Bug#838522: dahdi-linux: Please announce supported hardware using AppStream

[Tzafrir Cohen]

On 21/09/16 23:43, Petter Reinholdtsen wrote:
> Package: dahdi-linux
> Version: 2.11.1-1
> Severity: wishlist
> User: pere at hungry.com
> Usertags: appstream-modalias
>
> Hi.
>
> The dahdi-linux package is one of the packages in the Debian archive
> that should be proposed for installation when a given hardware dongle is
> inserted or available.  Thanks to the appstream system, this can now be
> announced in a way other tools can use and act on.  I've written the
> isenkram system to ask the current user if hardware specific packages
> should be installed when a new dongle is installed or already present on
> a machine, and isenkram now uses appstream as one source for hardware to
> package mappings.
>
> You can read more about this on my blog, 
> <URL: http://people.skolelinux.org/pere/blog/Using_appstream_with_isenkram_to_install_hardware_related_packages_in_Debian.html >.
>
> Instructions on how to create the metadata XML file can be found in
> <URL: https://wiki.debian.org/AppStream/Guidelines >.
>
> It would be great if you could add an appstream metainfo file to the
> dahdi-linux package, with content similar to this:
>
>   <?xml version="1.0" encoding="UTF-8"?>
>   <component>
>    [...]
>    <provides>
> 	<modalias>pci:v0000e159d00000001sv*</modalias>
> 	<modalias>lkmodule:dahdi</modalias>
>     </provides>
>   </component>
>
> If there are other hardware ids also supported by the package, please
> add those too. :)
DAHDI supports mostly PCI devices. There is a single type of USB
devices: the Xorcom Astribanks. The relevant PCI IDs are e4e4:11[3456]2
, although 1132 and 1142 are probably not around anymore, and 1152
should also be quite rate.

As for PCI devices (assuming you want them too):

Looking at /usr/share/misc/pci.ids , I see also:


Basically all of Digium's cards (vendor ID d161) are supported.

e159:0001 is an ISDN modem card. its chip was used by Digium for its
first cards. Thus it should not be used without the sub (vendor/product)
IDs. e159:0001 8086:0003 is the one of the old X100P/X101P or one of its
many clones (a modem reprogrammed as a Zaptel card). It does not show on
the pci.ids list but the wctdm driver (old fxs/fxo card. Not sold
anymore for a long time) uses a bunch of vendor sub-IDs under e159:0001.

static DEFINE_PCI_DEVICE_TABLE(wctdm_pci_tbl) = {
        { 0xe159, 0x0001, 0xa159, PCI_ANY_ID, 0, 0, (unsigned long)
&wctdm },
        { 0xe159, 0x0001, 0xe159, PCI_ANY_ID, 0, 0, (unsigned long)
&wctdm },
        { 0xe159, 0x0001, 0xb100, PCI_ANY_ID, 0, 0, (unsigned long)
&wctdme },
        { 0xe159, 0x0001, 0xb1d9, PCI_ANY_ID, 0, 0, (unsigned long)
&wctdmi },
        { 0xe159, 0x0001, 0xb118, PCI_ANY_ID, 0, 0, (unsigned long)
&wctdmi },
        { 0xe159, 0x0001, 0xb119, PCI_ANY_ID, 0, 0, (unsigned long)
&wctdmi },
        { 0xe159, 0x0001, 0xa9fd, PCI_ANY_ID, 0, 0, (unsigned long)
&wctdmh },
        { 0xe159, 0x0001, 0xa8fd, PCI_ANY_ID, 0, 0, (unsigned long)
&wctdmh },
        { 0xe159, 0x0001, 0xa800, PCI_ANY_ID, 0, 0, (unsigned long)
&wctdmh },
        { 0xe159, 0x0001, 0xa801, PCI_ANY_ID, 0, 0, (unsigned long)
&wctdmh },
        { 0xe159, 0x0001, 0xa908, PCI_ANY_ID, 0, 0, (unsigned long)
&wctdmh },
        { 0xe159, 0x0001, 0xa901, PCI_ANY_ID, 0, 0, (unsigned long)
&wctdmh },

under the vendor ID 10b5 (PLX Technology, Inc) there is 10b5:0557 10b5
9030, the tor2 card, and then near-by a bunch of Atcom (and other)
clones, most of which I believe we don't support. Another one is (the
original tor2 card) 10b5:d00d 10b5:9030 and a bunch of near-by (and
unsupported) clones (that did not notice 'd00d' was supposed to spell
'dude').

-- Tzafrir