On 2024-08-06, CrudeSausage <crude@sausa.ge> wrote:
> On 2024-08-05 9:01 p.m., RonB wrote:
>> On 2024-08-05, CrudeSausage <crude@sausa.ge> wrote:
>>
>> Lots of snips.
>>
>>> So, Dell it is if I want a machine capable of running Linux reliably and
>>> don't want to go for a System76 or ThinkPenguin.
>>
>> Well, hold on for a second... I went to start the D630 this evening using
>> battery power, i.e. not plugged in to the adapter. (I guess, in the past, I
>> had always started it when it was plugged in, though I have used it without
>> the battery after starting it — Or maybe this started when the new kernel
>> was installed?) I worked with it using the adapter because it hadn't been
>> charged in a while. At any rate, without the D630 being plugged in, it
>> locked when starting after the grub menu? Works fine with the power plugged
>> in — nothing wrong with the battery (shows all green lights when the test
>> button is pushed — used to have these test buttons on old laptop batteries).
>>
>> So I tried the recovery mode in grub... it would *sometimes* allow me to
>> "recover" and then resume, but without the right drivers. If I would restart
>> from there, it would boot with the right drivers coming up. Weird. (And this
>> wasn't consistent, sometimes recovery mode would hang after the grub menu
>> and sometimes it would lock when restarting without even getting to grub.)
>>
>> The "workaround" that works every time is to go into the BIOS (F2 key on a
>> Dell), do nothing and exit from the BIOS without saving. The D630 will then
>> boot normally on battery power. A quick Internet search tells me that is
>> probably a kernel issue. They suggest a BIOS update. I thought I was on the
>> newest BIOS, but it turns out I'm using A17, which is no longer on Dell's
>> site, but A18 came out in 2012 — and there's an A19 BIOS that was released
>> in 2020. So I'll update the BIOS and see if that fixes the issue.
>>
>> So, yeah, I do occasionally run into Linux bugs but (so far) always find
>> fixes. I'll report on whether the BIOS fixes this issue or not.
>
> I imagine that it will. I'm glad that you're honest about the bugs. In
> my case, I don't have the patience to deal with occasional bugs that I
> used to. I'd love to get away from a world where big tech has any
> influence over me, but I'm not willing to get surprised with a freeze
> while I'm working on something or teaching a class.

It doesn't look like the A19 BIOS was the answer. I do occasionally get the
D630 to boot from the previous kernel but it's intermittent (and it may have
intermittent before). Booting into BIOS and exiting from it still works.

More research.

>>> Hibernation is indeed useless on machines that suspend reliably in S3.
>>> However, most current machines not manufactured by Apple use S0 (Modern
>>> Standby) which wakes (barely) faster than a machine using the old S3.
>>> While it is "sleeping," it uses less power but not as little as an S3
>>> sleeper would. For those machines, using hibernation makes sense because
>>> continuous sleep would drain the battery continuously.
>>>
>>> To give you a quick example, Mac users are the kind of people who never
>>> shut down. When they're done with their work, they simply close the lid.
>>> If they wake two days later, they expect the machine to still have as
>>> much battery as it did when they put it to sleep. By default, MacOS no
>>> longer affords them this possibility but it is rather easy to disable
>>> the equivalent of modern standby on MacOS to make it possible. In
>>> Windows, Modern Standby is the default everywhere. If your computer uses
>>> about 10,000 mW every hour, closing the lid will drop its use to about
>>> 2,000 mW. However, it is still draining. If hibernation is not enabled,
>>> the machine won't eventually "turn off" so if you're trying to do what
>>> the Mac users do, the machine will no longer have any left to it when
>>> you open the lid. It is impossible to disable Modern Standby on most new
>>> machines. Even in my cases, I have to use a third-party utility to
>>> disable it in the BIOS and revert to the old S3. It will work properly
>>> when I go to sleep, but when it wakes the machine will use a ridiculous
>>> amount of power (like 25,000 mW) whereas sticking to S0 will get to run
>>> normally at 5,500 mW or so. For me, it makes more sense to stick to
>>> Modern Standby but enable hibernation if I'm away from the machine for a
>>> long time and don't want to shut down.
>>>
>>> Why is Modern Standby the default? Because Microsoft and manufacturers
>>> figured that most users want instant-on. Additionally, Modern Standby
>>> allows the network to still be connected even though the machine is
>>> sleeping. So you can push updates even during that time. Theoretically,
>>> you can kill whatever additional drain S0 causes by disabling the
>>> network connectivity, but manufacturers often disable that the same way
>>> they disable reverting to S3.
>>>
>>> In Linux, Modern Standby simply doesn't work so you _need_ for S3 to
>>> work reliably. On this machine, it works but not reliably. If I don't
>>> set it to use S3, closing the lid will do nothing other than maybe shut
>>> down the screen. If I make the mistake of putting my laptop in its bag
>>> after closing the lid, it will drain and overheat, potentially killing
>>> the computer entirely.
>>
>> Too much information for my feeble brain. I don't know which suspend mode
>> the D630 uses. I just know it comes on immediately when I open the lid and
>> the green light blinks slowly when in suspend mode. I'm guessing it does use
>> more power than the newer computers. It doesn't matter to me because I
>> almost never leave my laptops in sleep mode for more than a few minutes.
>
> Sleep is mostly useless to people who use their computer in one place at
> all times. In my case, it's not unusual to go from one room to another.
> Sleeping, moving to the next room and waking is a lot more convenient
> than constantly restarting the machine, especially if your bootup
> process includes logging into the system, then logging into Bitwarden,
> then logging into the Bitwarden extension in your browser before logging
> into a site. Sometimes, you just want everything to already be
> authenticated so you can get going.

Short term sleeping works fine on my laptops — actually what you describe
above (moving from room to room) is the only way I use sleep. Or, if I'm
going to get a cup of coffee or a snack, I'll close the lid of the laptop to
protect it.

>>> Because Intel prohibits manufacturers from putting an Intel wireless
>>> chip on a laptop that uses an AMD processor.
>>
>> I didn't know that. My D630 is labeled "Centrino," which, I've read, meant
>> it had an Intel motherboard, Intel CPU and Intel WiFi card. I've seen other
>> D630s without Centrino labels, because they came with Broadcom WiFi cards. I
>> have no idea why Dell didn't just always use Intel WiFi cards. Apparently
>> this has changed recently?
>
> I wouldn't know. I never really paid attention to the brand of any wifi
> chip I had until I had problems with the MediaTek I owned. Even the
> Atheros in my MSI was quite reliable.

I've rebuilt a lot of cheap Dell laptops (and given away a few of them to
extended relatives) so I've run into this a lot.

>>> I would bet that the wireless chips that uses 802.11ac work well but the
>>> 802.11ax do not.
>>
>> I don't know what the "ax" stands for.
>
> WiFi 6, so theoretical wireless speeds of 9.6Gbps vs the 3.5Gbps of WiFi
> 5. Good luck actually getting either of those speeds though.

Okay. So newer hardware than I'll ever need.

>>> I hope you have some luck because it would mean that I could eventually
>>> move to Linux with confidence. However, the wokeness of the developers
>>> at the moment is another hurdle. I figure it's a matter of time before
>>> they do to LibreOffice what they did to Microsoft Office with its
>>> automatic "inclusive language" functionality. Luckily, it's disabled by
>>> default.
>>
>> That update is going to be put on hold until I figure out the D630 issue. I
>> think I'm already running Linux Mint 20 on that computer, so it's still
>> supported until next April.
>
> I'm sure that you'll figure out the issue, but I worry that it will
> simply open up another one. That's been my experience with using
> Linux... it's always a game of whack a mole for me. Of course, that's
> also been my life. To give you an example, I ended up with a pretty
> large puck on the driver side car door on my SUV. I had it painted
> professionally for a pretty penny only to get a puck on the driver side
> again, this time near the gas container. This kind of stuff always
> happens to me.

I don't know if I'll figure it out or not. I guess I need to go to the logs,
which is what is what they always tell you to do. So far I've been able to
avoid doing that.

Except for the challenge of it, booting into the BIOS and out would probably
be my "solution" while hoping a future kernel would fix the issue properly.
Getting the D630 running Linux 22 was more to see if it would work than
anything else at this point.

--
[Self-centered, Woke] "pride is a life of self-destructive fakery, an
entrapment to a false and self-created matrix of twisted unreality."
"It was pride that changed angels into devils..." — St. Augustine