Re: [oi-dev] dmidecode package

2021-06-29 Thread Joshua M. Clulow via oi-dev
On Mon, 28 Jun 2021 at 10:06, s...@pandora.be  wrote:
> In order to fix DDU (Device Detection Utility) for the Dell Precision 
> workstation,
> and possibly for other computers with UEFI, perhaps dmidecode could be 
> packaged.
>
> http://savannah.nongnu.org/projects/dmidecode/
>
> This dmidecode utility compiles without problems and it works for me.
>
> It supports UEFI and their support for UEFI seems to work on OpenIndiana.
>
> Once dmidecode is packaged, the DDU probe script could use dmidecode,
> instead of dmi_info.
>
> I'm not familiar enough with DDU nor dmidecode to see whether this is a good 
> idea,
> but it could work I guess.

We actually have a library for dealing with SMBIOS in the OS today:

https://github.com/illumos/illumos-gate/tree/master/usr/src/lib/libsmbios


https://github.com/illumos/illumos-gate/blob/master/usr/src/uts/common/sys/smbios.h

It's used by a core OS command, smbios(1M), to interrogate the SMBIOS
tables for display.  The same code is also used by various core OS
libraries and the kernel as required.

It definitely works on UEFI machines; e.g., on an Intel NUC to which I
have access:

$ smbios -t 2
IDSIZE TYPE
2 115  SMB_TYPE_BASEBOARD (type 2) (base board)

  Manufacturer: Intel Corporation
  Product: NUC6i7KYB
  Version: H90766-405
  Serial Number: [REDACTED]
  Asset Tag:
  Location Tag: Default string

  Chassis: 3
  Flags: 0x9
SMB_BBFL_MOTHERBOARD (board is a motherboard)
SMB_BBFL_REPLACABLE (board is field-replacable)
  Board Type: 0xa (motherboard)

The source for the "dmi_info" command that comes with DDU seems to be here:

https://github.com/OpenIndiana/ddu/blob/oi/hipster/utils/dmi_info/dmi.c

It looks like it would be relatively easy to port it to use libsmbios
from the OS.  You would include  and link with -lsmbios and
you should be able to get going.

Note that there is not currently a strong stability guarantee around
the libsmbios ABI, but if you pass a specific version (e.g.,
SMB_VERSION_34) instead of just the latest version token (i.e.,
SMB_VERSION, as we do in the gate) then things will probably be OK.
In practice, we try pretty hard to keep existing binaries working --
and we may be able to make stronger guarantees and provide manual
pages and so on if there are consumers outside the gate.


Cheers.

-- 
Joshua M. Clulow
http://blog.sysmgr.org

___
oi-dev mailing list
oi-dev@openindiana.org
https://openindiana.org/mailman/listinfo/oi-dev


Re: [oi-dev] OI no longer automounts USB sticks

2021-06-29 Thread s...@pandora.be


When I initially said that I could reproduce the problem,
(I am testing on this completely new 2021.04 system)
I didn't check that rmvolmgr was enabled.

I should have verified that to begin with ...

I can still reproduce the problem that you reported,
but for example with an old Olympus camera SmartMedia card reader,
it is automounted fine as follows:
 
# svcadm restart hal
# svcadm restart rmvolmgr

# eject -l
/dev/dsk/c4t1d0s2cdrom,cdrom0,cd,cd0,sr,sr0
/dev/dsk/c6t0d0p0:1  rmdisk4,USBSLACK,/media/USBSLACK
/dev/dsk/c5t0d1p0:1  rmdisk1,/media/2.0 Reader-SM-xD

Of course restarting should not be necessary but this is a workaround.

David Stes

- Op 28 jun 2021 om 21:52 schreef gary mills gary_mi...@fastmail.fm:

> On Mon, Jun 28, 2021 at 06:53:33PM +0200, s...@pandora.be wrote:
>> 
>> It's good that you reported the issue and of course USB automount
>> is useful.
>> 
>> Andreas Wacknitz also confirmed this, and is trying to help as much
>> as possible.
> 
> I didn't know this.  In fact, I generally don't know when somebody
> else is working on something.  We should be collaborating, rather
> than duplicating effort.
> 
> I'm assuming the problem is in glib.  I have several snapshots of the
> OI source, including the current one of course.  I don't know yet if
> a patch disappeared.  I don't know yet if the glib developers have
> dropped support for solaris or illumos.  Going through the layers,
> I've gotten as far as:
> 
>status = channel->funcs->io_read (channel, channel->read_buf->str + 
> cur_len,
> channel->buf_size, &read_size, err);
> 
> I'm assuming that OS-specific versions of "io_read" exist in glib,
> but I haven't found them yet.
> 
> 
> --
> -Gary Mills-  -refurb--Winnipeg, Manitoba, Canada-
> 
> ___
> oi-dev mailing list
> oi-dev@openindiana.org
> https://openindiana.org/mailman/listinfo/oi-dev

___
oi-dev mailing list
oi-dev@openindiana.org
https://openindiana.org/mailman/listinfo/oi-dev