Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 21 Dec 2000 16:29:59 -0800 (PST)
From:      Matt Dillon <dillon@earth.backplane.com>
To:        Dennis <dennis@etinc.com>
Cc:        hackers@FreeBSD.ORG
Subject:   Re: FreeBSD vs Linux, Solaris, and NT
Message-ID:  <200012220029.eBM0TxE82553@earth.backplane.com>
References:  <5.0.0.25.0.20001220192150.01f42450@mail.etinc.com> <5.0.0.25.0.20001221120837.022ab0a0@mail.etinc.com> <5.0.0.25.0.20001221184852.03ceab10@mail.etinc.com>

next in thread | previous in thread | raw e-mail | index | archive | help
:No, the original writer was trying to use a very general argument about the 
:absolute uselessness of binary code, which is disgustingly wrong. Im sure 
:you dont disagree. Your argument is sound only if the manufacturer doesnt 
:implement those "fixes" in their binary drivers, which they usually do. Its 
:also more likely that they will use the correct workarounds and will know 
:about them before they bite end users in the arse, which is usually not the 
:case with "free" drivers typically found in free OSs.
:
:the previous writer used "objdump"  as an example of reverse engineering 
:software, the marketing of which would be illegal. Certainly you can figure 
:out how something works and write an original driver for it, but thats not 
:really reverse engineering to me. its still original code.
:
:Dennis
    
    You are correct about objdump ... that isn't reverse engineering.

    But while I generally agree that there is nothing wrong with binaries,
    I make a big distinction between user-level binaries and kernel-level
    modules.  I think user-level binaries are perfectly acceptable, but
    I have strong reservations in regards to kernel-level binaries.  Kernels
    change all the time... there is no 'API' per-say... at least nothing
    like the relatively stable syscall interface user-level binaries enjoy.
    And as has been pointed out time and time again, the vast majority of
    commercial device driver writers don't know jack about the OS they are
    writing for and proceed to do all sorts of illegal things in the driver
    code. 

    In that respect, I personally will not run anything inside my kernel that
    I don't have source for.  Now, I don't run frame-relay or T1's into
    FreeBSD boxes, so I'm not commenting on your software specifically.  I'm
    commenting in general.  The problem is not only support, but also
    protection against obsolescence.  Companies upgrade their products,
    companies go out of business, companies stop supporting products. 
    Without source you can wind up S.O.L. with a binary-only device driver.
    It's just too risky for me.

    Just look at all the poor windows bozos who are forced to throw away
    half their software every time they upgrade to a new version of Windows
    when Microsoft stops supporting the older releases.  That is not a cycle
    I will ever willingly get into.

						-Matt



To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-hackers" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200012220029.eBM0TxE82553>