> > I should note that when I tried to use a sun disk lable or a BSD disk
> > label to get 8 partitions ( which is wrong, sun has 16 partitions but
> > Debian or fdisk/parted can't do it .. weird ) I ended up with a
> > unbootable system everytime. Set a partition bootable got me no
> > where.
>
> No, it wouldn't, Linux has never made use of this flag.
oh .. well that would do it :-\
> >
> > Had to resort to DOS style partition table.
>
> If you have a need for a different partition system, that marks you as
> a fairly unusual user.
possibly even bizarre ! ;-)
Seriously, I come from a long history with Sun hardware and before
that, every other damn thing. Even IBM 3090 mainframes etc etc.
> Presumably you are multi-booting with something
> a bit exotic.
Not at all .. just normal regular boot and run.
> I've never used anything but a DOS (actually IBM, 'why
> would anyone ever want more than four operating systems on their
> mainframe?') partition system, but I've never needed compatibility with
> anything but Windows, nor more than 16 partitions, and I've never moved
> on from fdisk.
I think parted is as fancy as I get. the gnu parted gui can be nice if one
needs to have others looking over your shoulder.
> > Then wireless setup on the pefectly supported ath9k driver was a
> > nightmare for a day.
>
> Wireless can be entertaining, though a lot less so than it used to
> be. Even the Network Manager seems to more or less work these days.
The real bonkers bit for me is that wpa_supplicant throws an error message
at boot, every time, yet I still get dhcp working and here I am .. with wlan0
working just fine. No clue what black magic voodoo goes on there.
> >
> > Finally up and running yesterday on wheezy on this laptop and with
> > minimal packages.
> >
> > I would say that the process is still very very far out of reach of
> > the average user and Linux won't be getting into the mainstream any
> > time soon. This is what is a major blockage to getting linux into an
> > office. Users have no clue if it isn't windows and usualy, no clue
> > even if it is.
> >
> Well, to be honest, most Linux users don't care too much about that.
Yep. Nor do I.
> Linux is where it is without ever having needed widespread desktop use,
> it's not likely to wither away now if it doesn't get there. Widespread
> use would result in better driver access, but it would also result in
> the same kind of malware nightmare that afflicts Windows.
Strictly a philosophy point, but I don't know about that. I would think
that the nature of the Linux kernel would protect users from malware
quite neatly. Also the basic security model of UNIX/Linux protects the
system from one user running riot through /usr and wrecking the
system. An ordinary user can not just delete executables and even
a malicious user can't run riot with memory usage .. I tested that
with a nasty malloc loop and sure enough, the OOM steps in a kills
the process when swap runs out :
Jan 21 00:28:25 mars kernel: [35736.905848] maxalloc invoked oom-killer:
gfp_mask=0x280da, order=0, oom_adj=0, oom_score_adj=0
Jan 21 00:28:25 mars kernel: [35736.905860] maxalloc cpuset=/ mems_allowed=0
Jan 21 00:28:25 mars kernel: [35736.905869] Pid: 7160, comm: maxalloc Tainted:
G C O 3.2.0-4-amd64 #1 Debian 3.2.35-2
Jan 21 00:28:25 mars kernel: [35736.905874] Call Trace:
Jan 21 00:28:25 mars kernel: [35736.905889] [<ffffffff810b693e>] ?
dump_header+0x78/0x1bd
Jan 21 00:28:25 mars kernel: [35736.905899] [<ffffffff81097702>] ?
delayacct_end+0x72/0x7d
Jan 21 00:28:25 mars kernel: [35736.905908] [<ffffffff81163aa2>] ?
security_real_capable_noaudit+0x40/0x4f
Jan 21 00:28:25 mars kernel: [35736.905915] [<ffffffff810b6d07>] ?
oom_kill_process+0x49/0x271
Jan 21 00:28:25 mars kernel: [35736.905922] [<ffffffff810b7402>] ?
out_of_memory+0x2ea/0x337
Jan 21 00:28:25 mars kernel: [35736.905930] [<ffffffff810bb068>] ?
__alloc_pages_nodemask+0x629/0x7aa
Jan 21 00:28:25 mars kernel: [35736.905940] [<ffffffff8134bd27>] ?
_cond_resched+0x7/0x1c
Jan 21 00:28:25 mars kernel: [35736.905949] [<ffffffff810e59f0>] ?
alloc_pages_vma+0x12d/0x136
Jan 21 00:28:25 mars kernel: [35736.905957] [<ffffffff810d05ed>] ?
handle_pte_fault+0x165/0x79f
Jan 21 00:28:25 mars kernel: [35736.905965] [<ffffffff810cdcd1>] ?
pte_offset_kernel+0x16/0x35
Jan 21 00:28:25 mars kernel: [35736.905973] [<ffffffff8134fffd>] ?
do_page_fault+0x312/0x337
Jan 21 00:28:25 mars kernel: [35736.905980] [<ffffffff8134d5f5>] ?
page_fault+0x25/0x30
Jan 21 00:28:25 mars kernel: [35736.905985] Mem-Info:
Jan 21 00:28:25 mars kernel: [35736.905989] Node 0 DMA per-cpu:
Jan 21 00:28:25 mars kernel: [35736.905995] CPU 0: hi: 0, btch: 1 usd:
0
Jan 21 00:28:25 mars kernel: [35736.905999] CPU 1: hi: 0, btch: 1 usd:
0
Jan 21 00:28:25 mars kernel: [35736.906003] Node 0 DMA32 per-cpu:
Jan 21 00:28:25 mars kernel: [35736.906009] CPU 0: hi: 186, btch: 31 usd:
0
Jan 21 00:28:25 mars kernel: [35736.906013] CPU 1: hi: 186, btch: 31 usd:
0
Jan 21 00:28:25 mars kernel: [35736.906017] Node 0 Normal per-cpu:
Jan 21 00:28:25 mars kernel: [35736.906037] active_anon:1135388
inactive_anon:247310 isolated_anon:0
Jan 21 00:28:25 mars kernel: [35736.906040] active_file:20 inactive_file:42
isolated_file:0
Jan 21 00:28:25 mars kernel: [35736.906042] unevictable:0 dirty:0 writeback:0
unstable:0
Jan 21 00:28:25 mars kernel: [35736.906045] free:22961 slab_reclaimable:2610
slab_unreclaimable:4947
Jan 21 00:28:25 mars kernel: [35736.906048] mapped:1122 shmem:1116
pagetables:8054 bounce:0
Jan 21 00:28:25 mars kernel: [35736.906053] Node 0 DMA free:15912kB min:180kB
low:224kB high:268kB active_anon:0kB inactive_anon:0kB active_file:0kB
inactive_file:0kB unevictable:0kB isolated(anon):0kB isolated(file):0kB
present:15688kB mlocked:0kB dirty:0kB writeback:0kB mapped:0kB shmem:0kB
slab_reclaimable:0kB slab_unreclaimable:0kB kernel_stack:0kB pagetables:0kB
unstable:0kB bounce:0kB writeback_tmp:0kB pages_scanned:0 all_unreclaimable? yes
Jan 21 00:28:25 mars kernel: [35736.906076] lowmem_reserve[]: 0 3507 5637 5637
Jan 21 00:28:25 mars kernel: [35736.906084] Node 0 DMA32 free:50456kB
min:41928kB low:52408kB high:62892kB active_anon:2925096kB
inactive_anon:585028kB active_file:12kB inactive_file:8kB unevictable:0kB
isolated(anon):0kB isolated(file):0kB present:3591684kB mlocked:0kB dirty:0kB
writeback:0kB mapped:924kB shmem:916kB slab_reclaimable:1904kB
slab_unreclaimable:1936kB kernel_stack:144kB pagetables:11904kB unstable:0kB
bounce:0kB writeback_tmp:0kB pages_scanned:2648 all_unreclaimable? yes
Jan 21 00:28:25 mars kernel: [35736.906108] lowmem_reserve[]: 0 0 2130 2130
Jan 21 00:28:25 mars kernel: [35736.906115] Node 0 Normal free:25476kB
min:25468kB low:31832kB high:38200kB active_anon:1616456kB
inactive_anon:404212kB active_file:68kB inactive_file:160kB unevictable:0kB
isolated(anon):0kB isolated(file):0kB present:2181600kB mlocked:0kB dirty:0kB
writeback:0kB mapped:3564kB shmem:3548kB slab_reclaimable:8536kB
slab_unreclaimable:17852kB kernel_stack:1848kB pagetables:20312kB unstable:0kB
bounce:0kB writeback_tmp:0kB pages_scanned:352 all_unreclaimable? yes
.
.
.
etc etc
There is now even a cool feature :
if ( mallopt( M_CHECK_ACTION, mallop ) != 1 ) {
fprintf( stderr, "FAIL : %s at line %i : mallopt() failed\n",
argv[0], __LINE__ );
exit( EXIT_FAILURE );
}
printf ( " NOTE : M_CHECK_ACTION set to %i\n", mallop );
When did mallopt arrive on the scen ? Don't know but it sure can give great
data
on the bad stuff users and software may do.
If you wnt the source it is at the bottom of this email.
So I would disagree and say that widespread usage of Linux would not suffer the
same fate as Windows in terms of endless malware and virus's.
most likely a very different fate.
It would end up on cell phones everywhere as something called Android.
> There are too
> many home and business users who are just too important to use
> unprivileged accounts, or to be bothered elevating privilege on just
> the few occasions when it is necessary. This is changing, but very
> slowly. I help out a bit on an MS technical forum, and some people need
> beating with sticks to persuade them that running as root is a bad idea.
Personally I want to invent the taser keyboard. A simple command from the
sysadmin and the user gets an education.
> And as always, it's worth remembering that few people ever install
> Windows.
OKay .. never thought of that.
> I've only done it twice in the last five or six years. It's
> generally pretty difficult and unrewarding to install Windows on a PC
> that isn't very recent, and if it is recent, the latest Windows will be
> there already.
I did buy Windows 7 Ultimate 64-bit ( because why not? ) and then tried
to run it in VMware. What a pig.
> It would be reasonable to expect hypothetical Linux
> business computers to be sold with Linux already installed, and tailored
> to the machine.
I think these are called Samsung Galaxy cell phones. :-)
> The sticking point is that LibreOffice is never going
> to match the slickness of office software developed over more than
> twenty years, and there's still nothing Open Source that comes close to
> Access.
Well, I have managed to get along well in life without ever creating a Word
document after
MS Office 97. I generally look at spreadsheets with disdain and try to figure
out what the
hell is so cool about them. USers are forever getting in and changing a
spreadsheet
in one department that another department needs and then borking it up.
Thus is what business software is supposed to do ... isn't it?
The average business user today is too young to recall when software ran on some
big iron in the back and you could not simply mess with it.
> OK, I've had plenty of Linux installation problems, mostly concerning
> sound, but life is a lot easier now.
My Creative live cards sits unused. Sadly. Never could figure that out.
> It's still never possible to
> predict what may cause trouble. If you're not familiar with it, Knoppix
> is a Debian-based live Linux which has always had amazing hardware
> detection and drivers, and it is often worth booting the latest Knoppix
> to see how it deals with a 'difficult' piece of hardware.
Used it many times.
> Live CDs/USBs
> are mostly also installable, and this is a fairly painless way of
> installing something after you already know it drives your hardware.
This is how I managed to install wheezy. I created a live usb stick.
> Knoppix is not a good candidate for this, as it isn't maintainable, but
> there's a live Debian if you're not too keen on Ubuntu, and plenty of
> other choices. I've done this a couple of times, and haven't needed any
> manual surgery on the target hard drive. Back to the unusual partition
> tables...
Well, I have seen plenty of partition tables and to me it is just a simple
list of thing number foo and cylinder x to cylinder y. How damn hard
is that ? Trivial really. Unless the hardware is on fancy fibre arrays
with many many disks and Brocades etc but generally we use management
software for those. A single SAS disk or hardware mirror should just be
able to deal with any partition table that is well known and then the
BIOS/firmware needs the brains to look at it and ask :
Is there a partition table I recognize on the defined boot disk ?
Yes.
Is there a bootable partition in the list ?
Yes.
So load in the boot record off the damn cylinder listed first and then exec
it !
I think this is what GRUB is all about .. but .. I'm guessing. It has been
years
since I looked at the source code to GRUB2.
dc
--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/fb6b89231405.50fd9...@blastwave.org
