Re: Firefox slower on GMail since update to CentOS 5.5

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]



On 5/27/2010 2:47 PM, MHR wrote:
> On Thu, May 27, 2010 at 12:00 PM, Les Mikesell<lesmikesell@xxxxxxxxx>  wrote:
>> On 5/27/2010 1:49 PM, m.roth@xxxxxxxxx wrote:
>>>
>>> Sorry, I don't understand what you wrote. How does Linux make it
>>> "difficult to impossible"? It's an o/s, with POSIX calls, just like all
>>> the other unices. It's not M$, and it's not Apple... so is being neither
>>> making it hard?
>>
>> For one thing, the license terms do not permit including software with
>> differing terms (hence no zfs, etc. even though source is freely
>> available), and for another the interfaces keep changing so binaries
>> can't be expected to work after updates.
>>
>
> I'm not sure if I have the problem or not (don't think so), but I have
> to take issue with this last bit (changing interfaces).
>
> I'm running the nvidia driver
> dkms-nvidia-x11-drv-185.18.14-1.nodist.rf.x86_64 from Oct 22, 2009,
> and IIRC it has survived the 5.4 and 5.5 updates without any problems
> at all.  Perhaps that's because dkms rebuilds the driver for each new
> kernel, but even so, the interfaces can't have changed too much if
> they still build and work.
>
> When I was working on the Linux kernel directly, one of the least
> likely areas of change was the driver interface, precisely because
> changes at this level are detrimental to all new driver development
> and old driver compatibility.  The driver API is fairly well
> established and less likely to change than most other fly-by-night
> (i.e., M$) OSs.

No, dkms was a late add-in to help fix this problem.  It wouldn't be 
necessary if the OS itself provided a stable interface.  The 
'enterprise' versions of Linux also are an attempt to minimize the 
problem by freezing the kernel version for many years and only 
backporting minimal changes.

> Even the proprietary versions of UNIX don't mess with
> this (much).  They may be different, but they aren't too variable.
>
> Thus, I find this claim difficult to believe.
>
> Do you have examples?  Proof?

The nvidia driver is the obvious one.  They've done the best they can, 
Dell added dkms, and people still have trouble.  Most other vendors 
don't even bother.

-- 
   Les Mikesell
    lesmikesell@xxxxxxxxx




_______________________________________________
CentOS mailing list
CentOS@xxxxxxxxxx
http://lists.centos.org/mailman/listinfo/centos


[Index of Archives]     [CentOS]     [CentOS Announce]     [CentOS Development]     [CentOS ARM Devel]     [CentOS Docs]     [CentOS Virtualization]     [Carrier Grade Linux]     [Linux Media]     [Asterisk]     [DCCP]     [Netdev]     [Xorg]     [Linux USB]
  Powered by Linux