Linux-Development-Sys Digest #638
Linux-Development-Sys Digest #638, Volume #8 Thu, 12 Apr 01 11:13:15 EDT Contents: Re: CPU affinity in Red Hat Linux 6.2 (Kasper Dupont) Re: Suppressing Redhat bootup output (Wayne Pollock) Re: Need your recommendation for a full-featured text editor (Phillip Lord) Re: Need your recommendation for a full-featured text editor (Phillip Lord) Re: Need your recommendation for a full-featured text editor (Phillip Lord) Re: Need your recommendation for a full-featured text editor (Thore B. Karlsen) Re: Need your recommendation for a full-featured text editor (Phillip Lord) Re: Need your recommendation for a full-featured text editor (Phillip Lord) Re: Need your recommendation for a full-featured text editor (Ken Tough) Re: CPU affinity in Red Hat Linux 6.2 (bill davidsen) Makefile - Kernel dependancy ? ("Sean Bose") Re: A Linux emulator for Linux, does this exist? (bill davidsen) Re: Need your recommendation for a full-featured text editor (Phillip Lord) From: Kasper Dupont [EMAIL PROTECTED] Subject: Re: CPU affinity in Red Hat Linux 6.2 Date: Thu, 12 Apr 2001 13:50:33 + news.cso.uiuc.edu wrote: We have a large cluster of dual-processor Pentium II systems running Red Hat Linux 6.2. On a dual-processor system, we would like to tell the scheduler to place a given process on a specific processor and keep it there. This feature is sometimes called "CPU affinity." The idea is that we want to minimize the number of times a process bounces back and forth between processors. We could also like to monitor which processor a given process is running on. I couldn't find anything in the ps or proc man pages about how to do this. Are either of these features available on Red Hat Linux 6.2? Dave McWilliams University of Illinois Keeping a process on a given processor is not posible in a standard Red Hat 6.2 installation. Red Hat 6.2 uses kernel version 2.2.14 the feature you are asking for where added to the kernel in 2.4.x. If you upgrade your kernel to 2.4.x you will have the kernel support, but no part of the kernel uses this feature, no module uses this feature and there is no interface for using this feature from userspace. This was discussed some time ago and at that time I wrote a module implementing an interface through which userspace programs can use this feature. I will post the code in a few days, it is not currently online. I don't think there is any userspace interface for detecting which CPU a process is runing on, but the kernel know the information and uses it so it must also be posible to write a module implementing that interface. -- Kasper Dupont -- From: Wayne Pollock [EMAIL PROTECTED] Crossposted-To: comp.os.linux.misc,comp.os.linux.setup Subject: Re: Suppressing Redhat bootup output Date: Thu, 12 Apr 2001 10:34:50 -0400 I once looked into where these messages came from. Turns out there is a single file that each RC script runs to load some "shell functions" such as "daemon" that in turn start a service and produce the [OK] message output. (This was on SusE and Red Hat I believe, but I could be mistaken it was a while ago that I looked.) So it seems to me you have one shell script file to edit for this change. look at the top of any of the rc scripts in .../init.d for a line such as: . functions and then edit the file "functions" (it may not be called "functions"). You could for example send these messages to a file. Other possibilities include switching the virtual console early. The messages may be then sent to a single virtual console that the users don't see by default. I don't know how to do this or even if it is possible to do this early enough in the boot sequence to be of use to you. Another possibility is displaying a splash screen during boot up. I think there is software for this, I have seen systems that have such a screen, but I don't know where to get the software. Hope this helps! -Wayne Pollock Paul Haley wrote: Hello, I'm running Redhat 6.2 and would like to suppress the output on bootup, such as "Loading cron. [OK]", etc. All I want is to have lilo say "loading linux" and then give me a login prompt, though of course I still want all the programs/processes to run, I just don't want them to output to the screen. I checked out the init scripts and nothing jumped out at me. Can somebody help? Regards, Paul -- From: Phillip Lord [EMAIL PROTECTED] Crossposted-To: 24hoursupport.helpdesk,alt.comp.shareware.programmer,comp.editors,comp.lang.java.help,comp.lang.java.programmer,comp.lang.java.softwaretools,comp.os.linux.advocacy Subject: Re: Need your recommendation for a full-featured text editor Date: 12 Apr 2001 15:46:04 +0100 "Aaron" == Aaron R Kulkis [EMAIL PROTECTED] writes: Gamma And this is just for bare emacs; multiply all numbers by
Linux-Development-Sys Digest #639
Linux-Development-Sys Digest #639, Volume #8 Thu, 12 Apr 01 13:13:15 EDT Contents: Kernel summit Flask outcome? (Nicholas Geovanis) Re: Via + Maxtor + kernel 2.4.3 = crash? (Mark Hahn) Re: Need your recommendation for a full-featured text editor ("Aaron R. Kulkis") Re: Can we can change HZ from 100 to 1? (Mark Hahn) Re: Need your recommendation for a full-featured text editor ("Aaron R. Kulkis") Re: Need your recommendation for a full-featured text editor ("Aaron R. Kulkis") Re: Need your recommendation for a full-featured text editor ("Aaron R. Kulkis") BIOS Upgrades (Was: Via + Maxtor + kernel 2.4.3 = crash?) (Harold Stevens US.972.952.3293) Re: Limit in variable size in device driver? (Bob Hauck) Re: Need your recommendation for a full-featured text editor (Gamma) Re: Need your recommendation for a full-featured text editor (Gamma) From: Nicholas Geovanis [EMAIL PROTECTED] Subject: Kernel summit Flask outcome? Date: Thu, 12 Apr 2001 10:23:13 -0500 Reply-To: [EMAIL PROTECTED] What was the outcome of the NSA/Flask pitch to the kernel summit last week or the week before? I assume that the Network Associates takeover of work on it means that the answer was "no", but I didn't hear. Thanks. * Nick Geovanis | IT Computing Svcs | Northwestern Univ | [EMAIL PROTECTED] +--- -- From: Mark Hahn [EMAIL PROTECTED] Crossposted-To: comp.os.linux.hardware,comp.os.linux.questions,comp.os.linux.misc Subject: Re: Via + Maxtor + kernel 2.4.3 = crash? Date: 12 Apr 2001 15:48:24 GMT I tend to think your configuration may have trouble doing IDE DMA at its maximum specced speed. This wouldn't be the first time VIA chipsets have had it has nothing at all to do with speed. some bioses misconfigure the via hostbridge, and this corrupts most busmastering IO (including IDE, but also reported on some NICs.) this problem. I'd suggest first trying it with DMA disabled (this will severely hurt performance btw). If this works, then you might try looking gross. flash an updated bios or turn off "optimum" settings in bios. -- From: "Aaron R. Kulkis" [EMAIL PROTECTED] Crossposted-To: 24hoursupport.helpdesk,alt.comp.shareware.programmer,comp.editors,comp.lang.java.help,comp.lang.java.programmer,comp.lang.java.softwaretools,comp.os.linux.advocacy Subject: Re: Need your recommendation for a full-featured text editor Date: Thu, 12 Apr 2001 12:00:33 -0400 Phillip Lord wrote: "Aaron" == Aaron R Kulkis [EMAIL PROTECTED] writes: Gamma And this is just for bare emacs; multiply all numbers by Gamma around 8 or so for xemacs with news, ftp, fried okra, mashed Gamma potatoes, etc... You want the additional functionality you have to pay something for it. Given that the VI user probably also has a news reader, an ftp client, a mail reader, uses CVS an so on as well, then I am not sure that the memory usage is that significant. Aaron Maybe on academic machines. Aaron Most corporate Unix machines NEVER have a new reader on them. Aaron Many don't even have CVS, as they aren't used for Aaron development. I thought that we were talking about memory usage here. If you don't use the news reader and so forth then obvious they do not add to emacs' memory footprint. I'm talking about disk space. And don't give me the "it's only 10 cents worth of disk space" routine. Most applications are only a couple cents worth of disk space...but that doesn't mean you clutter up your servers with every piece of code you can find on the net. And if you are not using emacs for development, then I agree there is little point in using it. I mean I did use to use it only for email, but I am a little weird. Gamma And 120 more bucks worth of install time if you're in a Gamma consultant position and your client doesn't have emacs Gamma installed, or has the wrong version. It takes me about 2 minutes to install a new emacs on my local system. Thats because I have taken care to write my emacs customisations (which are pretty extensive) well. If you move machines a lot then the advantage of VI clearly gets greater. In the same way people who swap machines rarely customise their desktop, or even their shell. Aaron Clue for the clueless: consulting is where the BIG money is. Aaron That means working on-site at up to half-a dozen different Aaron client's locations within a week. My friend if money were my motivation I would not be working in the public sector now would I? I am fully aware of the of consultancy pay scales (not as good as they were now that IR35 means our consultants have to pay tax like everyone else!). Perhaps you feel I am clueless for not going after the big money? Perhaps I think you are clueless for going after the big
Linux-Development-Sys Digest #640
Linux-Development-Sys Digest #640, Volume #8 Thu, 12 Apr 01 17:13:14 EDT Contents: IP filtering ("Fabio Ventrone") supporting functions to the device driver writer? (philip) Re: Can we can change HZ from 100 to 1? ("tlin") Re: CPU affinity in Red Hat Linux 6.2 (Philip Armstrong) Re: Need your recommendation for a full-featured text editor (NO FTAA) Re: BIOS Upgrades ("Peter T. Breuer") Re: BIOS Upgrades (Harold Stevens US.972.952.3293) Re: Need your recommendation for a full-featured text editor (The Ghost In The Machine) Re: Need your recommendation for a full-featured text editor (The Ghost In The Machine) Re: Need your recommendation for a full-featured text editor (The Ghost In The Machine) Re: BUG: ftime inconsistent [RH7-fisher (kernel 2.4)] (Sten) Re: New directions for kernel development (The Ghost In The Machine) Re: Need your recommendation for a full-featured text editor (John Hawkins) Re: BIOS Upgrades ("Peter T. Breuer") From: "Fabio Ventrone" [EMAIL PROTECTED] Crossposted-To: comp.os.linux.networking Subject: IP filtering Date: Thu, 12 Apr 2001 18:31:37 +0200 I have to write a program that capture all the IP traffic and elaborate it (it has also to route it to a simulated radio bridge connected via TCP/IP...). How to capture the packets ? I initially tryied with "libpcap" but I had some problems (it gave me a segm. fault...). Then I though that for my kind of work it'd have been enough an easyer way to work: I can configure a firewall (ipchains) for all "input" packets that REDIRECT them to a local socket. this is the command I gave: ipchains -A input -d \! 127.0.0.1 -p TCP -j REDIRECT this should redirect all outgoing traffic to my local port , but it doesn't work. It just works as if there was no firewall. I have 3 "or-ed" questions: 1) do you have any other way to capture packets, elaborate them and redirect something else? 2) why my firewall config does not what I expect? 3) do you have any easy example for libpcap that does something similar?? Thank you FabioVentrone -- From: philip [EMAIL PROTECTED] Subject: supporting functions to the device driver writer? Date: Thu, 12 Apr 2001 17:30:36 GMT Hi! Is there anyone knows where I can find complete supporting function specification for device driver writer? from here I saw one:(seems was written in 1996) http://www.linuxdoc.org/LDP/khg/HyperNews/get/devices/reference.html and seems it is too old and not including something like "ioremap" that kind fucntion. any one can tell me where to get the newer and more complete one? thank you! philip -- From: "tlin" [EMAIL PROTECTED] Subject: Re: Can we can change HZ from 100 to 1? Date: Thu, 12 Apr 2001 13:34:02 -0600 Yeah, the new time interval is not guarantee to be 100 microseconds. Could you give me some more details about rtc? Thanks! Mark Hahn [EMAIL PROTECTED] wrote in message news:9b4it2$96n$[EMAIL PROTECTED]... I changed it from 100 to 1. It seems working now. I can get 100 microsecond time interval by the new HZ. this doesn't make much sense. first, rdtsc is the way to get nanosecond timing. if you want 100 us timers (alarms), there are also probably other better ways (for one, rtc can easily give you 8 KHz, without totally destroying the efficiency of the system/scheduler.) -- From: [EMAIL PROTECTED] (Philip Armstrong) Subject: Re: CPU affinity in Red Hat Linux 6.2 Date: 12 Apr 2001 10:18:57 +0100 In article H83B6.1015$[EMAIL PROTECTED], news.cso.uiuc.edu [EMAIL PROTECTED] wrote: sometimes called "CPU affinity." The idea is that we want to minimize the number of times a process bounces back and forth between processors. We could also like to monitor which processor a given process is running on. I couldn't find anything in the ps or proc man pages about how to do this. Are either of these features available on Red Hat Linux 6.2? Not without applying experimental kernel patches and recompiling your kernel, no. Note that processes already *have* a measure of cpu affinity built into the scheduler; it prefers to run then on the cpu they were last running on. If you're seeing processes bouncing between cpus it may be a masurement artifact (ie, the use of top / xosview to observe the processes is adding a process to the run queue which changes the scheduler decisions.) If you run a pair of cpu-bound test programs that find out for themselves (from /proc say) which cpu they are running on, you should find that they tend to remain on the same cpu. Phil -- http://www.kantaka.co.uk/ .oOo. public key: http://www.kantaka.co.uk/gpg.txt -- From: NO FTAA [EMAIL PROTECTED] Crossposted-To:
Linux-Development-Sys Digest #641
Linux-Development-Sys Digest #641, Volume #8 Fri, 13 Apr 01 02:13:12 EDT Contents: Re: BIOS Upgrades (Harold Stevens US.972.952.3293) Re: BIOS Upgrades (Was: Via + Maxtor + kernel 2.4.3 = crash?) ("Matt Ng") Re: IP filtering ([EMAIL PROTECTED]) TCP send buffer (Sam Liang) Re: usbutils (Phil Ehrens) kmod: failed to exec /sbin/modprobe -s -k binfmt-464c, errno=8 (Marc D Bumble) Re: BIOS Upgrades (Was: Via + Maxtor + kernel 2.4.3 = crash?) (Harold Stevens US.972.952.3293) Re: Need your recommendation for a full-featured text editor (Hansang Bae) Re: checking for non ansi extensions ("Raymond Limpus") Re: kmod: failed to exec /sbin/modprobe -s -k binfmt-464c, errno=8 (Paul Kimoto) Re: Setting Linux Disk Buffer Cache (Eric Taylor) ide vs. scsi why so much slower (Eric Taylor) Re: Need your recommendation for a full-featured text editor (Michael Vester) Re: New directions for kernel development (Brent R) From: [EMAIL PROTECTED] (Harold Stevens US.972.952.3293) Crossposted-To: comp.os.linux.hardware,comp.os.linux.questions,comp.os.linux.misc Subject: Re: BIOS Upgrades Date: Thu, 12 Apr 2001 21:46:13 GMT In [EMAIL PROTECTED], Peter T. Breuer: [Snip...] Which is it? The procedure I suggested will enable you to find out. Like I said: I'll do yast in 32 MB standing on my head in Hades first. Suddenly I don't care anymore about this futile exercise, OK? [Snip...] -- Regards, Weird (Harold Stevens) * IMPORTANT EMAIL INFO FOLLOWS * Pardon the bogus email domain (dseg etc.) in place for spambots. Really it's (wyrd) at raytheon, dotted with com. DO NOT SPAM IT. Standard Disclaimer: These are my opinions not Raytheon Company. -- From: "Matt Ng" [EMAIL PROTECTED] Crossposted-To: comp.os.linux.hardware,comp.os.linux.misc Subject: Re: BIOS Upgrades (Was: Via + Maxtor + kernel 2.4.3 = crash?) Date: Thu, 12 Apr 2001 18:01:38 -0400 On this topic, I need pointers on BIOS upgrades, especially where to find them, as I haven't had any luck at all with Google to date. Did you try http://www.ping.be/bios/index.html ? Found it using Google a while back when upgrading a BIOS. Rebooting resulted in a kernel panic which I almost expected on these old BIOS versions (Award 4.51PG i430VX) from reading threads I found. I'm using that same BIOS version and haven't had any problems with 96 and 128 MB of RAM, so I doubt that it's a BIOS problem... I didn't even need to change the LILO file. I'd think it's RAM related. Try just running the new sticks alone and see if that brings up the kernel panic. -- From: [EMAIL PROTECTED] Crossposted-To: comp.os.linux.networking Subject: Re: IP filtering Date: Thu, 12 Apr 2001 22:41:29 - In comp.os.linux.development.system Fabio Ventrone [EMAIL PROTECTED] wrote: | I have to write a program that capture all the IP traffic and elaborate it | (it has also to route it to a simulated radio bridge connected via | TCP/IP...). | | How to capture the packets ? Are you trying to "sniff" the traffic as in get a copy of everything which also continues to its normal destination, or are you trying to route everything to only go to a specific destination? If the former, see a 2 weeks old thread here labeled "opening ethernet as a raw device" where answers to my question were along these lines. Using the socket interface appropriately allows access to a copy of all raw packets. This is what libpcap does on Linux. The man pages do have most of the info needed. If the latter, look into the VTAP/VTUN interface, which allows a process to deploy a network to be routed to. It's often used for routed or bridged tunnels, but can be used for other creative ideas, too. The kernel source Documentation tree has some info. | 1) do you have any other way to capture packets, elaborate them and redirect | something else? The 2 above ways are good choices. | 2) why my firewall config does not what I expect? Unless I misinterpret what you expect, a firewall isn't for that. | 3) do you have any easy example for libpcap that does something similar?? Sorry, no. I personally don't like the libpcap API, so I don't use it. -- = | Phil Howard - KA9WGN | Dallas | http://linuxhomepage.com/ | | [EMAIL PROTECTED] | Texas, USA | http://phil.ipal.org/ | = -- From: Sam Liang [EMAIL PROTECTED] Subject: TCP send buffer Date: Thu, 12 Apr 2001 16:42:46 -0700 Reply-To: [EMAIL PROTECTED] I am trying figure out how TCP send buffer is managed in the kernel. Is struct sock { struct sk_buff_head write_queue} the TCP send buffer? I am talking about kernel 2.2.x. In tcp_send_skb(), a new skb is enqueued to write_queue. However, I couldn't find where it frees old data from the