Re: news 2010-04

[Arne Babenhauserheide]

Hi, 

Am Montag, 3. Mai 2010, 02:23:17 schrieb olafbuddenha...@gmx.net:
> > It is surely very valuable work that Sergio is doing (Go!, Go!, Go!),
> > but we previously agreed to only list items that have already (and
> > recently) been published as patches (or equivalent) -- such that
> > others can use / reproduce / verify it.
> 
> Not sure such a strict policy is really useful?
> 
> In my understanding the major point of the regular news is showing
> people that things are moving in the Hurd world -- which doesn't always
> have to be manifested in patches I'd say...

The reason for that is that the Hurd has a reputation for never getting 
finished, and that reputation is what hurts it most when it comes to adoption 
(and developer influx) — at least from my understanding. Did you ever hear an 
ironic phrase like “It will be released when the Hurd is finished”? 

The news are a direct push at fixing that problem: “This is what the Hurd 
developers accomplished last month” ⇒ ‘A month of the Hurd’. 

It’s not just about showing that things are moving, but about showing that 
things are being accomplished. 

And to maximize the impact of these news, every news item must be verifiable 
by every reader as not being vaporware. So the next time someone says “the 
Hurd will never be finished” people can just point to the news as solid 
evidence (that can’t debate away) which shows that the premonition is false (I 
already saw that happening once). 

That’s also why I often put in stuff like “this gets us a huge step towards …”: 
showing that the big picture is moving, too. 

In short: The month of the Hurd isn’t just a newsfeed. It is a take at fixing 
the public image of the Hurd. 

Best wishes, 
Arne


signature.asc
Description: This is a digitally signed message part.

news 2010-04 in identi.ca and twitter

[Arne Babenhauserheide]

Hi, 

If you use identi.ca or twitter, you can find our current news in there, too: 

→ https://identi.ca/notice/30857295
→ https://twitter.com/ArneBab/status/13288232394

(well, the title and a link)

Best wishes, 
Arne


signature.asc
Description: This is a digitally signed message part.

Re: news 2010-04

[Arne Babenhauserheide]

Am Montag, 3. Mai 2010, 02:25:41 schrieb olafbuddenha...@gmx.net:
> Apparently it didn't work out, if it needs explanation ;-)

Jepp :) 

- Arne


signature.asc
Description: This is a digitally signed message part.

Re: news 2010-04

[olafBuddenhagen]

Hi,

On Sun, May 02, 2010 at 02:34:42PM +0200, Thomas Schwinge wrote:

> You new have roughly six hours to further comment on
>  -- then I'm
> going to publish it.

So I missed is as usual.

> After a long break, he as well updated his toolchain cross-compilation
> script to the current source code packages, [...]

This is wrong grammer I believe: "as well" should go at the end -- or
just use "also" instead, which would be fine in that place :-)

BTW, the plug at the end sounds rather pathetic IHMO... But perhaps
that's just me ;-)

> It is surely very valuable work that Sergio is doing (Go!, Go!, Go!),
> but we previously agreed to only list items that have already (and
> recently) been published as patches (or equivalent) -- such that
> others can use / reproduce / verify it.

Not sure such a strict policy is really useful?

In my understanding the major point of the regular news is showing
people that things are moving in the Hurd world -- which doesn't always
have to be manifested in patches I'd say...

Well, perhaps I'm thinking "kernel traffic" too much here.

-antrik-

Re: news 2010-04

[olafBuddenhagen]

Hi,

On Sun, May 02, 2010 at 01:43:03PM +0200, Arne Babenhauserheide wrote:
> Am Sonntag, 2. Mai 2010, 09:10:06 schrieb olafbuddenha...@gmx.net:
> > On Thu, Apr 29, 2010 at 06:07:50PM +0200, Arne Babenhauserheide wrote:

> > > > Also the fix from *Samuel Thibault* got included in libxcb1, so
> > > > X.org [works out of the box
> > > > again](http://lists.debian.org/debian-hurd/2010/04/msg00034.html)
> > > > with a simple startx and *Philip Charles* [extended his
> > > > offerings](http://lists.debian.org/debian-hurd/2010/04/msg00019.html)
> > > > with an updated *GRUB usb stick for booting Debian/Hurd*.
> > 
> > Err... These seem to be two completely unrelated news items?
> 
> They were mostly pulled together, because they are both concerned with ease 
> of 
> running hurd, and because I wanted to keep the news easy to take in (too many 
> paragraphs make it harder to read). 

Apparently it didn't work out, if it needs explanation ;-)

> (my mind isn???t with me enough the last two weaks ??? I have to take 
> antihistaminica for the first time in my life :( )

Oh :-(

-antrik-

Re: cache coherency in DMA and MMIO

[Samuel Thibault]

Da Zheng, le Sun 02 May 2010 20:36:28 +0800, a écrit :
> On 10-4-23 下午7:37, Samuel Thibault wrote:
> > Depends on how you use the mem device. It it ends up using pmap_enter()
> > to actually fill the page table, it will notice whether your physical
> > addresses are within RAM or above, see the code there.
> I think I find the code to do that.
> if (machine_slot[cpu_number()].cpu_type >= CPU_TYPE_I486
> && pa >= phys_last_addr)
> template |= INTEL_PTE_NCACHE|INTEL_PTE_WTHRU;

Yes, that's it.

> The MMIO address of e1000 is at 0xd892 and 0xd890. I even insert print
> in the code to confirm that the MMIO address has been set non-cache.

Ok.

Samuel

Re: news 2010-04

[Arne Babenhauserheide]

Am Sonntag, 2. Mai 2010, 14:39:00 schrieb Thomas Schwinge:
> Get well soon!

I’ll try :)

thanks!

: Arne


signature.asc
Description: This is a digitally signed message part.

Re: news 2010-04

[Arne Babenhauserheide]

Am Sonntag, 2. Mai 2010, 14:39:00 schrieb Thomas Schwinge:
> I used ``perceivedly boring, yet extremely valuable'' -- what do you
> think?

I’d say „it might look less exciting from the outside…“

Best wishes, 
Arne


signature.asc
Description: This is a digitally signed message part.

Re: news 2010-04

[Thomas Schwinge]

Hello!

I saw this email only after I had just sent mine.


On Sun, May 02, 2010 at 01:43:03PM +0200, Arne Babenhauserheide wrote:
> Am Sonntag, 2. Mai 2010, 09:10:06 schrieb olafbuddenha...@gmx.net:
> > On Thu, Apr 29, 2010 at 06:07:50PM +0200, Arne Babenhauserheide wrote:
> > > > Also the fix from *Samuel Thibault* got included in libxcb1, so
> > > > X.org [works out of the box
> > > > again](http://lists.debian.org/debian-hurd/2010/04/msg00034.html)
> > > > with a simple startx and *Philip Charles* [extended his
> > > > offerings](http://lists.debian.org/debian-hurd/2010/04/msg00019.html)
> > > > with an updated *GRUB usb stick for booting Debian/Hurd*.
> > 
> > Err... These seem to be two completely unrelated news items?
> 
> They were mostly pulled together, because they are both concerned with ease 
> of 
> running hurd, and because I wanted to keep the news easy to take in (too many 
> paragraphs make it harder to read). 

I separated them.  I think that separate paragraphs are easier to parse.

> > I don't think "boring" is a suitable attribute :-)
> 
> > It may be less exciting 
> 
> Why didn’t I find that phrase?! :)
> And it’s so simple…

I used ``perceivedly boring, yet extremely valuable'' -- what do you
think?


> (my mind isn’t with me enough the last two weaks — I have to take 
> antihistaminica for the first time in my life :( )

Get well soon!


> @Thomas: can we still change it? 

Yes, until it appears on .


Regards,
 Thomas


signature.asc
Description: Digital signature

Re: cache coherency in DMA and MMIO

[Da Zheng]

On 10-4-23 下午7:37, Samuel Thibault wrote:
> Depends on how you use the mem device. It it ends up using pmap_enter()
> to actually fill the page table, it will notice whether your physical
> addresses are within RAM or above, see the code there.
I think I find the code to do that.
if (machine_slot[cpu_number()].cpu_type >= CPU_TYPE_I486
&& pa >= phys_last_addr)
template |= INTEL_PTE_NCACHE|INTEL_PTE_WTHRU;
The MMIO address of e1000 is at 0xd892 and 0xd890. I even insert print
in the code to confirm that the MMIO address has been set non-cache.

Zheng Da

Re: news 2010-04

[Thomas Schwinge]

Hello!

I filtered the news text through my brain.  You new have roughly six
hours to further comment on
 -- then I'm
going to publish it.


On Sun, May 02, 2010 at 09:10:06AM +0200, olafbuddenha...@gmx.net wrote:
> Perhaps this is a stealth reference to it ;-) -- but I wonder whether it
> wouldn't be good to explicitely mention Sergio's experiments on the
> effect of out-of-band data on IPC performance; work on fixing tmpfs; and
> various suggestions for improving the external pager interface?

It is surely very valuable work that Sergio is doing (Go!, Go!, Go!), but
we previously agreed to only list items that have already (and recently)
been published as patches (or equivalent) -- such that others can use /
reproduce / verify it.  But I'm pretty sure that there'll be something to
include in next month's news.  :-)


Regards,
 Thomas


signature.asc
Description: Digital signature

Re: news 2010-04

[Arne Babenhauserheide]

Am Sonntag, 2. Mai 2010, 09:10:06 schrieb olafbuddenha...@gmx.net:
> Hi,
> 
> On Thu, Apr 29, 2010 at 06:07:50PM +0200, Arne Babenhauserheide wrote:
> > > Also the fix from *Samuel Thibault* got included in libxcb1, so
> > > X.org [works out of the box
> > > again](http://lists.debian.org/debian-hurd/2010/04/msg00034.html)
> > > with a simple startx and *Philip Charles* [extended his
> > > offerings](http://lists.debian.org/debian-hurd/2010/04/msg00019.html)
> > > with an updated *GRUB usb stick for booting Debian/Hurd*.
> 
> Err... These seem to be two completely unrelated news items?

They were mostly pulled together, because they are both concerned with ease of 
running hurd, and because I wanted to keep the news easy to take in (too many 
paragraphs make it harder to read). 

> I don't think "boring" is a suitable attribute :-)

> It may be less exciting 

Why didn’t I find that phrase?! :)
And it’s so simple…

(my mind isn’t with me enough the last two weaks — I have to take 
antihistaminica for the first time in my life :( )

@Thomas: can we still change it? 

> Perhaps this is a stealth reference to it ;-) -- but I wonder whether it
> wouldn't be good to explicitely mention Sergio's experiments on the
> effect of out-of-band data on IPC performance; work on fixing tmpfs; and
> various suggestions for improving the external pager interface?

I ran out of “space”[^1], so I had to try and find the most interesting parts 
for readers. 

[^1]: I try to keep the news short enough, that people can read them in a 
short break. 

Best wishes, 
Arne

--- --- --- --- --- --- --- --- --- --- --- --- --- --- --- --- --- 
   - Ein Würfel System - einfach saubere Regeln -
   http://1w6.org



signature.asc
Description: This is a digitally signed message part.

Re: news 2010-04

[olafBuddenhagen]

Hi,

On Thu, Apr 29, 2010 at 06:07:50PM +0200, Arne Babenhauserheide wrote:

> > Also the fix from *Samuel Thibault* got included in libxcb1, so
> > X.org [works out of the box
> > again](http://lists.debian.org/debian-hurd/2010/04/msg00034.html)
> > with a simple startx and *Philip Charles* [extended his
> > offerings](http://lists.debian.org/debian-hurd/2010/04/msg00019.html)
> > with an updated *GRUB usb stick for booting Debian/Hurd*.

Err... These seem to be two completely unrelated news items?

> >> Also *Emilio Pozuelo Monfort* mentored by *Fredrik Hammar* will
> >> work on the possibly boring yet extremely valuable task of [fixing
> >> compatibility problems exposed by projects testsuites]
> (http://socghop.appspot.com/gsoc/student_project/show/google/gsoc2010/gnuproject/t127230759396)
> >> ([[details|community/gsoc/project_ideas/testsuites]]). 

I don't think "boring" is a suitable attribute :-)

It may be less exciting in the sense that it doesn't produce new
features, better performance etc. (Though it's quite possible that some
fixes will improve overall stability.) However, I don't think the actual
work involved is less interesting than work on other tasks.

> > So we had quite an exiting month with many starting threads, which
> > makes this the perfect time to subscribe to our [[mailing_lists]]
> > (bug-hurd and debian-hurd are the main lists) and see the Hurd make
> > another leap forward :)

Perhaps this is a stealth reference to it ;-) -- but I wonder whether it
wouldn't be good to explicitely mention Sergio's experiments on the
effect of out-of-band data on IPC performance; work on fixing tmpfs; and
various suggestions for improving the external pager interface?

-antrik-