Re: Announce: modutils 2.3.11 is available - the debugger's helper

From: Khimenko Victor (khim@sch57.msk.ru)
Date: Fri Apr 21 2000 - 17:50:37 EDT

  • Next message: willy@thepuffingroup.com: "Re: return values - 32, 64, 128 bits, type defs...etc/platform size"

    In <200004212053.e3LKrG505211@vindaloo.ras.ucalgary.ca> Richard Gooch (rgooch@ras.ucalgary.ca) wrote:
    > willy@thepuffingroup.com writes:
    >> On Fri, Apr 21, 2000 at 01:41:54PM -0600, Richard Gooch wrote:
    >> > willy@thepuffingroup.com writes:
    >> > > On Fri, Apr 21, 2000 at 11:00:26AM -0500, Stefan Monnier wrote:
    >> > > > >>>>> "Keith" == Keith Owens <kaos@ocs.com.au> writes:
    >> > > > > * Add ieee1394 directory, requested by Andreas Bombe.
    >> > > >
    >> > > > I know this has been mentioned before, but I'd like to reiterate
    >> > > > my request for a way to just look in *every* subdirectory.
    >> > > > I does not have to be the default, but it should be at least possible
    >> > > > to get this behavior with some /etc/modules.conf editing.
    >> > >
    >> > > I'm not entirely sure why the modules are segregated into separate
    >> > > directories anyway. The namespace collision already happens at
    >> > > build time (when they're all symlinked into /usr/src/linux/modules)
    >> > > and modprobe is invoked (by me anyway) with just the name of the
    >> > > module, so why pretend it's not a flat namespace? Eschew
    >> > > obfuscation.
    >> >
    >> > Because a flat directory is harder to view for the human. When I ask
    >> > myself "do I have the ne2k-pci module", it's obvious to just look in
    >> > the "net" directory.
    >>
    >> surely
    >>
    >> ls *ne2k*
    >>
    >> is not too hard for someone who wants to know the answer to that
    >> question?

    > So then I'll change the question to "what network drivers do I have".
    > :-)

    And HOW THIS will help you to find out if you have driver for
    "Z8530 based HDLC cards" (scc.o) or "Packet Engines G-NIC PCI Gigabit
    Ethernet adapter" (yellowfin.o) ? And WHERE my COM20020 arcnet driver at
    all ??? I'm pretty sure I answered "compile as module" when was asked ...
    -- cut --
    $ ls /lib/modules/2.3.36/net
    3c501.o
    3c503.o
    3c505.o
    3c507.o
    3c509.o
    3c515.o
    3c523.o
    3c527.o
    3c59x.o
    6pack.o
    82596.o
    8390.o
    ac3200.o
    acenic.o
    aironet4500_card.o
    aironet4500_core.o
    aironet4500_proc.o
    arlan-proc.o
    arlan.o
    at1700.o
    baycom_epp.o
    baycom_par.o
    baycom_ser_fdx.o
    baycom_ser_hdx.o
    bpqether.o
    bsd_comp.o
    cops.o
    cs89x0.o
    de4x5.o
    de600.o
    de620.o
    depca.o
    dgrs.o
    dmascc.o
    dmfe.o
    dummy.o
    e2100.o
    eepro.o
    eepro100.o
    eexpress.o
    epic100.o
    eql.o
    es3210.o
    eth16i.o
    ethertap.o
    ewrk3.o
    fmv18x.o
    hdlcdrv.o
    hp-plus.o
    hp.o
    hp100.o
    ipddp.o
    ircomm.o
    irda.o
    irda_deflate.o
    irlan.o
    lance.o
    lne390.o
    ltpc.o
    mkiss.o
    ne.o
    ne2.o
    ne2k-pci.o
    ne3210.o
    ni5010.o
    ni52.o
    ni65.o
    pcnet32.o
    plip.o
    ppp_async.o
    ppp_deflate.o
    ppp_generic.o
    ppp_synctty.o
    rcpci.o
    rrunner.o
    rtl8139.o
    sb1000.o
    scc.o
    seeq8005.o
    sis900.o
    sk98lin.o
    sk_mca.o
    slip.o
    smc-mca.o
    smc-ultra.o
    smc-ultra32.o
    smc9194.o
    soundmodem.o
    starfire.o
    strip.o
    tlan.o
    tulip.o
    via-rhine.o
    wavelan.o
    wd.o
    yam.o
    yellowfin.o
    -- cut --
    And now my IDE-SCSI emulation does not work. But where to look: in
    /lib/modules/2.3.36/block or /lib/modules/2.3.36/scsi ? And why my IDE cd does
    not have driver in /lib/modules/kernel/2.3.36/cdrom ??? Ahh.
    It's in /lib/modules/2.3.36/block (.../ide in recent kernels).
    Then IPX & AX.25. It's network protocols for sure and PPP is also
    network protocol... Then WHY ipx.o is in /lib/modules/kernel/2.3.36/misc
    while ppp*.o are in /lib/modules/kernel/2.3.36/net ??? Hmm. Is it really
    helpfull for anyone ??? Sorry. This mess ALWAYS created more problems then
    solutions for me...

    >> > The subdirectories are a natural categorisation
    >> > that makes life easier. It's not about pretending there is a
    >> > hierarchical namespace.
    >>
    >> but it makes life somewhat more complex than necessary for the
    >> kernel build scripts and for modutils.

    > The kernel build scripts already have this "complexity" (which isn't
    > really that much), and modutils do too. In fact, by just scanning all
    > directories, modutils can be simplified. So we have nothing to lose.

    Of course if I'll be able to find out list of supported networks cards with
    ls it'll be usefull. But I can not: names of files do not resemble names of
    supported chipsets in A LOT OF cases and not all drivers are in /net directory.
    So I need to scan documentation anyway. What the benefit ?

    > And at the end of the day, computers are supposed to make things
    > easier for humans. I don't care if if takes me 5 milliseconds longer
    > to compile and install my kernel modules, if it makes my life a bit
    > easier.

    In which way ?

    > I can't see why we would change things in the direction of making it
    > harder for humans, since we already have a reasonable system
    > implemented. It's not like we need to invest days or weeks of coding
    > effort.

    For me `find /lib/modules/<version> modulename.o` always looked
    longer then `find /lib/modules/<version> modulename.o` so I can not
    see benefits.

    -
    To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
    the body of a message to majordomo@vger.rutgers.edu
    Please read the FAQ at http://www.tux.org/lkml/



    This archive was generated by hypermail 2b29 : Fri Apr 21 2000 - 18:00:01 EDT