I just recently transitioned from static device nodes (simple and efficient) to dynamic
nodes (complex and pointless) using e/udev (not because of desire but because
of necessity due to the libinput fiasco wrought by GNOME/freedesktop).

My root directory is on partition /dev/sda3.

When I execute " df -h /dev/sda3" I get the message:

devtmpfs 7.8G 0 7.8G 0% /dev

WTF! This is nonsense!

The command "cfdisk /dev/sda" shows this:

Device Boot Start End Sectors Size Id Type
/dev/sda1 * 2048 314574847 314572800 150G 83 Linux
/dev/sda2 314574848 629147647 314572800 150G 83 Linux
/dev/sda3 629147648 681576447 52428800 25G 83 Linux
/dev/sda4 681576448 1953525167 1271948720 606.5G 5 Extended
├─/dev/sda5 681578496 786436095 104857600 50G 83 Linux
....

The size of /dev/sda3 is 25G.

Also, the mount command shows:

/dev/sda3 on / type ext3 (rw,noatime,nodiratime)
....

No other partition shows this behavior.

If I execute "df -h /dev/sda5" I get:

/dev/sda5 49G 5.1G 42G 11% /usr

which is the correct usage information.

Why does "df -h /dev/sda3" fail to show the correct usage?

With static nodes this does not happen. Only the "new improved Linux," courtesy of
the GNOME/freedesktop faggots, does this lunacy prevail.

--
Scratch your technical itch:
https://www.linuxfromscratch.org/

On 11/29/21 4:40 PM, Diego Garcia wrote:
> When I execute " df -h /dev/sda3" I get the message:
>
> devtmpfs 7.8G 0 7.8G 0% /dev
>
> WTF! This is nonsense!

That's what I'm used to. (More details below.)

> If I execute "df -h /dev/sda5" I get:
>
> /dev/sda5 49G 5.1G 42G 11% /usr
>
> which is the correct usage information.

That's unexpected to me.

> Why does "df -h /dev/sda3" fail to show the correct usage?

It does show the correct usage to me.

I'm used to df (disk free) showing the free space of the file system
containing the file that you ask about. So when you ask about
/dev/sda3, that's device file named "sda3" in the "/dev" file system.
The "/dev" file system happens to be tmpfs. Hence why you see what you
see in the first output.

I would have expected "df -h /dev/sda5" to also show the free space of
the "/dev" file system.

Try /just/ "df -h" and see if it shows your root file system. I expect
that it will.

--
Grant. . . .
unix || die

On Mon, 29 Nov 2021 18:40:46 -0500, Diego Garcia <dg@chaos.info> wrote:
> Why does "df -h /dev/sda3" fail to show the correct usage?
>
> With static nodes this does not happen. Only the "new improved Linux," courtesy of
> the GNOME/freedesktop faggots, does this lunacy prevail.

On my Mageia 8 system, which uses udev ...
$ df /dev/sdb2
Filesystem Size Used Avail Use% Mounted on
/dev/sdb2 63G 48G 13G 80% /

So it isn't a caused by using udev.

What's the full output of "df -h" without a device specified?

Regards, Dave Hodgins

On 29.11.2021 at 17:19, Grant Taylor scribbled:

> On 11/29/21 4:40 PM, Diego Garcia wrote:
>
> > When I execute " df -h /dev/sda3" I get the message:
> >
> > devtmpfs 7.8G 0 7.8G 0% /dev
> >
> > WTF! This is nonsense!
>
> That's what I'm used to. (More details below.)

I don't know what distribution the OP is using, but he explicitly
mentioned eudev, which is a standalone udev fork created by Gentoo
because of earlier (meanwhile proven-false) concerns that the mainstream
udev — which now comes bundled as part of systemd — would no longer be
supporting distributions not based upon systemd.

I don't know what distribution the OP is using, but you yourself are a
Slackware user, and Slackware also used eudev. The OP is himself most
likely not using Slackware, because his distribution allows for the
installation of GNOME, which Slackware does not carry in its repository
— maybe he's using Devuan, because even Gentoo has already long ago
transitioned away from static device nodes.

So, what we have here are two presumably different distributions, both
using eudev, and both manifesting the same oddity in the output of the
"df" command, while at the same time, this oddity does not occur in
Mageia (as testified by Dave W. Hodgins), nor in Manjaro (as verified
by myself), both of which use the mainstream udev — Mageia is a
Mandriva fork, while Manjaro is based upon Arch and uses many
unmodified Arch packages, among which systemd and udev.

The above all said, logic therefore suggests that this odd behavior of
the "df" command would be due to a bug in eudev, and if so, then it
would probably be wise to report this to the upstream eudev developers
over at Gentoo.

--
With respect,
= Aragorn

On 30/11/2021 00.40, Diego Garcia wrote:
> I just recently transitioned from static device nodes (simple and efficient) to dynamic
> nodes (complex and pointless) using e/udev (not because of desire but because
> of necessity due to the libinput fiasco wrought by GNOME/freedesktop).
>
> My root directory is on partition /dev/sda3.
>
> When I execute " df -h /dev/sda3" I get the message:
>
> devtmpfs 7.8G 0 7.8G 0% /dev

That's not your root, that's the size of the filesystem where
"/dev/sda3" is mounted, ie, "/dev".

I rather prefer that to see a copy paste showing the complete command
prompt sequence from start to end. Like this:

Telcontar:~ # df -h /dev/sda3
Filesystem Size Used Avail Use% Mounted on
devtmpfs 16G 36K 16G 1% /dev
Telcontar:~ #

....

> Why does "df -h /dev/sda3" fail to show the correct usage?
>
> With static nodes this does not happen. Only the "new improved Linux," courtesy of
> the GNOME/freedesktop faggots, does this lunacy prevail.

It works fine here.

Look, the command doesn't tell you the size of a partition, but of a
mount point. Or filesystem, rather.

Telcontar:~ # df -h /dev/sda3
Filesystem Size Used Avail Use% Mounted on
devtmpfs 16G 36K 16G 1% /dev
Telcontar:~ # df -h /dev/sda2
Filesystem Size Used Avail Use% Mounted on
devtmpfs 16G 36K 16G 1% /dev
Telcontar:~ # df -h /dev/sda1
Filesystem Size Used Avail Use% Mounted on
devtmpfs 16G 36K 16G 1% /dev
Telcontar:~ #

Before, it happened that the "mount point" /dev/sda3 happened to be the
same as root, so it gave the result you expected. But now, it is a
different partition (/dev), so it gives you the size of "/dev". Exactly
what you asked.

The correct command you want is not "df -h /dev/sda3", but "df -h /":

Telcontar:~ # df -h /
Filesystem Size Used Avail Use% Mounted on
/dev/nvme0n1p5 148G 59G 82G 43% /
Telcontar:~ #

So, stop blaming udev, gnome, freedesktop, etc :-)

--
Cheers, Carlos.

On 11/29/21 9:57 PM, Carlos E.R. wrote:
> On 30/11/2021 00.40, Diego Garcia wrote:
>> I just recently transitioned from static device nodes (simple and
>> efficient) to dynamic
>> nodes (complex and pointless) using e/udev (not because of desire but
>> because
>> of necessity due to the libinput fiasco wrought by GNOME/freedesktop).
>>
>> My root directory is on partition /dev/sda3.
>>
>> When I execute " df -h /dev/sda3" I get the message:
>>
>> devtmpfs        7.8G     0  7.8G   0% /dev
>
> That's not your root, that's the size of the filesystem where
> "/dev/sda3" is mounted, ie, "/dev".
>
>
> I rather prefer that to see a copy paste showing the complete command
> prompt sequence from start to end. Like this:
>
> Telcontar:~ # df -h /dev/sda3
> Filesystem      Size  Used Avail Use% Mounted on
> devtmpfs         16G   36K   16G   1% /dev
> Telcontar:~ #
>
>
> ...
>
>> Why does "df -h /dev/sda3" fail to show the correct usage?
>>
>> With static nodes this does not happen.  Only the "new improved
>> Linux," courtesy of
>> the GNOME/freedesktop faggots, does this lunacy prevail.
>
> It works fine here.
>
> Look, the command doesn't tell you the size of a partition, but of a
> mount point. Or filesystem, rather.
>
> Telcontar:~ # df -h /dev/sda3
> Filesystem      Size  Used Avail Use% Mounted on
> devtmpfs         16G   36K   16G   1% /dev
> Telcontar:~ # df -h /dev/sda2
> Filesystem      Size  Used Avail Use% Mounted on
> devtmpfs         16G   36K   16G   1% /dev
> Telcontar:~ # df -h /dev/sda1
> Filesystem      Size  Used Avail Use% Mounted on
> devtmpfs         16G   36K   16G   1% /dev
> Telcontar:~ #
>
>
> Before, it happened that the "mount point" /dev/sda3 happened to be the
> same as root, so it gave the result you expected. But now, it is a
> different partition (/dev), so it gives you the size of "/dev". Exactly
> what you asked.
>
>
> The correct command you want is not "df -h /dev/sda3", but "df -h /":
>
> Telcontar:~ # df -h /
> Filesystem      Size  Used Avail Use% Mounted on
> /dev/nvme0n1p5  148G   59G   82G  43% /
> Telcontar:~ #
>
>
> So, stop blaming udev, gnome, freedesktop, etc :-)

Has he tried "sudo df -h" ??? I've run into permissions
issues with df before.

On 11/29/21 7:49 PM, Aragorn wrote:
> but you yourself are a Slackware user
Actually, I'm not.

I'm replying form a Gentoo system (still using udevd).

I have terminals open administering Debian, Ubuntu, and FreeBSD systems.

I've seen the behavior I described using `df` on Solaris and AIX too.

> The above all said, logic therefore suggests that this odd behavior
> of the "df" command would be due to a bug in eudev,

I can't agree with that.

Assuming that it's a bug in eudev is somewhat like thinking the fridge
has stopped cooling when the milk smells bad. There can be multiple
things going on, like /dev being it's own file system, or not, and the
milk having a best buy date /before/ the two week vacation that you go
back from a week ago.

I maintain that `df` has always shown the free space for the file system
the referenced file is on. If /dev/<bla> is on an independent /dev file
system, then it's perfectly natural that `df` will show the /dev file
system and device, /not/ the / (root) file system and device. If /dev
is not an independent file system and instead part of the / (root) file
system, that's a horse of a different color.

I'd be curious what the type, major, and minor numbers are for the
devices in question in /dev. After all, (e)udev etc. create device
nodes. The device nodes are actually the representation into the
kernel. If the device nodes are correct, then chances are quite good
that (e)udev is perfectly fine.

It may even be something associated with the way that things are mounted
causing df to mis-report the underlying file system that the device node
lives on vs the file system the node has on it.

There could be all sorts of things going on that may not be a bug in
(e)udev.

--
Grant. . . .
unix || die

On 29.11.2021 at 21:22, Grant Taylor scribbled:

> On 11/29/21 7:49 PM, Aragorn wrote:
>
> > but you yourself are a Slackware user
>
> Actually, I'm not.

I deducted this from the fact that you appear to be mainly posting in
alt.os.linux.slackware.

My apologies, then. ;)

> > The above all said, logic therefore suggests that this odd behavior
> > of the "df" command would be due to a bug in eudev,
>
> I can't agree with that.
>
> Assuming that it's a bug in eudev is somewhat like thinking the
> fridge has stopped cooling when the milk smells bad. There can be
> multiple things going on, like /dev being it's own file system, or
> not, and the milk having a best buy date /before/ the two week
> vacation that you go back from a week ago.
>
> I maintain that `df` has always shown the free space for the file
> system the referenced file is on. If /dev/<bla> is on an independent
> /dev file system, then it's perfectly natural that `df` will show the
> /dev file system and device, /not/ the / (root) file system and
> device. If /dev is not an independent file system and instead part
> of the / (root) file system, that's a horse of a different color.

[nx-74205:/dev/pts/3][/home/aragorn]
[05:40:51][aragorn] > mount | grep sda3
/dev/sda3 on / type btrfs
(rw,nodev,noatime,nodiratime,sync,compress=zstd:3,ssd,space_cache,subvolid=257,subvol=/@)

[nx-74205:/dev/pts/3][/home/aragorn]
[05:41:28][aragorn] > mount | grep devtmpfs
devtmpfs on /dev type devtmpfs
(rw,nosuid,size=8121260k,nr_inodes=2030315,mode=755)

[nx-74205:/dev/pts/3][/home/aragorn]
[05:41:49][aragorn] > df -h /dev/sda3
Filesystem Size Used Avail Use% Mounted on
/dev/sda3 1.0G 13M 781M 2% /

[nx-74205:/dev/pts/3][/home/aragorn]
[05:42:08][aragorn] >

In other words...

° My /dev is distinct from the root filesystem — its contents reside
on devtmpfs.

° I am not seeing the odd output that the OP is seeing, and that you
yourself report to be seeing as well.

Now, devtmpfs is initialized in the kernel itself and the device nodes
for the internally mounted — in the physical/mechanical sense — block
devices is enumerated by the kernel even before userspace is
initialized, so maybe the skewed output you are both seeing has
something to do with the kernel version?

[nx-74205:/dev/pts/3][/home/aragorn]
[05:48:47][aragorn] > uname -r
5.4.159-1-MANJARO

[nx-74205:/dev/pts/3][/home/aragorn]
[05:48:55][aragorn] >

--
With respect,
= Aragorn

On 30/11/2021 00.40, Diego Garcia wrote:
> I just recently transitioned from static device nodes (simple and efficient) to dynamic
> nodes (complex and pointless) using e/udev (not because of desire but because
> of necessity due to the libinput fiasco wrought by GNOME/freedesktop).
>
> My root directory is on partition /dev/sda3.
>
> When I execute " df -h /dev/sda3" I get the message:
>
> devtmpfs 7.8G 0 7.8G 0% /dev
>
> WTF! This is nonsense!

df works only on properly mounted file systems, it fallbacks to /dev as
there ain't a mounted /dev/sda3.

I think the 7.8G is based on the amount of ram you have and those the
maximum space the /dev theoretically could occupy.

> The size of /dev/sda3 is 25G.
>
> Also, the mount command shows:
>
> /dev/sda3 on / type ext3 (rw,noatime,nodiratime)
> ...

Seems your mountpoint is wrong, the mount says it's /, so it seems you
have managed to mount it to / and on top of that you have mounted
/dev/sda2 (I'm guessing that is your system partition)

How does your /etc/fstab look like?

> With static nodes this does not happen. Only the "new improved Linux," courtesy of
> the GNOME/freedesktop faggots, does this lunacy prevail.

Switch to a distribution more fitted to your needs...

--

//Aho

On 29/11/2021 23:40, Diego Garcia wrote:
> I just recently transitioned from static device nodes (simple and efficient) to dynamic
> nodes (complex and pointless) using e/udev (not because of desire but because
> of necessity due to the libinput fiasco wrought by GNOME/freedesktop).
>
> My root directory is on partition /dev/sda3.
>
> When I execute " df -h /dev/sda3" I get the message:
>
> devtmpfs 7.8G 0 7.8G 0% /dev
>
> WTF! This is nonsense!
>
> The command "cfdisk /dev/sda" shows this:
>
> Device Boot Start End Sectors Size Id Type
> /dev/sda1 * 2048 314574847 314572800 150G 83 Linux
> /dev/sda2 314574848 629147647 314572800 150G 83 Linux
> /dev/sda3 629147648 681576447 52428800 25G 83 Linux
> /dev/sda4 681576448 1953525167 1271948720 606.5G 5 Extended
> ├─/dev/sda5 681578496 786436095 104857600 50G 83 Linux
> ...
>
> The size of /dev/sda3 is 25G.
>
> Also, the mount command shows:
>
> /dev/sda3 on / type ext3 (rw,noatime,nodiratime)
> ...
>
> No other partition shows this behavior.
>
> If I execute "df -h /dev/sda5" I get:
>
> /dev/sda5 49G 5.1G 42G 11% /usr
>
> which is the correct usage information.
>
> Why does "df -h /dev/sda3" fail to show the correct usage?
>
> With static nodes this does not happen. Only the "new improved Linux," courtesy of
> the GNOME/freedesktop faggots, does this lunacy prevail.
>
Gnome is far above the kernel/filesystem code. Why are you raving at Gnome?

--
“But what a weak barrier is truth when it stands in the way of an
hypothesis!”

Mary Wollstonecraft

Grant Taylor <gtaylor@tnetconsulting.net> writes:
> On 11/29/21 4:40 PM, Diego Garcia wrote:
>> When I execute " df -h /dev/sda3" I get the message:

What does ‘df --version’ say?

What is the full contents of /proc/mounts?

>> devtmpfs 7.8G 0 7.8G 0% /dev
>>
>> WTF! This is nonsense!
>
> That's what I'm used to. (More details below.)
>
>> If I execute "df -h /dev/sda5" I get:
>>
>> /dev/sda5 49G 5.1G 42G 11% /usr
>>
>> which is the correct usage information.
>
> That's unexpected to me.
>
>> Why does "df -h /dev/sda3" fail to show the correct usage?
>
> It does show the correct usage to me.
>
> I'm used to df (disk free) showing the free space of the file system
> containing the file that you ask about. So when you ask about
> /dev/sda3, that's device file named "sda3" in the "/dev" file
> system.

The documented behavior is:

If an argument FILE resolves to a special file containing a mounted
file system, ‘df’ shows the space available on that file system
rather than on the file system containing the device node. GNU ‘df’
does not attempt to determine the disk usage on unmounted file
systems, because on most kinds of systems doing so requires
extremely nonportable intimate knowledge of file system structures.

i.e. if you’re talking about coreutils df then you only got used to that
behavbior with unmounted filesystems.

Anyway the implication is that the OP’s df can’t determine that
/dev/sda3 is mounted for some reason. That’s a puzzle given mount can
apparently work it out.

--
https://www.greenend.org.uk/rjk/

On 30/11/2021 04:54, Aragorn wrote:
> On 29.11.2021 at 21:22, Grant Taylor scribbled:
>
>> On 11/29/21 7:49 PM, Aragorn wrote:
>>
>>> but you yourself are a Slackware user
>>
>> Actually, I'm not.
>
> I deducted this from the fact that you appear to be mainly posting in
> alt.os.linux.slackware.
>
> My apologies, then. ;)
>
No Aragorn, strictly you *inferred* it.

*deductive* logic is of the nature of:

IF a straight line is the shortest distance between two points THEN (all
of geometry, etc) is *demonstrably* true.

If you want to view it in terms of information, one can prove that e.g.
the *deductions* of formal geometry contain no more information than is
expressed by it's axioms: That is the formal proofs of geometry are in
fact no more than explications of implicit information contained in the
axioms.

This is a completely different case from *inductive* logic, where
instead of working from axioms (or prime causes) to effects, we try to
work the reverse magic: go from effects, to *possible causes*.

The problem here is that there is a many-to-one set of potential causes,
for any effect.

e.g. the *demonstrable fact* of his posting in alt.os.linux.slackware
*might* mean that he is secretly enamoured of another poster there, he
once installed slackware and hasn't unsubscribed yet, he is off his
trolley and thinks slackware is a form of clothing....

This is the generalised form of Hume's Problem Of Induction.
Winnie Ille Pu said:

"De heffalumpis semper dubitandum est. Cura ut bene valeas!"*

Which of course was plagiarised by Kierkegaard as "De omnibus
dubitandum est".

But the message is clear. Doubt must the default condition when applied
to inductive hypothesises, or inferences.

Everything from Divine Intervention to the Russians must be considered a
suspect, And as Sherlock Holmes incorrectly remarked 'when you have
eliminated all the impossible, whatever is left must be the truth' Sadly
for Sherlock, he was involved not in deduction, but in induction, and
only in fiction does the *possible* truth become *single valued*.

In short there are an infinite number of explanations that *don't* fit
the facts and an infinite number that *do*.

And anyone who claims they have arrived at 'The scientific truth' is
dangerously deluded.

William of Ockham along with other mediæval thinkers was well aware of
this :
"Numquam ponenda est pluralitas sine necessitate"

Which basically boils down to '[since we can't know the truth], keep
theories simple enough to work and no more'

A view the Church tried correctly to instil in Galileo, when he
attempted to ascribe Truth content to what was simply another inductive
theory, and as we know from Einstein, there is no centre round which
everything rotates anywhere in the Universe, it's all relative, so you
just pick the most convenient axes**.

So it is important to understand that deduction is logic, but science
and casual explanations lie in the realm of induction, or inference, and
the only advantage science ultimately has, is not the establishment of
truth content, but in the efficacy of its predictions. Science that
doesn't predict what happens, isn't science. ***

*Winnie Ille Pu:
https://kupdf.net/download/winnie-ille-pupdf_59c0fd2f08bbc57e12686fb6_pdf

**http://vps.templar.co.uk/Cartoons%20and%20Politics/centrifugal_force.png

*** except of course ClimateScience, where the 'science' has singularly
failed to predict anything that has actually happened to any degree of
accuracy better than chance.

--
"First, find out who are the people you can not criticise. They are your
oppressors."
- George Orwell

On Mon, 29 Nov 2021 17:19:36 -0700, Grant Taylor wrote:

>
> Try /just/ "df -h" and see if it shows your root file system. I expect
> that it will.
>

It does:

[~]# df -h
Filesystem Size Used Avail Use% Mounted on
/dev/root 25G 16G 8.3G 65% /
tmpfs 7.8G 508K 7.8G 1% /run
devtmpfs 7.8G 0 7.8G 0% /dev
/dev/sda5 49G 5.1G 42G 11% /usr

But this does not:

[~]# df -h /dev/sda3
Filesystem Size Used Avail Use% Mounted on
devtmpfs 7.8G 0 7.8G 0% /dev

But this also does:

[News]# df -h /
Filesystem Size Used Avail Use% Mounted on
/dev/root 25G 16G 8.3G 65% /

--
Scratch your technical itch:
https://www.linuxfromscratch.org/

On Tue, 30 Nov 2021 03:49:50 +0100, Aragorn wrote:

>
> I don't know what distribution the OP is using,
>

A distro, in my view, should only be a stepping stone toward creating ones
own standalone GNU/Linux system.

Linux is a lot simpler than it may appear and if one not be dependent on
a distro after a few years of using and learning Linux.

Every user should have his own personal distro.

--
Scratch your technical itch:
https://www.linuxfromscratch.org/

On Tue, 30 Nov 2021 08:59:45 +0000, The Natural Philosopher wrote:

>
> Gnome is far above the kernel/filesystem code. Why are you raving at Gnome?
>

I gave the reason.

GNOME, via freedesktop.org, has succeeded in mandating the use of libinput as
the only input driver for X Window, which thereby destroys the possibility of using
static device nodes which thereby overly complicates the entire GNU/Linux system.

--
Scratch your technical itch:
https://www.linuxfromscratch.org/

On Tue, 30 Nov 2021 05:54:11 +0100, Aragorn wrote:

>
> Now, devtmpfs is initialized in the kernel itself and the device nodes
> for the internally mounted — in the physical/mechanical sense — block
> devices is enumerated by the kernel even before userspace is
> initialized,
>

That may be the way most distros do it, but it's not the only way and
it's not the way that I do it.

The end result, in any case, is the same.

--
Scratch your technical itch:
https://www.linuxfromscratch.org/

On 30/11/2021 at 12.07, Diego Garcia wrote:
>> Try /just/ "df -h" and see if it shows your root file system. I expect
>> that it will.
>>
>
> It does:
>
> [~]# df -h
> Filesystem Size Used Avail Use% Mounted on
> /dev/root 25G 16G 8.3G 65% /
> tmpfs 7.8G 508K 7.8G 1% /run
> devtmpfs 7.8G 0 7.8G 0% /dev
> /dev/sda5 49G 5.1G 42G 11% /usr
>
> But this does not:
>
> [~]# df -h /dev/sda3
> Filesystem Size Used Avail Use% Mounted on
> devtmpfs 7.8G 0 7.8G 0% /dev
>
> But this also does:
>
> [News]# df -h /
> Filesystem Size Used Avail Use% Mounted on
> /dev/root 25G 16G 8.3G 65% /

You have /dev/root mounted on /, not /dev/sda3.

What does 'ls -l /dev/sda3 /dev/root' show you?
I'd say these are both device files with the same
major and minor numbers… Like this:

# ls -l /dev/sdf1 /dev/foo
brw-rw---- 1 root root 8, 81 Nov 30 13:04 /dev/foo
brw-rw---- 1 root disk 8, 81 Nov 28 19:21 /dev/sdf1
# # mount | grep ' /boot '
/dev/sdf1 on /boot type ext4 (rw,relatime)
# # df -h /dev/sdf1
Filesystem Size Used Avail Use% Mounted on
/dev/sdf1 487M 163M 295M 36% /boot
# # df -h /dev/foo
Filesystem Size Used Avail Use% Mounted on
devtmpfs 4.0M 8.0K 4.0M 1% /dev
# # umount /boot
# mount /dev/foo /boot
# # mount | grep ' /boot '
/dev/foo on /boot type ext4 (rw,relatime)
# # df -h /dev/sdf1
Filesystem Size Used Avail Use% Mounted on
devtmpfs 4.0M 8.0K 4.0M 1% /dev
# # df -h /dev/foo
Filesystem Size Used Avail Use% Mounted on
/dev/foo 487M 163M 295M 36% /boot
#

--
mrg

Diego Garcia <dg@chaos.info> writes:
> On Mon, 29 Nov 2021 17:19:36 -0700, Grant Taylor wrote:
>> Try /just/ "df -h" and see if it shows your root file system. I expect
>> that it will.
>>
>
> It does:
>
> [~]# df -h
> Filesystem Size Used Avail Use% Mounted on
> /dev/root 25G 16G 8.3G 65% /
^^^^^^^^^

This is why. df can’t find a filesystem corresponding to /dev/sda3,
because / is mounted on /dev/root. That ultimately resolves to the same
thing but not in a way that df can see through.

--
https://www.greenend.org.uk/rjk/

Diego Garcia <dg@chaos.info> writes:
> Aragorn wrote:
>> I don't know what distribution the OP is using,
>
> A distro, in my view, should only be a stepping stone toward creating ones
> own standalone GNU/Linux system.
>
> Linux is a lot simpler than it may appear and if one not be dependent on
> a distro after a few years of using and learning Linux.
>
> Every user should have his own personal distro.

That would be impractical for most users and a ridiculous waste of time
and effort for the rest.

--
https://www.greenend.org.uk/rjk/

On 30/11/2021 00.40, Diego Garcia wrote:

Subject is a bit misleading... those my original post didn't include the
following:

> I just recently transitioned from static device nodes (simple and efficient) to dynamic
> nodes (complex and pointless) using e/udev (not because of desire but because
> of necessity due to the libinput fiasco wrought by GNOME/freedesktop).
>
> My root directory is on partition /dev/sda3.
>
> When I execute " df -h /dev/sda3" I get the message:
>
> devtmpfs 7.8G 0 7.8G 0% /dev

/dev/sda3 ain't mounted, you have /dev/root mounted, sure they are the
same partition this time. The /dev/root was mounted by the initramfs for
your kernel as there was a lack of root= option in the boot loader for
your kernel.

Add root=/dev/sda3 to the kernel options and you should have /dev/sda3
mounted instead of /dev/root.

This is unrelated to the redhat gnome gang.

--

//Aho

On 30/11/2021 12:40, Richard Kettlewell wrote:
> Diego Garcia <dg@chaos.info> writes:
>> Aragorn wrote:
>>> I don't know what distribution the OP is using,
>>
>> A distro, in my view, should only be a stepping stone toward creating ones
>> own standalone GNU/Linux system.
>>
>> Linux is a lot simpler than it may appear and if one not be dependent on
>> a distro after a few years of using and learning Linux.
>>
>> Every user should have his own personal distro.
>
> That would be impractical for most users and a ridiculous waste of time
> and effort for the rest.
>
+ 1000

Yes I customise my desktop, no it is not a 'new distro'

--
“Politics is the art of looking for trouble, finding it everywhere,
diagnosing it incorrectly and applying the wrong remedies.”
― Groucho Marx

On Tue, 30 Nov 2021 12:40:56 +0000, Richard Kettlewell wrote:

>>
>> Every user should have his own personal distro.
>
> That would be impractical for most users and a ridiculous waste of time
> and effort for the rest.
>

Not so.

Once the initial phase of customization is complete, maintenance
becomes very easy provided one uses an automated build system such
as Gentoo Portage or Pauludis which will keep track of all dependencies.

The Linux kernel is very stable and the user only has to be troubled
to enter "make oldconfig" at the keyboard with each new release.

I devote about 30 minutes maximum each week in maintaining my custom
"distro."

There is no real argument against doing this.

The only exception is when some rouge, renegade, psuedo-authorithy
like RedHat, GNOME, or freedesktop.org decides to break the stability
with some new-fangled "improvement" and to abandon support for all other
options.

That's exactly what happened with libinput. The user then has to scramble
to accommodate these tyrannanical changes.

On 30/11/2021 13.40, Richard Kettlewell wrote:
> Diego Garcia <dg@chaos.info> writes:
>> Aragorn wrote:
>>> I don't know what distribution the OP is using,
>>
>> A distro, in my view, should only be a stepping stone toward creating ones
>> own standalone GNU/Linux system.
>>
>> Linux is a lot simpler than it may appear and if one not be dependent on
>> a distro after a few years of using and learning Linux.
>>
>> Every user should have his own personal distro.
>
> That would be impractical for most users and a ridiculous waste of time
> and effort for the rest.
>

Yep.

--
Cheers, Carlos.

On Tue, 30 Nov 2021 03:49:50 +0100, Aragorn wrote:

>
> eudev, which is a standalone udev fork created by Gentoo
> because of earlier (meanwhile proven-false) concerns that the mainstream
> udev — which now comes bundled as part of systemd — would no longer be
> supporting distributions not based upon systemd.
>

That's not true. The reason for the eudev fork was that non-glibc
systems were having problems. Read this from Gentoo news:

"The integration of udev into the systemd git repo introduced numerous
problems for non-glibc systems, such as musl and uclibc. Several
options were considered, and the one chosen was to fork and maintain udev
independent of the rest of systemd."

These problems have been recently eliminated but the freedom-of-choice
community still has strong suspicions about the motives of systemd
developers.

--
Scratch your technical itch:
https://www.linuxfromscratch.org/

On Tue, 30 Nov 2021 13:06:54 +0100, marrgol wrote:

>
> What does 'ls -l /dev/sda3 /dev/root' show you?
> I'd say these are both device files with the same
> major and minor numbers… Like this:
>

[~]# ls -l /dev/sda3 /dev/root
ls: cannot access '/dev/root': No such file or directory
brw-rw---- 1 root disk 8, 3 Nov 30 17:15 /dev/sda3

--
Scratch your technical itch:
https://www.linuxfromscratch.org/