__/ [ The Ghost In The Machine ] on Tuesday 29 August 2006 18:00 \__
> In comp.os.linux.advocacy, Stuart Krivis
> <jd@xxxxxxxxxxxxxxxx>
> wrote
> on Mon, 28 Aug 2006 15:34:16 -0400
> <18h6f29akchnjsilkt2qs2ggsph6scfsah@xxxxxxx>:
>> On Mon, 28 Aug 2006 15:27:47 -0400, Stuart Krivis
>> <jd@xxxxxxxxxxxxxxxx> wrote:
>>
>>>On Thu, 24 Aug 2006 18:00:18 GMT, The Ghost In The Machine
>>><ewill@xxxxxxxxxxxxxxxxxxxxxxx> wrote:
>>>
>>>>And only if the user accepts them. However, kernel updates
>>>>aren't all that rare.
>>>>
>>>>This is the ChangeLog list from
>>>>http://www.kernel.org/pub/linux/kernel/v2.6 .
>>>>
>>>
>>><snip>
>>>
>>>
>>>>
>>>>94 entries between 2003-12-18 and 2006-08-23, a span of 979 days.
>>>>This is one kernel patch every 10 1/2 days, give or take.
>>>>
>>>>I don't know if this is good or bad -- probably not all
>>>>that bad. It's one metric among many, and I frankly don't
>>>>know how to measure the Quality Problem(tm).
>>>
>>>I don't think that all 94 have applied to all users though. I've got
>>>an uptime on my SuSE 10 box right now of 78 days. So I'm obviously not
>>>seeing an update for every kernel patch.
>>
>> My mistake. I looked at the wrong machine. :-)
>>
>> My SuSE 10 box has actually been up for only 27 days.
Not even Linux can endure power outages. *smile*
Personally, I give KDE a kick every month or two, due to aggregated memory
leaks (Thunderbird 1.0.x is a culprit). This does not require a reboot and
KDE restores sessions perfectly.
> This is true...at least for Linux the latest kernel is
> the only one really needed. One can, presumably, take
> the base version (2.6.0) and apply *all* the patches if
> one is a masochist -- or a pedant. :-)
Many of the so-called important kernel patches address DDOS vulnerabilities,
which can be exploited locally, e.g. by mischievous users.
> Me, I'm not sure I want to bother, and the number
> of issues on Linux the kernel are not all that big.
> Linux applications get a fair amount of patch traffic
> but patching them doesn't need a reboot of the entire
> system -- at worst, one gets an outage of that service.
> However, there is an interesting exception for things such
> as the X server, which for me was down for 2-3 hours as I
> switched from 6.8.2 to 7.1. But I for one would consider
> that slightly unusual anyway, and X isn't really needed
> for servers.
>
> (It's my work machine; I don't get to patch it that often.)
I haven't patched mine for a while. If I recall correctly, a Linux kernel
cycle that's concerned with bogfixes (recent decision made by Morton and
Torvalds) addresses hardware incompatibilities with very old hardware.
Compare that with loopholes that compromise the entire system, _remotely_.
The concern would not be 15-year-old keyboards, which in Windows, for
example, would probably require drivers that complement the kernel's remit.
That in itself is an issue because third-party drivers cannot be trusted.
Intel's driver recently leaked memory in Windows and, in some cases, these
compromise the system too. Also think of AMD's recent rebuttal to an alleged
64-bit issue at chip-level.
Best wishes,
Roy
--
Roy S. Schestowitz | make install -not war
http://Schestowitz.com | Open Prospects ¦ PGP-Key: 0x74572E8E
Tasks: 125 total, 2 running, 122 sleeping, 0 stopped, 1 zombie
http://iuron.com - knowledge engine, not a search engine
|
|
