eudev hwdb?

eudev hwdb?

[Gary Thomas]

I see that the replacement for udev (eudev) in OE-core takes nearly
12MB on my i.MX6 (ARM CortexA7).  This is nearly 25% of my total
storage space (64MB NAND)!

# ls -lR /etc/udev
/etc/udev:
total 6520
-rw-r--r-- 1 root root    1600 Jan  1  1970 cache.data
-r--r--r-- 1 root root 6660476 Apr 26 13:08 hwdb.bin
drwxr-xr-x 2 root root    1296 Apr 26 11:34 hwdb.d
-rw-r--r-- 1 root root      51 Apr 26 11:31 mount.blacklist
drwxr-xr-x 2 root root     160 Apr 26 11:35 mount.blacklist.d
drwxr-xr-x 2 root root     608 Apr 26 13:08 rules.d
drwxr-xr-x 2 root root     304 Apr 26 11:35 scripts
-rw-r--r-- 1 root root      49 Apr 26 11:34 udev.conf

/etc/udev/hwdb.d:
total 5224
-rw-r--r-- 1 root root 1292931 Apr 26 11:34 20-OUI.hwdb
-rw-r--r-- 1 root root  126596 Apr 26 11:34 20-acpi-vendor.hwdb
-rw-r--r-- 1 root root   41289 Apr 26 11:34 20-bluetooth-vendor-product.hwdb
-rw-r--r-- 1 root root     111 Apr 26 11:34 20-net-ifname.hwdb
-rw-r--r-- 1 root root   13710 Apr 26 11:34 20-pci-classes.hwdb
-rw-r--r-- 1 root root 2648842 Apr 26 11:34 20-pci-vendor-model.hwdb
-rw-r--r-- 1 root root     783 Apr 26 11:34 20-sdio-classes.hwdb
-rw-r--r-- 1 root root    4067 Apr 26 11:34 20-sdio-vendor-model.hwdb
-rw-r--r-- 1 root root    8070 Apr 26 11:34 20-usb-classes.hwdb
-rw-r--r-- 1 root root 1113001 Apr 26 11:34 20-usb-vendor-model.hwdb
-rw-r--r-- 1 root root    3702 Apr 26 11:34 60-evdev.hwdb
-rw-r--r-- 1 root root   52967 Apr 26 11:34 60-keyboard.hwdb
-rw-r--r-- 1 root root   14217 Apr 26 11:34 70-mouse.hwdb
-rw-r--r-- 1 root root    4627 Apr 26 11:34 70-pointingstick.hwdb

/etc/udev/mount.blacklist.d:
total 0

/etc/udev/rules.d:
total 20
-rw-r--r-- 1 root root 1449 Apr 26 11:31 10-imx.rules
-rw-r--r-- 1 root root    0 Apr 26 11:34 80-net-name-slot.rules
-rw-r--r-- 1 root root  847 Apr 26 11:31 automount.rules
-rw-r--r-- 1 root root  757 Apr 26 11:31 autonet.rules
-rw-r--r-- 1 root root  885 Apr 26 11:34 local.rules
-rw-r--r-- 1 root root  843 Apr 26 11:31 localextra.rules

/etc/udev/scripts:
total 8
-rwxr-xr-x 1 root root 2469 Apr 26 11:31 mount.sh
-rwxr-xr-x 1 root root 1402 Apr 26 11:31 network.sh

Is there any way to not have both the hwdb.bin (which I assume
is a binary version of the hardware databases) and /etc/udev/hwdb.d?
Perhaps there is some way to only use one?

Or maybe there is a better choice, such as mdev?  I'm not sure about
this (no experience) but my system needs to handle plug&play devices
which the current eudev does well.  It's just a huge "price" to pay.

Any ideas or pointers would be great

-- 
------------------------------------------------------------
Gary Thomas                 |  Consulting for the
MLB Associates              |    Embedded world
------------------------------------------------------------

Re: eudev hwdb?

[Burton, Ross]

[-- Attachment #1: Type: text/plain, Size: 805 bytes --]

On 27 April 2016 at 09:48, Gary Thomas <gary@mlbassoc.com> wrote:

> Is there any way to not have both the hwdb.bin (which I assume
> is a binary version of the hardware databases) and /etc/udev/hwdb.d?
> Perhaps there is some way to only use one?
>
> Or maybe there is a better choice, such as mdev?  I'm not sure about
> this (no experience) but my system needs to handle plug&play devices
> which the current eudev does well.  It's just a huge "price" to pay.
>

You could not install hwdb at all, it's a recommends so you can
BAD_RECOMMENDS it away.  It's certainly possible to wipe away the source
files for hwdb if you don't care about extending them at runtime but nobody
has done that yet: a rootfs postproccess hook could delete it and confirm
that everything works fine.

Ross

[-- Attachment #2: Type: text/html, Size: 1233 bytes --]

Re: eudev hwdb?

[Yegor Yefremov]

Hi Gary,

On Wed, Apr 27, 2016 at 10:48 AM, Gary Thomas <gary@mlbassoc.com> wrote:
> I see that the replacement for udev (eudev) in OE-core takes nearly
> 12MB on my i.MX6 (ARM CortexA7).  This is nearly 25% of my total
> storage space (64MB NAND)!
>
> # ls -lR /etc/udev
> /etc/udev:
> total 6520
> -rw-r--r-- 1 root root    1600 Jan  1  1970 cache.data
> -r--r--r-- 1 root root 6660476 Apr 26 13:08 hwdb.bin
> drwxr-xr-x 2 root root    1296 Apr 26 11:34 hwdb.d
> -rw-r--r-- 1 root root      51 Apr 26 11:31 mount.blacklist
> drwxr-xr-x 2 root root     160 Apr 26 11:35 mount.blacklist.d
> drwxr-xr-x 2 root root     608 Apr 26 13:08 rules.d
> drwxr-xr-x 2 root root     304 Apr 26 11:35 scripts
> -rw-r--r-- 1 root root      49 Apr 26 11:34 udev.conf
>
> /etc/udev/hwdb.d:
> total 5224
> -rw-r--r-- 1 root root 1292931 Apr 26 11:34 20-OUI.hwdb
> -rw-r--r-- 1 root root  126596 Apr 26 11:34 20-acpi-vendor.hwdb
> -rw-r--r-- 1 root root   41289 Apr 26 11:34 20-bluetooth-vendor-product.hwdb
> -rw-r--r-- 1 root root     111 Apr 26 11:34 20-net-ifname.hwdb
> -rw-r--r-- 1 root root   13710 Apr 26 11:34 20-pci-classes.hwdb
> -rw-r--r-- 1 root root 2648842 Apr 26 11:34 20-pci-vendor-model.hwdb
> -rw-r--r-- 1 root root     783 Apr 26 11:34 20-sdio-classes.hwdb
> -rw-r--r-- 1 root root    4067 Apr 26 11:34 20-sdio-vendor-model.hwdb
> -rw-r--r-- 1 root root    8070 Apr 26 11:34 20-usb-classes.hwdb
> -rw-r--r-- 1 root root 1113001 Apr 26 11:34 20-usb-vendor-model.hwdb
> -rw-r--r-- 1 root root    3702 Apr 26 11:34 60-evdev.hwdb
> -rw-r--r-- 1 root root   52967 Apr 26 11:34 60-keyboard.hwdb
> -rw-r--r-- 1 root root   14217 Apr 26 11:34 70-mouse.hwdb
> -rw-r--r-- 1 root root    4627 Apr 26 11:34 70-pointingstick.hwdb
>
> /etc/udev/mount.blacklist.d:
> total 0
>
> /etc/udev/rules.d:
> total 20
> -rw-r--r-- 1 root root 1449 Apr 26 11:31 10-imx.rules
> -rw-r--r-- 1 root root    0 Apr 26 11:34 80-net-name-slot.rules
> -rw-r--r-- 1 root root  847 Apr 26 11:31 automount.rules
> -rw-r--r-- 1 root root  757 Apr 26 11:31 autonet.rules
> -rw-r--r-- 1 root root  885 Apr 26 11:34 local.rules
> -rw-r--r-- 1 root root  843 Apr 26 11:31 localextra.rules
>
> /etc/udev/scripts:
> total 8
> -rwxr-xr-x 1 root root 2469 Apr 26 11:31 mount.sh
> -rwxr-xr-x 1 root root 1402 Apr 26 11:31 network.sh
>
> Is there any way to not have both the hwdb.bin (which I assume
> is a binary version of the hardware databases) and /etc/udev/hwdb.d?
> Perhaps there is some way to only use one?
>
> Or maybe there is a better choice, such as mdev?  I'm not sure about
> this (no experience) but my system needs to handle plug&play devices
> which the current eudev does well.  It's just a huge "price" to pay.
>
> Any ideas or pointers would be great

eudev provides following configure option:

 --enable-hwdb  install hwdb.d files

this way you can omit hwdb.d installation by providing:

--disable-hwdb

Yegor

Re: eudev hwdb?

[Gary Thomas]

On 2016-04-27 10:59, Burton, Ross wrote:
>
> On 27 April 2016 at 09:48, Gary Thomas <gary@mlbassoc.com <mailto:gary@mlbassoc.com>> wrote:
>
>     Is there any way to not have both the hwdb.bin (which I assume
>     is a binary version of the hardware databases) and /etc/udev/hwdb.d?
>     Perhaps there is some way to only use one?
>
>     Or maybe there is a better choice, such as mdev?  I'm not sure about
>     this (no experience) but my system needs to handle plug&play devices
>     which the current eudev does well.  It's just a huge "price" to pay.
>
>
> You could not install hwdb at all, it's a recommends so you can BAD_RECOMMENDS it away.  It's certainly possible to wipe
> away the source files for hwdb if you don't care about extending them at runtime but nobody has done that yet: a rootfs
> postproccess hook could delete it and confirm that everything works fine.

Can you explain a bit more? - BAD_RECOMMENDS doesn't seem to be documented.

-- 
------------------------------------------------------------
Gary Thomas                 |  Consulting for the
MLB Associates              |    Embedded world
------------------------------------------------------------

Re: eudev hwdb?

[Burton, Ross]

[-- Attachment #1: Type: text/plain, Size: 289 bytes --]

On 27 April 2016 at 10:22, Gary Thomas <gary@mlbassoc.com> wrote:

> Can you explain a bit more? - BAD_RECOMMENDS doesn't seem to be documented.
>

My mistake - BAD_RECOMMENDATIONS.

http://www.yoctoproject.org/docs/2.0.1/ref-manual/ref-manual.html#var-BAD_RECOMMENDATIONS

Ross

[-- Attachment #2: Type: text/html, Size: 845 bytes --]

Re: eudev hwdb?

[Martin Jansa]

[-- Attachment #1: Type: text/plain, Size: 3344 bytes --]

On Wed, Apr 27, 2016 at 10:48:03AM +0200, Gary Thomas wrote:
> I see that the replacement for udev (eudev) in OE-core takes nearly
> 12MB on my i.MX6 (ARM CortexA7).  This is nearly 25% of my total
> storage space (64MB NAND)!

See
http://permalink.gmane.org/gmane.comp.handhelds.openembedded.core/77244
and following discussions

> 
> # ls -lR /etc/udev
> /etc/udev:
> total 6520
> -rw-r--r-- 1 root root    1600 Jan  1  1970 cache.data
> -r--r--r-- 1 root root 6660476 Apr 26 13:08 hwdb.bin
> drwxr-xr-x 2 root root    1296 Apr 26 11:34 hwdb.d
> -rw-r--r-- 1 root root      51 Apr 26 11:31 mount.blacklist
> drwxr-xr-x 2 root root     160 Apr 26 11:35 mount.blacklist.d
> drwxr-xr-x 2 root root     608 Apr 26 13:08 rules.d
> drwxr-xr-x 2 root root     304 Apr 26 11:35 scripts
> -rw-r--r-- 1 root root      49 Apr 26 11:34 udev.conf
> 
> /etc/udev/hwdb.d:
> total 5224
> -rw-r--r-- 1 root root 1292931 Apr 26 11:34 20-OUI.hwdb
> -rw-r--r-- 1 root root  126596 Apr 26 11:34 20-acpi-vendor.hwdb
> -rw-r--r-- 1 root root   41289 Apr 26 11:34 20-bluetooth-vendor-product.hwdb
> -rw-r--r-- 1 root root     111 Apr 26 11:34 20-net-ifname.hwdb
> -rw-r--r-- 1 root root   13710 Apr 26 11:34 20-pci-classes.hwdb
> -rw-r--r-- 1 root root 2648842 Apr 26 11:34 20-pci-vendor-model.hwdb
> -rw-r--r-- 1 root root     783 Apr 26 11:34 20-sdio-classes.hwdb
> -rw-r--r-- 1 root root    4067 Apr 26 11:34 20-sdio-vendor-model.hwdb
> -rw-r--r-- 1 root root    8070 Apr 26 11:34 20-usb-classes.hwdb
> -rw-r--r-- 1 root root 1113001 Apr 26 11:34 20-usb-vendor-model.hwdb
> -rw-r--r-- 1 root root    3702 Apr 26 11:34 60-evdev.hwdb
> -rw-r--r-- 1 root root   52967 Apr 26 11:34 60-keyboard.hwdb
> -rw-r--r-- 1 root root   14217 Apr 26 11:34 70-mouse.hwdb
> -rw-r--r-- 1 root root    4627 Apr 26 11:34 70-pointingstick.hwdb
> 
> /etc/udev/mount.blacklist.d:
> total 0
> 
> /etc/udev/rules.d:
> total 20
> -rw-r--r-- 1 root root 1449 Apr 26 11:31 10-imx.rules
> -rw-r--r-- 1 root root    0 Apr 26 11:34 80-net-name-slot.rules
> -rw-r--r-- 1 root root  847 Apr 26 11:31 automount.rules
> -rw-r--r-- 1 root root  757 Apr 26 11:31 autonet.rules
> -rw-r--r-- 1 root root  885 Apr 26 11:34 local.rules
> -rw-r--r-- 1 root root  843 Apr 26 11:31 localextra.rules
> 
> /etc/udev/scripts:
> total 8
> -rwxr-xr-x 1 root root 2469 Apr 26 11:31 mount.sh
> -rwxr-xr-x 1 root root 1402 Apr 26 11:31 network.sh
> 
> Is there any way to not have both the hwdb.bin (which I assume
> is a binary version of the hardware databases) and /etc/udev/hwdb.d?
> Perhaps there is some way to only use one?
> 
> Or maybe there is a better choice, such as mdev?  I'm not sure about
> this (no experience) but my system needs to handle plug&play devices
> which the current eudev does well.  It's just a huge "price" to pay.
> 
> Any ideas or pointers would be great
> 
> -- 
> ------------------------------------------------------------
> Gary Thomas                 |  Consulting for the
> MLB Associates              |    Embedded world
> ------------------------------------------------------------
> -- 
> _______________________________________________
> yocto mailing list
> yocto@yoctoproject.org
> https://lists.yoctoproject.org/listinfo/yocto

-- 
Martin 'JaMa' Jansa     jabber: Martin.Jansa@gmail.com

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 188 bytes --]

Re: eudev hwdb?

[Gary Thomas]

On 2016-04-27 11:23, Burton, Ross wrote:
>
> On 27 April 2016 at 10:22, Gary Thomas <gary@mlbassoc.com <mailto:gary@mlbassoc.com>> wrote:
>
>     Can you explain a bit more? - BAD_RECOMMENDS doesn't seem to be documented.
>
>
> My mistake - BAD_RECOMMENDATIONS.
>
> http://www.yoctoproject.org/docs/2.0.1/ref-manual/ref-manual.html#var-BAD_RECOMMENDATIONS

Thanks, that worked.  Took away all of the hwdb files (binary and sources)
but my system is working just fine without them (so far)

-- 
------------------------------------------------------------
Gary Thomas                 |  Consulting for the
MLB Associates              |    Embedded world
------------------------------------------------------------