerge @world will probably take overnight.] I've read the merge-usr page, and it looks pretty simple. But, I'm going to run 23.0 split-usr for a few weeks first -- just to make sure that the new profile hasn't broken anything. If you run OpenRC, it doesn't sound like there's any real reason you need to do the merge any time soon. If you run systemd, there's some version number cutoff where it will refuse to work on a split-usr install (IIRC). After all, the systgemd motto is "all your computer are belong to us!" -- Grant
[gentoo-user] Re: Terminal emulator to replace Konsole
On 2024-03-23, Mickaël Bucas wrote: > I think it's not a terminal emulator feature, but rather a shell > feature. > > Some terminal programs are designed to interact with the mouse, but > bash command line, based on readline, doesn't react to mouse clicks. Agreed. > I've tried Midnight Commander in Konsole and xterm, and it actually > moves the cursor to the click position in its own command line ! > > Maybe there's an extension or set of parameters for bash or other > shells to handle mouse clicks, but I'm not aware of it. I vaguely recall that there was some sort zsh hack/addon/module that added extra mouse functionality, but I don't recall the details (I was never a zsh user). -- Grant
[gentoo-user] Re: gentoo-sources 5.15.151 breaks amdgpu support?
On 2024-03-11, Grant Edwards wrote: > I upgraded gentoo-sources from 5.15.147 to 5.15.151 this morning and > amdgpu support is now borked on my system with an AMD Ryzen 5 3400G > with Radeon Vega Graphics. > > Everything worked fine with 5.15.147, but when 5.15.151 (built with > same .config via "make oldconfig") boots there's always a kernel oops, > and video output goes blank. > > I suppose maybe it's time to see if 6.1 works... Gentoo-sources 6.1 seems to work fine. I had masked it when it first went stable because it wouldn't boot back then. Whatever was wrong with it then seems to have been fixed. Hopefully it won't catch whatever malady is affecting 5.15.151. -- Grant
[gentoo-user] gentoo-sources 5.15.151 breaks amdgpu support?
I upgraded gentoo-sources from 5.15.147 to 5.15.151 this morning and amdgpu support is now borked on my system with an AMD Ryzen 5 3400G with Radeon Vega Graphics. Everything worked fine with 5.15.147, but when 5.15.151 (built with same .config via "make oldconfig") boots there's always a kernel oops, and video output goes blank. I suppose maybe it's time to see if 6.1 works... [2.888353] [drm] amdgpu kernel modesetting enabled. [2.896597] amdgpu: Topology: Add APU node [0x0:0x0] [2.896736] [drm] initializing kernel modesetting (RAVEN 0x1002:0x15D8 0x1462:0x7C02 0xC8). [2.896742] amdgpu :2a:00.0: amdgpu: Trusted Memory Zone (TMZ) feature enabled [2.896751] [drm] register mmio base: 0xFCA0 [2.896752] [drm] register mmio size: 524288 [2.896762] [drm] add ip block number 0 [2.896764] [drm] add ip block number 1 [2.896766] [drm] add ip block number 2 [2.896767] [drm] add ip block number 3 [2.896769] [drm] add ip block number 4 [2.896770] [drm] add ip block number 5 [2.896772] [drm] add ip block number 6 [2.896774] [drm] add ip block number 7 [2.896775] [drm] add ip block number 8 [2.896779] Loading firmware: amdgpu/picasso_gpu_info.bin [2.919851] [drm] BIOS signature incorrect 5b 7 [2.919872] amdgpu :2a:00.0: amdgpu: Fetched VBIOS from ROM BAR [2.919874] amdgpu: ATOM BIOS: 113-PICASSO-115 [2.919883] Loading firmware: amdgpu/picasso_sdma.bin [2.920090] [drm] VCN decode is enabled in VM mode [2.920091] [drm] VCN encode is enabled in VM mode [2.920092] [drm] JPEG decode is enabled in VM mode [2.920111] amdgpu :2a:00.0: vgaarb: deactivate vga console [2.920841] Console: switching to colour dummy device 80x25 [2.920885] [drm] vm size is 262144 GB, 4 levels, block size is 9-bit, fragment size is 9-bit [2.920892] amdgpu :2a:00.0: amdgpu: VRAM: 2048M 0x00F4 - 0x00F47FFF (2048M used) [2.920895] amdgpu :2a:00.0: amdgpu: GART: 1024M 0x - 0x3FFF [2.920898] amdgpu :2a:00.0: amdgpu: AGP: 267419648M 0x00F8 - 0x [2.920904] [drm] Detected VRAM RAM=2048M, BAR=2048M [2.920906] [drm] RAM width 128bits DDR4 [2.921005] [drm] amdgpu: 2048M of VRAM memory ready [2.921008] [drm] amdgpu: 3072M of GTT memory ready. [2.921012] [drm] GART: num cpu pages 262144, num gpu pages 262144 [2.921136] [drm] PCIE GART of 1024M enabled. [2.921137] [drm] PTB located at 0x00F40090 [2.921208] Loading firmware: amdgpu/picasso_asd.bin [2.921688] Loading firmware: amdgpu/picasso_ta.bin [2.921897] amdgpu :2a:00.0: amdgpu: PSP runtime database doesn't exist [2.921939] Loading firmware: amdgpu/picasso_pfp.bin [2.922059] Loading firmware: amdgpu/picasso_me.bin [2.922246] Loading firmware: amdgpu/picasso_ce.bin [2.922344] Loading firmware: amdgpu/picasso_rlc_am4.bin [2.922846] Loading firmware: amdgpu/picasso_mec.bin [2.923079] Loading firmware: amdgpu/picasso_mec2.bin [2.924074] amdgpu: hwmgr_sw_init smu backed is smu10_smu [2.924081] Loading firmware: amdgpu/raven_dmcu.bin [2.924221] Loading firmware: amdgpu/picasso_vcn.bin [2.924504] [drm] Found VCN firmware Version ENC: 1.15 DEC: 3 VEP: 0 Revision: 0 [2.924510] amdgpu :2a:00.0: amdgpu: Will use PSP to load VCN firmware [2.945168] [drm] reserve 0x40 from 0xf47fc0 for PSP TMR [3.004533] amdgpu :2a:00.0: amdgpu: RAS: optional ras ta ucode is not available [3.009017] amdgpu :2a:00.0: amdgpu: RAP: optional rap ta ucode is not available [3.011810] [drm] psp gfx command LOAD_TA(0x1) failed and response status is (0x7) [3.011919] [drm] psp gfx command INVOKE_CMD(0x3) failed and response status is (0x4) [3.011922] amdgpu :2a:00.0: amdgpu: Secure display: Generic Failure. [3.011924] amdgpu :2a:00.0: amdgpu: SECUREDISPLAY: query securedisplay TA failed. ret 0x0 [3.013330] [drm] kiq ring mec 2 pipe 1 q 0 [3.014477] [drm] DM_PPLIB: values for F clock [3.014478] [drm] DM_PPLIB: 160 in kHz, 4399 in mV [3.014481] [drm] DM_PPLIB: values for DCF clock [3.014482] [drm] DM_PPLIB: 30 in kHz, 3099 in mV [3.014484] [drm] DM_PPLIB: 60 in kHz, 3574 in mV [3.014485] [drm] DM_PPLIB: 626000 in kHz, 4250 in mV [3.014487] [drm] DM_PPLIB: 654000 in kHz, 4399 in mV [3.014716] [drm] Display Core initialized with v3.2.149! [3.100173] [drm] VCN decode and encode initialized successfully(under SPG Mode). [3.101219] kfd kfd: amdgpu: Allocated 3969056 bytes on gart [3.101437] amdgpu: Topology: Add APU node [0x15d8:0x1002] [3.101440] kfd kfd: amdgpu: added device 1002:15d8 [3.101447] kfd kfd: amdgpu: Failed to resume IOMMU for device 1002:15d8 [3.101449] amdgpu :2a:00.0: amdgpu: amdgpu_device_ip_init failed [3.101452] amdgpu :2a:00.0: amdgpu: Fatal error during G
[gentoo-user] Re: is a global use flag necessary for python?
On 2024-03-09, Walter Dnes wrote: > On Sat, Mar 09, 2024 at 07:55:13PM +0100, n952162 wrote >> I just synced my system after a long delay, > > That's your problem right there. Yep, to quote Olivia Rodrigo... Bad idea, right? >> Is there a way to do it globally? > > First of all python targets should not need to be mentioned in > make.conf or package.use. Like the girl said. > Gentoo manages versions automatically... if you update often enough. This has been pointed out here (I think even to the OP) more than once. Gentoo is not appropriate for machines that don't get updated regularly. I recommend at least once a week, but once a month will probably be OK. Much longer than a few months and you're asking for trouble. Hard-wiring PYTHON versions is also asking for trouble. Combining the two is demanding trouble. Just back up your user data and re-install. There are ways to get out of your mess, but if you have to ask... As they used to say at the billiards tournaments: "If you could make a shot like that, you wouldn't need to make a shot like that." -- Grant
[gentoo-user] Re: How to set up drive with many Linux distros?
On 2024-03-10, Michael wrote: > Perhaps I'm picking up on semantics, but shouldn't this sentence: > > "... The gap between the DOS disklabel and the first partition" > > read: > > "The gap between the MBR and the first partition"? Yes, thanks -- MBR is more accurate, I've changed that sentence. > Your next paragraph pointed out something which I hadn't considered at any > length. Namely, the installation of GRUB's boot.img in a MBR or VBR also > hardcodes in a block list format the location of the first sector where the > core.img is stored and more importantly, the physical position of this sector > can be altered both by COW fs (and by the wear levelling firmware of flash > storage devices). > > I had assumed both the COW fs and/or the flash controller will in > both cases translate any physical data position to the logical layer > and presented this to inquiring software. Have you actually tried > using btrfs as a distro's root fs to see if the VBR installed GRUB > boot.img will ever lose access to the core.img? No, I haven't. I agree that the flash controller can't change the logical address of a filesystem data block without the knowledge of the filesystem, so I don't think controller layer wear-leveling would be a problem. But, the filesystem layer is allowed to move data blocks around, so flash-aware filesystems that attempt to do wear-leveling or defragmentation could move data blocks. Some of the descriptions I've read of "fancier" filesystem internals have also implied implied that does happen under certain conditions, but I may have misunderstood or the descriptions may have been wrong. My use of these multi-boot installs have no need for anything beyond exnN, so I've never tried using block lists with anything other than extN filesystems. Since I am confident extN filesystems won't cause problems, I've always stuck with that. -- Grant
[gentoo-user] Re: How to set up drive with many Linux distros?
On 2024-02-22, Grant Edwards wrote: > For many years, I've used a hard drive on which I have 8-10 Linux > distros installed -- each in a separate (single) partition. > > [...] > > Is there an easier way to do this? After some additional studying of UEFI and boot managers like rEFInd, I decided that my current approach was still the easiest method. I did switch from DOS to GPT disklabel (I bricked my old drive tring to update the firmware, so I had to start over anyway). In case anybody is interested in the gory details, I documented my scheme and the helper shell scripts at https://github.com/GrantEdwards/Linux-Multiboot -- Grant
[gentoo-user] Re: Problem with "GRUB upgrades" news item
On 2024-03-06, Walter Dnes wrote: > I've got a UEFI system. According to the news item... > >> Re-runing grub-install both with and without the --removable option >> should ensure a working GRUB installation. > > I tried that... > > [i3][root][~] grub-install I believe you have to run grub-install using all the same options you did originally. AFAICT, grub doesn't remember what you did the last time.
[gentoo-user] Re: How to set up drive with many Linux distros?
On 2024-02-26, Wol wrote: > On 26/02/2024 20:51, Grant Edwards wrote: > >> The simple answer is to quit wasting time trying to multi-boot like >> that and just buy a dozen USB flash drives. >> > And then, if USB isn't the default boot media, he might as well sort out > UEFI boot, and multi-boot that way. Except that every time I've found a write-up about multibooting a lot of Linux distros with UEFI, it turns out that it doesn't actually work very well and is more work to maintain than what I'm doing now. -- Grant
[gentoo-user] Re: How to set up drive with many Linux distros?
On 2024-02-26, eric wrote: > I agree, using the custom.cfg file would not work if needing to boot > different kernels of the same OS and those kernels were being updated. The simple answer is to quit wasting time trying to multi-boot like that and just buy a dozen USB flash drives. -- Grant
[gentoo-user] Re: How to set up drive with many Linux distros?
On 2024-02-26, eric wrote: > On 2/26/24 04:57, gentoo-u...@krasauskas.dev wrote: >> You could also write a script that keeps all the distros up to date >> from within whichever one you're currently booted by mounting >> subvolumes to /mnt or wherever, chrooting in and running the update. > > To avoid grub not being able to point to a newly updated kernel on one > of the OS's installed, I use a "custom.cfg" file in all my /boot/grub/ > directories for each OS where the "linix" and "initrd" point to the > symbolic links of the kernel and init files which point to the newly > updated files on most major distributions like ubuntu, arch, suse, and > debian. The name of the symbolic links stay the same over upgrades. It > works great when using UUID to identify the partition that has root and > I can always boot into any of the OS's installed no matter which one > hijacked the MBR. Except I generally have multiple kernels installed for each of the distros, and need to be able to choose which kernel to boot. There are also various other boot options (e.g. "safe mode") offered by some distros that I occasionally need to use. > https://forums.linuxmint.com/viewtopic.php?t=315584 Interesting article, thanks. After reading up more on UEFI, it looks like that would be even more work and more mess. So, there seem to be two options: 1) Stick to the dual-stage chainloading scheme I'm using now (though I'll probably switch from DOS to GTP disklabel). That way after selecting which parition (distro) to boot, I get all the boot options normally offered by that distro's install. Installing a distro involves letting it install to MBR and BIOS-boot, installing grub manually to the root partition, then restoring MBR and BIOS-boot. 2) Use a single master grub to boot any distro. I think I'd need to write my own OS-prober. All of the distros I care about seem to now be using grub2 now. Instead of looking for kernels and initrd images and adding them to the master grub.cfg, I would probe for grub.cfg files, and for each one found incorporate the entire set of choices in that .cfg file as a submenu in the main grub.cfg menu. I think that in order to generate the distro's grub.cfg files, I still have to allow the distros to install grub to the MBR and BIOS-boot (or to a second disk that I don't care about), then restoring MRB/BIOS-boot.
[gentoo-user] Re: How to set up drive with many Linux distros?
On 2024-02-23, Mark Knecht wrote: > The only other idea I had was to install to a different > disk and then use something like Clonezilla to move it to the partition > you want it in on your system. > > While I suspect you were being sarcastic I do not think any solution > that involves a 'pocketful of USB 3 thumb drives' will be satisfying. Actually (assuming the thumb drives are relaible) it probably would work fine. That's more-or-less the typical solution that my colleagues use -- though it's a drawer full of hard drives in caddies instead of USB thumb drives. Back in the day when we supported a couple versions of SCO, various BSDs, Novell Netware and Solaris, it required a fait bit of drawer space. For now, I guess I'll stick with the scheme I'm using but switch from DOS disklabel and gap to GPT disklabel and BIOS boot partition. It seems ugly, but it's managable with the help of a few shell scripts that are stored in the parition that belongs to the master copy of grub.
[gentoo-user] Re: How to set up drive with many Linux distros?
On 2024-02-23, Mark Knecht wrote: > On Fri, Feb 23, 2024 at 11:59 AM Grant Edwards > wrote: >> >> The simple solution is to give up on multi-booting a dozen different >> distros on a single disk and buy a pocketful of USB 3 thumb drives. >> > > Given performance does drop a bit and there can be issues with allocating > hardware, why not use Virtualbox which allows you to run both your base > distro and then as many of the others as your hardware can handle as VMs? The machine is used for testing PCI and PCI-express boards and their drivers. I don't really trust PCI passthrough (especially when it comes to interrupt handling) enought to do such testing in a VM. IFAICT, it's very rare for customers to use VMs. If customers do start to use VMs, then we'll have to test with VMs also. -- Grant
[gentoo-user] Re: How to set up drive with many Linux distros?
On 2024-02-23, Wols Lists wrote: > On 23/02/2024 00:28, Grant Edwards wrote: >> In my experience, 's bootloader does not boot other >> installations by calling other bootloaders. It does so by rummaging >> through all of the other partitions looking for kernel images, >> intird files, grub.cfg files, etc. It then adds menu entries to >> the config file for 's bootloader which, when selected, >> directly load the kernel image and initrd from those other >> partitions. Sometimes, it works -- at least until those other >> installations get updated without the knowlege of the distro that >> currently "owns" the MBR's bootloader config. Then it stops working >> until you tell that bootloader to re-do it's rummaging about >> routine. > > IME distros that try that (SUSE, anyone!) generally get confused as > to which kernel belongs to which root partition. > > Hence needing to boot with a live distro to edit the resulting mess > and get the system to actually come up without crashing ... IIRC, all of the big distros used to do that. It didn't work very well, but at least it took a really long time. However, I read recently that Ubuntu had disabled the os-prober by default in 22.04. Disabling it was always one of the first things I did after installing a new distro. The simple solution is to give up on multi-booting a dozen different distros on a single disk and buy a pocketful of USB 3 thumb drives. -- Grant
[gentoo-user] Re: How to set up drive with many Linux distros?
On 2024-02-23, Michael wrote: > The problem starts if/when kernel images are overwritten by > successive Linux OS distros. This is likely when derivatives of the > same main distros e.g. Ubuntu all create a directory called > /EFI/ubuntu/ in the ESP and drop their kernels & initrd images in > there potentially overwriting other distro's files. Yes, that's the problem that I've read about when trying to multi-boot with UEFI. I usually have 3-4 different Ubuntu installations and 3-4 different RedHat installations. Ubuntu in particular causes a log of complaints about overwriting ESP files belonging to other Ubuntus. > When using a distro's installer menu on a legacy BIOS MoBo you can > select a partition (PBR) to install GRUB, You used to be able to. I can no longer find the option to do that in Ubuntu or RedHat. I've been told that Suse still allows it. > but GRUB will complain and suggest you could use blocklists but it > is unreliable. Last time I received an error like this, I installed > grub in a PBR manually with the '--force' option, without using the > installer GUI. After that, whenever I updated GRUB it complained > again about blocklists, but it worked fine. Using --force will work fine as long as the grub 1.5 files don't get moved afterwards. That's why I lock them in place. Locking them will cause future upgrades to Grub for that distro to fail, but that doens't happen very often. When it does, you unlock them, updated grub, re-install using --force, and lock them again. >> I'd welcome pointers to where those advanced options are in the RH >> and Ubunutu installers -- I've searched everywhere I can think >> of. Various things Google has found lead me to believe that they no >> longer support installing grub in a partition. > > Try using '--force' to make GRUB install its image in some distro's > boot/root partition PBR instead of the disk MBR, but you'll probably > have to perform this outside the installer script. I've done this > with VMs. Yes, in my OP describing what I'm doing now it explains that's what I do. Then I lock the Grub files that are located using the blocklists created by the --force option. >> I guess I'll stick with my current setup. >> >> Or perhaps I'll switch from a DOS disklabel to a GPT disklabel. >> Instead of backing up and restoring the MBR and the gap, I would >> backup and restore the MBR and the BIOS boot partition. And I could >> use UUIDs and partition labels. > > These days I use disks with GPT even on MoBos with legacy BIOS. Same here -- except for this one machine. I think I'll switch it over soon. > Instead of backing up and restoring the MBR/BIOS Boot Partition you > could just reinstall grub and run grub-mkconfig, as long as the > latter involves fewer key-presses. ;-) I don't use grub-mkconfig for the "main" grub. It has a fixed grub.cfg file that does nothing but chainload the user-selected partition. Currently, backing up MBR+gap only happens once when I install/setup the main grub. Restoring BMR+gap is one command (which is actually in a shell script) that's run after any new distro is installed. MBR+Bios-boot-partition would work pretty much the same way. -- Grant
[gentoo-user] Re: How to set up drive with many Linux distros?
On 2024-02-23, Wojciech Kuzyszyn wrote: > I guess most (all) of the distro's you are talking about use GRUB (or > at least they allow to do it). Yes, I belive that they are all now using Grub2. > If that's true, I'm pretty sure you can happily let them overwrite > the GRUB in MBR as many times as they want, since it's the same (or > just probably minor version differences) bootloader. Just make a > copy of /boot/grub/grub.cfg and make sure it's the same on every > partition. That means I have to update all of those grub.cfg files manually every time I install or update a distro. That's a lot of work. > Or, even better, if that's possible right now, make a common /boot > partition and after installing the new distro just merge the > (probably new) /boot/grub/grub.cfg with your old one. "just merge the new grub.cfg file with the old one" is the problem. I tried that for a while: every time a distro got installed, I would add it to the "main" grub config file. Every time a distro got a kernel update, I'd modify the main grub config file. It was a lot more work that my current scheme (and a lot more error prone). > I really think that *should* work! It would work, but maintaining the grub.cfg files is a lot of work. The scheme I'm using now doens't require me to mess with any of the grub.cfg files when distros get updated or installed. -- Grant
[gentoo-user] Re: How to set up drive with many Linux distros?
On 2024-02-22, Wol wrote: > On 22/02/2024 21:45, Grant Edwards wrote: >> I've been reading up on UEFI, and it doesn't seem to be any >> better. People complain about distro's stomping on each other's files >> in the ESP partiton and multiple distro's using the same name in the >> boot slots stored in NVM. And then the boot choice order changes >> (though it may not be apparent to the naked eye) when one of the >> distros decides to update/reinstall its boot stuff. > > At least if you use UEFI *as* your bootloader, then that won't > happen. That assumes you're using UEFI, though! According to what I've read UEFI isn't a bootloader. It's a boot manager which can load and run EFI bootloaders (of which there can be multiple installed). > In which case, 's bootloader doesn't get a look-in. Yes, AFAICT, it does (sometimes?). When you install under UEFI it installs EFI bootloader files (either kernels wrapped in EFI bootloader executables or the grub EFI bootloader) in the EFI Systgem Partition (ESP), and then adds one or more entries in the EFI NVM that points to those files (or something like that). The Linux UEFI systems I have all still use grub2 (which gets written to the ESP). It's entirely possible for one distro to overwrite files in the ESP that belong to other distros. I've read multiple complaints about exactly that when trying to do multi-boot with UEFI. In practice it's just like the fight over who owns the MBR and the DOS disklable gap. One recipe I read about for doing what I wanted to do with UEFI involved installing a Linux distro (didn't really matter which), then installing rEFInd. After that, some manual renaming and deleting of the files in the ESP was required. Then he started installing various distros. After each distro installation, the author had to re-install rEFInd, and after many of them he had to manually remove or rename files in the ESP (or adjust the rEFInd config file). And in the end, he ended up with multiple menu entries (for different installations) that had identical names. It was more complicated and difficult than my current scheme. > As for "'s obviously superior bootloader", well > is using the exact same boot-loader, and when IT installs, how is it > going to be able to boot if it can't call 's boot > loader because it's just trashed it by overwriting it? In my experience, 's bootloader does not boot other installations by calling other bootloaders. It does so by rummaging through all of the other partitions looking for kernel images, intird files, grub.cfg files, etc. It then adds menu entries to the config file for 's bootloader which, when selected, directly load the kernel image and initrd from those other partitions. Sometimes, it works -- at least until those other installations get updated without the knowlege of the distro that currently "owns" the MBR's bootloader config. Then it stops working until you tell that bootloader to re-do it's rummaging about routine. > To me, you seem to be describing the *default* installer setup, that's > been there for ever. Last I installed SUSE, iirc I had to specify > "advanced bootloader installation", most of who's options I didn't even > understand!, but it did do what I told it to (apart from not recognising > my weird disk stack!). So SuSe still allows you to install grub to a partition instead of MBR? That's encouraging. RH and Ubuntu used to allow that, but AFAIK, now they do not. > If you can find, and understand!, that advanced options, I think > you'll find you can do what you want. I'd welcome pointers to where those advanced options are in the RH and Ubunutu installers -- I've searched everywhere I can think of. Various things Google has found lead me to believe that they no longer support installing grub in a partition. I guess I'll stick with my current setup. Or perhaps I'll switch from a DOS disklabel to a GPT disklabel. Instead of backing up and restoring the MBR and the gap, I would backup and restore the MBR and the BIOS boot partition. And I could use UUIDs and partition labels.
[gentoo-user] Re: How to set up drive with many Linux distros?
On 2024-02-22, Wol wrote: > On 22/02/2024 19:17, Grant Edwards wrote: > >> However, the choice to install bootloaders in partitions instead of >> the MBR has been removed from most (all?) of the common installers. >> This forces me to jump through hoops when installing a new Linux >> distro: > > File a bug! LOL, good one! As if a normal person filing a bug with RedHat or Ubuntu actually accomplishes anything. I'll tell them to make systemd optional while I'm at it. > If that's true, it basically borks any sort of dual boot, unusual disk > layout, whatever. Yep, it does. The answer from is: You should really just (shut up and) install 's (obviously superior) bootloader in the MBR. It will auto-detect (some of) the other already installed (obviously inferior) OSes, and will add (some subset of) them to the boot menu that will (sometimes) allow you to boot them (maybe -- if you kneel, bow your head and ask nicely). > Last time I installed SUSE, it trashed my boot totally because it > didn't recognise my disk stack, failed to load necessary drivers, > and worse trashed my gentoo boot too... > > Cue one big rescue job to get the system up and working again. At > least it was only the boot that was trashed. I've been reading up on UEFI, and it doesn't seem to be any better. People complain about distro's stomping on each other's files in the ESP partiton and multiple distro's using the same name in the boot slots stored in NVM. And then the boot choice order changes (though it may not be apparent to the naked eye) when one of the distros decides to update/reinstall its boot stuff. -- Grant
[gentoo-user] How to set up drive with many Linux distros?
For many years, I've used a hard drive on which I have 8-10 Linux distros installed -- each in a separate (single) partition. There is also a single swap partition (used by all of the different Linux installations). There is also a small partition devoted only to the "master" instance of Grub that lives in the MBR and the space between the MBR and the first partition (the drive uses a DOS disklabel). That master instance of Grub has a menu which contains entries which "chainload" each of the other partitions. For many years, this worked great. All of the various distro installers offered the option of installing the bootloader in the MBR (e.g. /dev/sda) or in a partition (e.g /dev/sdaN). I would tell the installer to install the bootloader in the root partition, and everything "just worked". However, the choice to install bootloaders in partitions instead of the MBR has been removed from most (all?) of the common installers. This forces me to jump through hoops when installing a new Linux distro: 1. Back up the MBR and gap between the MBR and the first partition. 2. Let the installer install it's bootloader (seems it's always grub these days) in the MBR. 3. Boot into the newly installed Linux. 4. Manually install grub in the root partition (e.g. /dev/sdaN) using the --force option to tell grub to use blocklists to find it's files. 5. Find those grub files and lock them so they can't be moved. 6. Restore the MBR/gap backup from step 1. It seems like there should be a better way to do this. One might hope that UEFI offers a solution to this problem. Google has found me others asking the same question but no real answers. Is there an easier way to do this? -- Grant
[gentoo-user] Re: Re-run grub-install to update installed boot code!
On 2024-02-17, Dale wrote: > Grant Edwards wrote: >> Today's routine update says: >> >> Re-run grub-install to update installed boot code! >> >> Is "sudo grub-install" really all I have to do? [...] >> >> Or do I have to run grub-install with all the same options that >> were originally used to install grub? > > I been wondering the same since I saw this posted on -dev. The news > item seems to mention the EFI booting but I'm sure us legacy booting > users need to do the same. At this point, I may skip updating grub > this week until I know exactly what I'm supposed to do as well. I'd > think we need to reinstall like when we first did our install but > not sure. :/ That was my guess. I should have recorded the options originally passed to grub-install. Now that I have BIOS boot partitions (instead of using embedded blocklists) on all my machines, reinstalling grub should be trivial. I think all I have to do is tell grub-install the boot device. > It would suck to have a unbootable system. More than once I've had to boot from either systemrescuecd or minimal gentoo install ISO so I could re-install (or re-configure) grub after someting gets messed up. It's not difficult, but it is annoying. -- Grant
[gentoo-user] Re-run grub-install to update installed boot code!
Today's routine update says: Re-run grub-install to update installed boot code! Is "sudo grub-install" really all I have to do? Grub knows where/how everthing was originally installed and will do the right thing without any options? Or do I have to run grub-install with all the same options that were originally used to install grub? [I use a manually generated grub.cfg file, so I'm ignoring the message that tells me I to run "grub-mkconfig".] -- Grant
