Re: [gPXE] I want to render the lan boot rom blank
On Dec 23, 2013 6:03 AM, Jim Michaels jmich...@yahoo.com wrote: how do I do this? this is the flash rom built into a system. is there a command for this? I could sure use one. I am trying to make it so if the lan boot rom does get booted again, it will skip the lan boot and default to the hard drive and ignore the LAN boot ROM I would like to flash. it's built into a motherboard. I am trying to render it inert. help appreciated. software even more appreciated, especially if it's an iso I can just burn to clear the rom. Use the BIOS setup to remove the option and set a BIOS setup password. Most LoMs get their OROM from the system firmware and attempts to wipe it will likely render the system unbootable. Seting the password I'm my experience is often the only way to remove the option from manual selection. Some systems still allow its manual select but require the BIOS setup password to continue. --Gene ___ gPXE mailing list gPXE@etherboot.org http://etherboot.org/mailman/listinfo/gpxe
Re: [gPXE] Problem using gPXE to load PXELINUX with HTTP server
On Fri, Dec 6, 2013 at 6:39 PM, Allan Wu al...@cs.ucla.edu wrote: Hello, I got a problem of starting PXELINUX with gPXE using HTTP server. I am trying to setup a diskless system using gPXE. I intend for all the client machines to use static IP in the network, and our network already have a DHCP server which is difficult for me to modify. So I want to use gPXE to assign static IP for my client machine and obtain image from a HTTP server. In gPXE, I can successfully load the pxelinux.0 file and execute it. However, PXELINUX then reported an error saying Unable to locate configuration file. I tried with several versions but got similar problems. In general, I want to set each client with static IP to get around the DHCP server, which seems to work for getting the images. I also tried to load the initrd image and vmlinuz kernel directly. But I got error saying the kernel cannot find ramdisk. I think I need to use the parameters such as 'initrd=' and 'root=' when booting the kernel. Is there a way to do that with gPXE along? I'm open to other alternatives, since my approach may be limited on my lack of knowledge. My setup looks like this: - On the HTTP server: http://.xxx.xxx/xxx/tftpboot/ ~ this is the root directory for the pxelinux file ./pxelinux.0 ./pxelinux.cfg/default ./centos6/ ~ this directory contains the ramdisk image and kernel The pxelinux configure file looks something like this: default rhel6 label rhel6 kernel centos6/vmlinuz append initrd=centos6/diskless_initrd.img root=nfs:xxx.xxx.xxx.xxx:/diskless/x86_64/RHEL6/root/ rw - For my client, I embedded a static.gpxe script looks something like this: #! gpxe ifclose net0 set net0/ip xxx.xxx.xxx.xxx set net0/netmask xxx.xxx.xxx.xxx set net0/gateway xxx.xxx.xxx.xxx set net0/dns xxx.xxx.xxx.xxx ifopen net0 set 209:string pxelinux.cfg/default set 210:string http://xxx.xxx.xxx/xxx/tftpboot/ chain ${210:string}pxelinux.0 After this the PXELINUX is booted, but then failed on the error. I was using PXELINUX 4.02. For newer version, I also tried 5.10, but then PXELINUX stucks on the first line (version and copyright stuff). How are you loading gPXE, from floppy/ISO/USB? For a first test, I'd suggest a different version of PXELINUX. Having seen numerous bugfixes on sysli...@zytor.com, I'd suggest starting with the precompiled binaries from the binary/source archive at kernel.org and using 4.07. If you want to try 5.1x, you can either use gPXE to load core/pxelinux.0 via HTTP then com32/elflink/ldlinux/ldlinux.c32 (yes, it needs that before the configuration file) OR set DHCP option 210 (direct or encapsulated) on your DHCP server and load core/lpxelinux.0 from TFTP then all other files relative to the prefix in 210. I'd suggest version 5.11-pre9 for your first test of a 5.1x version. Also, if you wanted to try 6.0x, either 6.02 or 6.01, either bios/core/pxelinux.0 or bios/core/lpxelinux.0 with bios/com32/elflink/ldlinux/ldlinux.c32 -- -Gene ___ gPXE mailing list gPXE@etherboot.org http://etherboot.org/mailman/listinfo/gpxe
Re: [gPXE-devel] Attempting to add support for Broadcom BCM5755M (14e4:1673) ethernet adapter
On Tue, Jan 3, 2012 at 06:07, Quinn Plattel qie...@gmail.com wrote: Hi, Did you do a git pull? git clone git://git.etherboot.org/scm/gpxe.git git pull git://git.etherboot.org/people/meteger/gpxe.git branch tg3 is what you want. git remote add meteger git://git.etherboot.org/people/meteger/gpxe.git git checkout -b meteger-tg3 remotes/meteger/tg3 This should get you more consistent results. -- -Gene ___ gPXE-devel mailing list gPXE-devel@etherboot.org http://etherboot.org/mailman/listinfo/gpxe-devel
Re: [gPXE] [gPXE-devel] Attempting to add support for Broadcom BCM5755M (14e4:1673) ethernet adapter
On Tue, Jan 3, 2012 at 06:07, Quinn Plattel qie...@gmail.com wrote: Hi, Did you do a git pull? git clone git://git.etherboot.org/scm/gpxe.git git pull git://git.etherboot.org/people/meteger/gpxe.git branch tg3 is what you want. git remote add meteger git://git.etherboot.org/people/meteger/gpxe.git git checkout -b meteger-tg3 remotes/meteger/tg3 This should get you more consistent results. -- -Gene ___ gPXE mailing list gPXE@etherboot.org http://etherboot.org/mailman/listinfo/gpxe
Re: [gPXE] gPXE in uEFI environments
On Sun, Dec 11, 2011 at 02:49, muralidhar.appa...@emulex.com wrote: Hello, I would like to know whether gPXE has support for network boot and iSCSI boot in uEFI environments. If yes, could someone help me in getting the documentation for same? Thanks Muralidhar I'm assuming that you're talking loading only EFI-compatible payloads. 'make bin/driver.efi' should work and should be capable of network boot and iSCSI boot. As far as docs, I'd say start with 'make doc'. -- -Gene ___ gPXE mailing list gPXE@etherboot.org http://etherboot.org/mailman/listinfo/gpxe
Re: [gPXE] how to use gpxe w/o dhcp
On Mon, Nov 7, 2011 at 00:41, Yixuan Huang yixuan...@gmail.com wrote: Hello Everyone, I have one question, if my environment doesn't have dhcp server, how can I set ip for gpxe and load remote kernel? Thanks, yixuan Yes, you can embed a script into gPXE. If a static script is insufficient, you can have a static script in gPXE reference a static or dynamic script generated from an HTTP server. You could include attributes like UUID, NIC MAC address, and IP as portions of the URL when fetching the HTTP generated script. -- -Gene ___ gPXE mailing list gPXE@etherboot.org http://etherboot.org/mailman/listinfo/gpxe
Re: [gPXE] Question About using GPXE in my environment
On Tue, Oct 18, 2011 at 19:09, Regis Chapman (Prithvi Catalytic Inc) v-rec...@microsoft.com wrote: I’m in a completely Microsoft environment, except for the lab machines I have running RHEL 5 and 6. I’m trying to set up a PXE server on one Linux box that inherits it’s PXE authority from the main WDS server on this network (see below). This machine is also the DHCP server. I’m currently running in a test environment that we will make production after working out the bugs. I've seen true PXE servers (which use UDP port 4011) actually supersede the boot fields in DHCP when the PXE server responds in time. Configuring the network for them involves either putting the PXE server on the same subnet or having the DHCP proxy also ask the PXE server for a response. For example, in Cisco IOS devices like Catalyst switches used in an IP routing capability, this is done with the 'ip helper-address' configuration parameter. Almost all documentation for achieving PXE booting assumes your PXE server and your DHCP server are on the same machine, and I’m looking for a way to configure my way out of this mess. I have no control over the DHCP server nor the WDS server, but I’ve gotten them to follow this process: http://www.vcritical.com/2011/06/peaceful-coexistence-wds-and-linux-pxe-servers/#comment-12655 To allow a WDS server to hand off via PXE Chain, to my Linux server. Using a TFTP server seems to be the common thing, but at the same time, maybe I should choose and setup an HTTP server for this instead? My Linux boxes are spread across subnets and in different physical locations. So now I have some level of control over what goes on. However, I have never set up a PXE server before and I’d like some advice on how to understand what is happening, and how to act upon it. Hope this helps. -- -Gene ___ gPXE mailing list gPXE@etherboot.org http://etherboot.org/mailman/listinfo/gpxe
Re: [gPXE] native gPXE loading custom gPXE
On Sep 29, 2011 12:52 PM, Richman, Steven C steven.c.rich...@intel.com wrote: I am using a modified version of gPXE that allows 256 characters to be passed to the kernel command line. gPXE needs to pass a very long string of arguments to the kernel command line. I am chain-loading gPXE from PXE. DHCP is configured to look for the gpxe.bus-id in the dhcpd.conf file to break out of the infinite loop. Depending on your real goals, have you considered using gpxelinux.0 from Syslinux or pxelinux.0 from Syslinux-4.10-pre15 (current beta which includes lwIP, a lightweight IPv4 stack) which can handle kernel command lines up to about 2047 total characters? gpxecmd.c32 can also be used to issue commands to gPXE. The IPAPPEND option also gives a nice way to identify the NIC that you're booting from. Generally, an IPAPPEND value of 2 is the recommended option but doc/syslinux.txt can help with more details. The new Qlogic NICs are coming with gPXE. This poses a problem because it doesn't load the custom gPXE (w/Kernel CmdLine patch) because DHCP sees that the gpxe.bus-id is already defined so it doesn't hand-out the custom gPXE like it does for NICs that PXE boot. Is there a known solution for loading a custom (patched) gPXE from gPXE (native) without falling into an infinite loop? If you need gPXE without PXELINUX, embedding a script into your customized gPXE build to perform your task(s) and break this cycle is an easy option. Can an attribute be set to tell dhcp to always offer the custom bootloader unless the attribute defined in the custom gPXE version is present? Thanks, Steve Good luck. --Gene ___ gPXE mailing list gPXE@etherboot.org http://etherboot.org/mailman/listinfo/gpxe
Re: [gPXE] Convert TFTP boot to gPXE
On Fri, Aug 19, 2011 at 15:53, Andrew Bobulsky rule...@gmail.com wrote: Ben, Not entirely sure what the problem is there. In my experience, 150 megs or so should come down within 20 seconds on gigabit when using the undionly.kpxe build. The best suggestion that I have is to try monitoring the download from the server end, if possible. If your webserver doesn't let you easily do that in real time, I can suggest HFS. It's a very small Windows app that makes a great HTTP File Server in a pinch: http://www.rejetto.com/hfs/ Otherwise, perhaps someone else may have an idea :) Cheers, Andrew Bobulsky 1) What about Wireshark on the server? This would be especially handy if something else by chance is interfering (although unlikely) 2) Why not a middle of the road file (say 14 MiB but 10-20 would suffice nicely)? 3) Is your initrd completely uncompressed? If so, look into what decompression your kernel supports (recent kernels can even support XZ compressed initrds). 4) Does it eventually load that initrd or does it eventually just fail? What HTTP daemon? Tried using your web browser to ensure you can appropriately download it (including a post-download check like sha1sum)? 5) By bigger blocksize, are you thinking the data size that each period character represents? I know PXELINUX changed in the last year or so to a much larger block size per period character. -- -Gene On Fri, Aug 19, 2011 at 3:19 PM, Ben Francis wireless...@gmail.com wrote: Dimitri, The same initrd works when I boot over TFTP using pxelinux.0, but it doesn't work from the web directory, so I think the initrd is good. Thanks for archlinux link. Andrew, I tried loading the initrd first before the kernel, i.e. gPXE dhcp net0 DHCP (net0 mac-address) ok gPXE initrd http://ip-addres/gpxe/initrd . . . ...(20 lines of this) is the result for about ten minutes now. Same problem. The initrd IS a huge file at 143MB, but my network downloads that size file over tftp in about a minute. gPXE was loading the 1.4 MB vmlinuz file in one second, so it seems gPXE knows how to download a file quickly. Can the initrd command take parameters like a bigger blocksize, say 4096? Ben On Fri, Aug 19, 2011 at 6:08 PM, DiP aspamkil...@yandex.ru wrote: I think you have to check your initrd. For me wiki.archlinux.org directions have been very usefull in that way. Look at the Network boot NFS root article. I hope my working configuration will be usefull for you: dd if=/srv/mixer/boot/gpxe-1.0.1-gpxe.usb of=/dev/flash_card for booting client from that flashcard (I use it for clients that cannot boot from network) dhcpd.conf allow booting; allow bootp; group{ option domain-name home; option domain-name-servers 192.168.0.254; server-name server.home; next-server 192.168.0.254; if exists dhcp-parameter-request-list { option dhcp-parameter-request-list = concat(option dhcp-parameter-request-list,d0,d1,d2,d3); } host mixer { hardware ethernet 00:30:4f:12:d9:a6; fixed-address 192.168.0.51; option routers 192.168.0.254; site-option-space pxelinux; option pxelinux.magic f1:00:74:7e; option pxelinux.configfile pxelinux.cfg/default; option pxelinux.pathprefix http://server.home/;; filename http://server.home/boot.txt;; } } /srv/mixer/boot/boot.txt #!gpxe echo Greetings! kernel http://server.home/vmlinuz-linux rootfstype=nfs root=/dev/nfs nfsroot=192.168.0.254:/srv/mixer,v3,rsize=16384,wsize=16384 ip=::dhcp initrd http://server.home/initramfs-linux.img boot /etc/exports /srv/mixer 192.168.0.0/24(rw,fsid=0,no_root_squash,no_subtree_check,async) and I have working archlinux installation in /srv/mixer/ from nginx.conf server { listen 192.168.0.254; server_name server.home; charset utf-8; access_log logs/tftp.access.log; root /srv/mixer/boot; autoindex on; } If you want to boot your client from the standard network card ROM you have to change filename http://server.home/boot.txt;; to filename http://server.home/gpxelinux.0;; so the BIOS of the card will boot gpxelinux, who will get paramethers from pxelinux.cfg/default where contents is: default linux label linux kernel http://server.home/vmlinuz-linux append initrd=http://server.home/initramfs-linux.img rootfstype=nfs root=/dev/nfs
Re: [gPXE] chainloading xxx.c32
On Sun, Jul 3, 2011 at 11:34, Tom Baumann t...@thomasbaumann.com wrote: Hello list, I currently try to chainload the Menu from kernel.org's bko. But when performing a chain http://boot3.kernel.org/bko/menu.c32 I always get Requested memory not available (0x46038001). I tried as well with undionly.kpxe. Version is gPXE 1.0.1. gPXE 1.0.1 -- Open Source Boot Firmware -- http://etherboot.org Features: HTTP HTTPS iSCSI DNS TFTP bzImage COMBOOT ELF Multiboot NBI PXE PXEXT How could this being solved? How will it be possible to chainload the comboot.c32? IIRC, BKO has been updated to Syslinux 4.xx. Those .c32 files are in the new COM32R format which gPXE does not support. Loading PXELINUX first with appropriate DHCP options should solve this issue. -- -Gene ___ gPXE mailing list gPXE@etherboot.org http://etherboot.org/mailman/listinfo/gpxe
Re: [gPXE] chainloading xxx.c32
On Sun, Jul 3, 2011 at 13:01, Tom Baumann t...@thomasbaumann.com wrote: But how can I load c32 then without the use of tftp or dhcp-options? Are there plans to solve this issue? And, will it work with Syslinux-3 c32-Files? Files from Syslinux 3.86 should work. Bear in mind that gPXE can set these DHCP options (manually or via a script) such that they need not be set in the DHCP server. Am 03.07.2011 17:41, schrieb Gene Cumm: On Sun, Jul 3, 2011 at 11:34, Tom Baumannt...@thomasbaumann.com wrote: Hello list, I currently try to chainload the Menu from kernel.org's bko. But when performing a chain http://boot3.kernel.org/bko/menu.c32 I always get Requested memory not available (0x46038001). I tried as well with undionly.kpxe. Version is gPXE 1.0.1. gPXE 1.0.1 -- Open Source Boot Firmware -- http://etherboot.org Features: HTTP HTTPS iSCSI DNS TFTP bzImage COMBOOT ELF Multiboot NBI PXE PXEXT How could this being solved? How will it be possible to chainload the comboot.c32? IIRC, BKO has been updated to Syslinux 4.xx. Those .c32 files are in the new COM32R format which gPXE does not support. Loading PXELINUX first with appropriate DHCP options should solve this issue. -- Thomas Baumann Johann-Adolf-Hasse-Str. 2b 21493 Schwarzenbek 0172 6646466 ___ gPXE mailing list gPXE@etherboot.org http://etherboot.org/mailman/listinfo/gpxe
Re: [gPXE] gpxe booting
On Wed, Apr 20, 2011 at 05:35, Markus Stopper mar...@stopper.at wrote: Hello, I tried to use syslinux 4.04 components (http://www.kernel.org/pub/linux/utils/boot/syslinux/) with gPXE generated image with all drivers (kkpxe -http://rom-o-matic.net/gpxe/gpxe-git/gpxe.git/contrib/rom-o-matic/) on the Application Note: User specific boot menus (http://www.etherboot.org/wiki/appnotes/authmenus) and the chainloading worked right away. My webserver provides all data generated out of the php scripts (checked in browser), but at the time as it comes to chainload (chain vesamenu.c32 menu.php) the system freezes. You can't load Syslinux-4.04 vesamenu.c32 directly; you must use PXELINUX to load vesamenu.c32. I just updated that authmenus page earlier this week but I think it still needs a little bit of refinement. My Question here since I just started with linux: Do I need to match the rom-o-matic hook image somehow with syslinux version, since nowhere was a version given. If this is true, can somebody give me a quick tutorial how to build on what system? (I already got me a couple of vms in case I need to build something Debian, ubuntu, fedora…etc.) So please help me since I really just want to do the example on the user specific boot menus. I think it must have worked somehow because screenshots are there and they look great. An easy option is to use the pre-built binary gpxe/gpxelinux.0 which is gPXE-1.0.0's undionly.kkpxe (with minor patches), core/pxelinux.0 and a script to run PXELINUX. -- -Gene ___ gPXE mailing list gPXE@etherboot.org http://etherboot.org/mailman/listinfo/gpxe
Re: [gPXE] Win 7 install pxe chain loader nic won't auto negotiate 1000Mbps
On Tue, Apr 12, 2011 at 20:25, Schafer, Christopher L CTR USAF AFMC AFRL/RDSM christopher.schafer@maui.afmc.af.mil wrote: Installing a HP z400 using win 7 x 64 ent onto a Oracle 7320 iscsi lun. Machine gets stuck at 10Mbit….. Suggestions? First would probably be providing more details. What's the sequence that you're using? What gPXE driver are you using? Have you been able to test with any other OS setup that negotiates the NICs to other speeds (ie WinPE or a Linux LiveCD)? What's the PCI ID of your NIC (perhaps a bug if you're using a full driver)? -- -Gene ___ gPXE mailing list gPXE@etherboot.org http://etherboot.org/mailman/listinfo/gpxe
Re: [gPXE] Win 7 install pxe chain loader nic won't auto negotiate 1000Mbps
On Tue, Apr 12, 2011 at 22:31, Schafer, Christopher L CTR USAF AFMC AFRL/RDSM christopher.schafer@maui.afmc.af.mil wrote: HP PXE downloads gPXE from tftp downloads gpxe script connects to SAN/NAS iscsi lun. I have been able to hard code 100 full in the hp PXE bios but it won't hard code to 1000 nor does auto work. Switch is a 2960G copper. Not sure which driver but it does appear to be a Broadcom Broadcom NetXtreme Gigabit PCI\VEN_14E4DEV_1684SUBSYS_1309103CREV_10 PCI\VEN_14E4DEV_1684SUBSYS_1309103C PCI\VEN_14E4DEV_1684CC_02 PCI\VEN_14E4DEV_1684CC_0200 I'd guess that you must be using the undionly driver in a PXE format (.pxe, .kpxe, or .kkpxe) as 14e4:1684 does not appear to be recognized by any driver in gPXE at this time. If you haven't tried the .kkpxe, it might be worth a shot. But since then I have been now having troubles getting the lun to stick around after the installer loads. It is up to the OS to reconnect to the LUN after it loads its own drivers. -Original Message- From: Gene Cumm [mailto:gene.c...@gmail.com] Sent: Tuesday, April 12, 2011 2:48 PM To: Schafer, Christopher L CTR USAF AFMC AFRL/RDSM Cc: gpxe@etherboot.org Subject: Re: [gPXE] Win 7 install pxe chain loader nic won't auto negotiate 1000Mbps On Tue, Apr 12, 2011 at 20:25, Schafer, Christopher L CTR USAF AFMC AFRL/RDSM christopher.schafer@maui.afmc.af.mil wrote: Installing a HP z400 using win 7 x 64 ent onto a Oracle 7320 iscsi lun. Machine gets stuck at 10Mbit. Suggestions? First would probably be providing more details. What's the sequence that you're using? What gPXE driver are you using? Have you been able to test with any other OS setup that negotiates the NICs to other speeds (ie WinPE or a Linux LiveCD)? What's the PCI ID of your NIC (perhaps a bug if you're using a full driver)? -- -Gene -- -Gene No one ever says, 'I can't read that ASCII(plain text) e-mail you sent me.' ___ gPXE mailing list gPXE@etherboot.org http://etherboot.org/mailman/listinfo/gpxe
Re: [gPXE] Booting a physical machine using a virtual image and gPXE
On Thu, Apr 7, 2011 at 14:42, Clive Eisen cl...@hildebrand.co.uk wrote: um unless I have really mis understood your question then NO A given image is for a given hypervisor - gPXE or indeed iPXE is not a hypervisor Boot the hypervisor then boot the image -- Clive On 7 Apr 2011, at 19:32, Swapnesh Chaubal wrote: Hey Folks, I am working on a project in which I would like to boot physical machines using virtual images stored on a server. I understand that we can use gPXE to boot ISO images over the network, and to boot Virtual Machine Hyper-visors like KVM/VirtualBox using gPXE, however, can we use gPXE to boot a physical machine using a .vmdk/.vhd (or any other format) virtual image? http://erwan.l.free.fr/iscsi/ It appears that you could present a .vmdk/.vhd as an iSCSI target however 1) You'd need to reconfigure the OS in the disk image to know to attach to the iSCSI target, almost defeating the purpose of the VM image file when used with a hypervisor. 2) If your OS happens to be Windows, it will likely detect a huge change in hardware and require activation again, likely preventing you from running it as a VM again. -- -Gene ___ gPXE mailing list gPXE@etherboot.org http://etherboot.org/mailman/listinfo/gpxe
Re: [gPXE] int 10h problem
2011/4/2 何闯 justhechu...@163.com: Hi all, As Gene suggested, I browsed vesamenu code and found nothing useful. However, I figured out how to use the 640X480 mode in mode 12h. It turns out that one pixel takes one bit place. Codes are: // void setvga() { __asm__ __volatile__ ( REAL_CODE ( movw $0x12, %%ax\n\t int $0x10\n\t)::); } void putpixel(int x,int y, int c __unused ) { char v; char idx; if(vediobuf == NULL) { vediobuf = (unsigned char*)phys_to_user(0xa); } v = *(vediobuf+x/8+80*y); idx = x % 8; //One byte divided into eight bits for pixels switch(idx) { case 0: *(vediobuf+x/8+80*y) = (17)|v; break; case 1: *(vediobuf+x/8+80*y) = (16)|v; break; case 2: *(vediobuf+x/8+80*y) = (15)|v; break; case 3: *(vediobuf+x/8+80*y) = (14)|v; break; case 4: *(vediobuf+x/8+80*y) = (13)|v; break; case 5: *(vediobuf+x/8+80*y) = (12)|v; break; case 6: *(vediobuf+x/8+80*y) = (11)|v; break; case 7: *(vediobuf+x/8+80*y) = (10)|v; break; } } And this works, I finally entered 640X480 mode!!! Nevertheless, one bit can only hold two colors(black and white), but document says that video mode 12h will enter 640X480X16 resolution, which means 4bit for one pixel, this conflicts with my practise. I am confused again.. Your video card might not know mode 12h and used 11h instead. Also, I'd advise against writing the video buffer directly unless you've asked your video card where its buffer is. See also INT 10h AH=0Ch, AH=0Dh and AH=0Fh. At 2011-04-02,Gene Cumm gene.c...@gmail.com wrote: On Fri, Apr 1, 2011 at 15:15, Vaza gpxe vaza.g...@gmail.com wrote: I'd suggest you find out how many bytes there are per pixel. I guess you have 4 (not 1 you are using) and the resolution you think you see is actually 640x120. You may also have limited video memory that makes different resolutions use different pixel sizes I believe there should be only 4 bits at that video mode. Soforth, I'd suggest examining http://www.ctyme.com/intr/int-10.htm for starters. I believe INT 10h AX 4F00h is the start of VESA related info that Syslinux uses for vesamenu.c32 (and possibly gfxboot.c32) -- -Gene In Reply to: 2011/4/1 何闯 justhechu...@163.com: Deal all, I just tested int 0x10 BIOS interrupt to draw some pictures. However, when I coded the follow to enter 640X480 video mode, it seems that the actual video mode is 640X100, I draw a rectangle with range(0,0, 640, 480) to verify, and it resulted in range(0,0, 640,100), about a quarter square of a rectangle. And I noticed that 320X200 video mode is OK, and 320X200 == 640X100, is that mean my machine can hand only 64000 pixels? Can anyone tell me why? /// void setvga() { __asm__ __volatile__ ( REAL_CODE ( movw $0x4f02, %%ax\n\t //ax=0x0012 here seem stay in the text mode movw $0x0101, %%bx\n\t int $0x10\n\t )::); } void putpixel(int x,int y,int c) { if(vediobuf == NULL) { vediobuf = (unsigned char*)phys_to_user(0xa); //physic address of graphic mode } *(vediobuf+x+640*y)=c; } / yours, soforth I'd suggest looking at other boot loaders and how they write to the screen (including Syslinux's vesamenu.c32 and gfxboot.c32). 64,000 sounds like the 16-bit limit of 65,536 per segment. -- -Gene No one ever says, 'I can't read that ASCII(plain text) e-mail you sent me.' ___ gPXE mailing list gPXE@etherboot.org http://etherboot.org/mailman/listinfo/gpxe
Re: [gPXE] int 10h problem
2011/4/1 何闯 justhechu...@163.com: Deal all, I just tested int 0x10 BIOS interrupt to draw some pictures. However, when I coded the follow to enter 640X480 video mode, it seems that the actual video mode is 640X100, I draw a rectangle with range(0,0, 640, 480) to verify, and it resulted in range(0,0, 640,100), about a quarter square of a rectangle. And I noticed that 320X200 video mode is OK, and 320X200 == 640X100, is that mean my machine can hand only 64000 pixels? Can anyone tell me why? /// void setvga() { __asm__ __volatile__ ( REAL_CODE ( movw $0x4f02, %%ax\n\t //ax=0x0012 here seem stay in the text mode movw $0x0101, %%bx\n\t int $0x10\n\t )::); } void putpixel(int x,int y,int c) { if(vediobuf == NULL) { vediobuf = (unsigned char*)phys_to_user(0xa); //physic address of graphic mode } *(vediobuf+x+640*y)=c; } / yours, soforth I'd suggest looking at other boot loaders and how they write to the screen (including Syslinux's vesamenu.c32 and gfxboot.c32). 64,000 sounds like the 16-bit limit of 65,536 per segment. -- -Gene ___ gPXE mailing list gPXE@etherboot.org http://etherboot.org/mailman/listinfo/gpxe
Re: [gPXE] int 10h problem
On Fri, Apr 1, 2011 at 15:15, Vaza gpxe vaza.g...@gmail.com wrote: I'd suggest you find out how many bytes there are per pixel. I guess you have 4 (not 1 you are using) and the resolution you think you see is actually 640x120. You may also have limited video memory that makes different resolutions use different pixel sizes I believe there should be only 4 bits at that video mode. Soforth, I'd suggest examining http://www.ctyme.com/intr/int-10.htm for starters. I believe INT 10h AX 4F00h is the start of VESA related info that Syslinux uses for vesamenu.c32 (and possibly gfxboot.c32) -- -Gene In Reply to: 2011/4/1 何闯 justhechu...@163.com: Deal all, I just tested int 0x10 BIOS interrupt to draw some pictures. However, when I coded the follow to enter 640X480 video mode, it seems that the actual video mode is 640X100, I draw a rectangle with range(0,0, 640, 480) to verify, and it resulted in range(0,0, 640,100), about a quarter square of a rectangle. And I noticed that 320X200 video mode is OK, and 320X200 == 640X100, is that mean my machine can hand only 64000 pixels? Can anyone tell me why? /// void setvga() { __asm__ __volatile__ ( REAL_CODE ( movw $0x4f02, %%ax\n\t //ax=0x0012 here seem stay in the text mode movw $0x0101, %%bx\n\t int $0x10\n\t )::); } void putpixel(int x,int y,int c) { if(vediobuf == NULL) { vediobuf = (unsigned char*)phys_to_user(0xa); //physic address of graphic mode } *(vediobuf+x+640*y)=c; } / yours, soforth I'd suggest looking at other boot loaders and how they write to the screen (including Syslinux's vesamenu.c32 and gfxboot.c32). 64,000 sounds like the 16-bit limit of 65,536 per segment. ___ gPXE mailing list gPXE@etherboot.org http://etherboot.org/mailman/listinfo/gpxe
Re: [gPXE] Need help with gPXE set up
On Tue, Mar 29, 2011 at 01:31, Pandey, Hariom Sharan hariom-sharan.pan...@hp.com wrote: Hello Team, We are setting up PXE boot using NX3031 NIC card which is now coming with gPXE code. NexXen FlexLOM. Nice. Our servers are using DHCP assigned IP for PXE boot and works great for PXE, using gPXE I am coming across through timeout error, please refer the screen shot. My questions are: 1. How exactly to start OS installation using gPXE? Just like any PXE other boot system. However, gPXE has additional enhancements including scripting, HTTP/FTP download, iSCSI attach and probably others I can't think of. 2. Does it require any specific settings on DHCP server and what changes need to be done on client side. This depends on what you want to do. 3. I am also not able to see the ctrl-B, so not sure how to get into cmd. Using cmd what I need to do on client side to boot through PXE. It may be that it's been disabled in that build. 4. Is there any option to bring gPXE back to normal PXE. I'd guess that the card you have has gPXE as its onboard PXE ROM. -- -Gene ___ gPXE mailing list gPXE@etherboot.org http://etherboot.org/mailman/listinfo/gpxe
Re: [gPXE] unicode support in gPXE
2011/3/29 何闯 justhechu...@163.com: What are you looking to do? Use the echo command within a script to print out something for an end user to see perhaps? Anything after gPXE boots to something else (ie Linux kernel, an NBP (Network Boot Program) like PXELINUX, SAN disk) will depend on that target as gPXE will no longer have control. Thanks for your quick response. I see that menu in syslinux supports unicode with very beautiful background pcitures. Can welcome message in gPXE have a chinese version? We may be *VERY* excited with that!!! I'm not sure what you mean. To my knowledge, Syslinux's menu.c32 also depends on the underlying BIOS. It's possible that you may be referring to gfxboot, a graphical boot selection system whose intention (I believe) was to make it possible to create the same looking menu in Syslinux, Grub and Lilo. As implemented with Syslinux, there has been a significant change from an assembly patch outside of Syslinux's package to a COMBOOT module to a COM32 module. http://syslinux.zytor.com/wiki/index.php/GFXBoot.c32 gives some background. If you could provide a reference, perhaps some light can be shed on the topic. Pictures might do but the addition of other configuration details (even by links to web pages) would help more solidly determine the situation behind what you have seen. -- -Gene ___ gPXE mailing list gPXE@etherboot.org http://etherboot.org/mailman/listinfo/gpxe
Re: [gPXE] unicode support in gPXE
Reverting to plain text (although I think unicode) for archive/search and in case the mail gets blocked by the mail server. 2011/3/29 何闯 justhechu...@163.com If you could provide a reference, perhaps some light can be shed on the topic. Pictures might do but the addition of other configuration details (even by links to web pages) would help more solidly determine the situation behind what you have seen. -- -Gene What I want is like this effect in gPXE: 1. AoE 启动 2.iSCSI 启动 3.HTTP 启动 please choose: [1~3] sample pictures: Thanks for the picture. Along the bottom it plainly states GFXBOOT. Knowing what the other Syslinux menu systems look like, I'm pretty sure it's not one of those. gPXE does not really have an end-user selection system however it does load PXELINUX rather easily and PXELINUX can provide a menu through menu.c32/vesamenu.c32/gfxboot.c32 (or the cmenu system). 2011/3/30 何闯 justhechu...@163.com If possible, I want do something for this. Please give me suggestions on how to achieve that. yours, soforth To achieve what you see in that picture though network booting, the sequence would be gPXE (loaded from local media, including a ROM, or via PXE) loading PXELINUX loading gfxboot.c32 with any needed parts. You'll want to look at gpxecmd.c32 (the COM32 module to issue commands back to gPXE like sanboot) and you'll really want a recent version of the Syslinux package (either a testing version of 4.04-pre14 or 3.86; older 4.xx have a bug solved in recent testing versions). There have been several enhancements to gfxboot.c32 since 3.86 that would lean me towards 4.04-pre* (hopefully 4.04 will be released soon). See http://syslinux.zytor.com -- -Gene ___ gPXE mailing list gPXE@etherboot.org http://etherboot.org/mailman/listinfo/gpxe
Re: [gPXE] Boot from remote SAN as quickly as possible
On Sat, Nov 13, 2010 at 05:05, Erik Loman e...@surfright.com wrote: Hello all, I am new to this mailing list and I have several questions. I saw a YouTube video (from Phoenix) were they show how to boot Windows 7 in 10 seconds. I want to do that as well, but this time using a diskless PC and boot from a remote SAN. I plan to use Intel Atom boards with EFI bios that support Intel Rapid BIOS Boot to minimize bios boot time (Intel D510MO). Could you reference which video? http://www.youtube.com/watch?v=z9E_Qri8b2E perhaps? This sounds very interesting. I have several questions: 1. Should I either use iSCSI or AoE? Which one is faster? (the machines are on a local network) AoE doesn't have the overhead of iSCSI (TCP headers, TCP checksum, etc) but it'll probably also depend on buffering (both sides), in-flight transactions, the hardware backing the target and other optimizations. At least the board you referenced has a Gigabit LoM for a possibility of speed. 2. Which iSCSI or AoE target software (for Windows) supports non-persistent writes? Which one do you recommend? (either commercial or free software) I'm not sure if you really want non-persistent writes so much as something like a COW (Copy-On-Write) or snapshot as you'll need to define when the writes are forgotten. If you do it based on attaching to the disk, a network interruption will probably cause the the client OS to freak out and crash as the disk will not be in the state it expects. http://www.etherboot.org/wiki/appnotes/cow is one AoE COW example for Linux. I'm not familiar with any Windows targets. 3. I read that gPXE uses DHCP to determine how and what file to boot. Is using a boot script significantly faster? Generally, a script should be faster, especially if you statically configure your network interface and statically define your boot target. However, it'll depend on how you load gPXE in the first place. Using the undionly driver means you PCs will PXE boot (1 DHCP round) but you won't have any options outside of what PXE normally requests (ie root-path) unless your DHCP daemon forcibly adds unrequested options (in which case you'll want gPXE to do a second DHCP round). Loading gPXE (with appropriate drivers) from local media (ie, UFD; USB Flash Drive) could eliminate the first round of DHCP. I know, many questions. But if you have an answer to at least one of these questions I am very grateful. Best regards, Erik -- -Gene ___ gPXE mailing list gPXE@etherboot.org http://etherboot.org/mailman/listinfo/gpxe
Re: [gPXE] is it bug or feature for gPXE?
On Thu, Nov 11, 2010 at 02:53, Andrew Bobulsky rule...@gmail.com wrote: Hello Oleg, I can't say with any certainty what your problem is, but I have personally had to deal with DHCP problems that were impossible to solve without doing some packet capture. For the benefit of those who will likely read your email in the morning, my only suggestion is that you load up Wireshark and capture the traffic occurring during the test scenario you're describing. I've found that the cause of funny DHCP client behavior is usually easy to pinpoint, even if it's not easy to fix. Cheers, Andrew Bobulsky This is definitely a good idea. Capture all packets going to/from the machine that is using gPXE. On Nov 11, 2010, at 2:32 AM, oleg-gumenyuk oleg-gumen...@yandex.ru wrote: Hello ALL! Historically our customer has got network where 2 DHCP servers are up and running for redundancy purpose. DHCPD config file is common for both servers. No issues if we boot down the wire WinPE, any Linux. We observe issue when gPXE is booted. I find it understandable that there are two as I have a similar situation at my work. As long as they don't try to assign the same IP to two different hosts, this is generally good. One way to prevent this issue is to configure them to hand out a different pool of IP addresses in the same subnet. Another is that the two must some how coordinate to prevent this issue (this coordination can be more complex and is DHCPD-specific). When gPXE requests IP address both DHCP servers send responces. The concequence of such behaviour is that gPXE lost network settings. We have localized problem to the following small scenario: This is a very nice small scenario. 1. boot to gPXE 1.0.1 on host A and enter to the interactive mode (Ctrl-B). 2. from another host B start pinging IP address which will be assigned to the host A How do you know that host A will be assigned a given IP? Is this a lease reservation configured in both servers? Is it a pre-existing unexpired lease that both servers are trying to use? 3. execute on Host A gPXE command: dhcp net0 4. ping from host B are successful. ...waiting a little bit about 5-7 secs Host A is not pingable. Does it drop just a few pings (even for something as extreme as 30 seconds) or does it never respond again? Have you ensured that there are no new hosts that are attempting to obtain a new DHCP lease that might steal that IP? Have you tried running route (with no parameters) from gPXE just after running dhcp net0 and after host A stops responding to pings? Make sure it says the same info both times. It should show something like: net0: 10.0.0.6/255.255.255/0 gw 10.0.0.1 If we stop one of DHCP server such issue is eliminated. I'd assume you've tried them individually with identical results? Stop DHCP server C, run test, start DHCP server C, stop DHCP server D and run the test again. Question: is it a bug of gPXE? _or_ Is it a feature and our customer breaks some RFCs? Thanks, Oleg. Another idea is to also examine the logs of the DHCP servers. This may be easier to start doing than performing the packet capture as the packet capture may require some setup to get it to capture the relevant packets however analyzing the packet capture will probably reveal the issue quickly and more reliably. -- -Gene ___ gPXE mailing list gPXE@etherboot.org http://etherboot.org/mailman/listinfo/gpxe
Re: [gPXE] is it bug or feature for gPXE?
2010/11/11 oleg-gumenyuk oleg-gumen...@yandex.ru: Hello Gene, You have asked very valid questions. Sorry, I have not clarified these aspects in my initial email. How do you know that host A will be assigned a given IP? we use something like this host glab-56-25 { hardware ethernet 00:0C:29:04:2F:28; fixed-address glab-56-25; } In the DNS we have A record glab-56-25 - 172.28.56.25 So, there is no way to stil IP address. Good. That eliminates one possible issue Does it drop just a few pings (even for something as extreme as 30 seconds) or does it never respond again? Never I'd assume you've tried them individually with identical results? Stop DHCP server C, run test, start DHCP server C, stop DHCP server D and run the test again. Frankly speaking NO, due to the fact that servers are idential. The same Linux, the same dhcpcd, different host with same HW. It might be worth a shot, just in case there's an issue with one of them (doubtful, though). Nothing suspiciuos in the dhcpd logs. It's quick and easy but I figured it would probably not yield anything useful. Thanks, Oleg. PS: I've attached screenshot with gPXE in the previous response in the mailthread. I saw. On Thu, Nov 11, 2010 at 07:08, oleg-gumenyuk oleg-gumen...@yandex.ru wrote: Hello, I'll try to do the packet capturing. At this moment it is not easy due to specific of the lab environment where set of boards and bunch of ESXi with virtual systems ar deployed. tcpdump on ESXi will be ideal but imposible. :) http://kb.vmware.com/kb/1000880 I'm not sure if there's another KB article that's more relevant to 4.x (in case you're running 4.x). This article makes me think that with changes to the vSwitch in ESXi, another guest may be able to run the capture. I've never tried this before but it seems plausible. I'd like to calify that dhcpd.conf is configured without pool of IP address. We can treat as bootp. So, it is not possible to have IP address conflict, we always know which IP address will be assigned to the respective MAC address. Thank you for that clarification. Also I've attached screenshot of gPXE commands. Pings from Host B never resumes. Interesting. gPXE closes the interface. What happens if you try dhcp net0 again after noticing that the interface is closed? Oleg. If you're comfortable compiling gPXE, you may want to consider it but I'm not sure which DEBUG options you'll want to enable. 11.11.10, 14:43, Gene Cumm gene.c...@gmail.com: On Thu, Nov 11, 2010 at 02:53, Andrew Bobulsky wrote: Hello Oleg, I can't say with any certainty what your problem is, but I have personally had to deal with DHCP problems that were impossible to solve without doing some packet capture. For the benefit of those who will likely read your email in the morning, my only suggestion is that you load up Wireshark and capture the traffic occurring during the test scenario you're describing. I've found that the cause of funny DHCP client behavior is usually easy to pinpoint, even if it's not easy to fix. Cheers, Andrew Bobulsky This is definitely a good idea. Capture all packets going to/from the machine that is using gPXE. On Nov 11, 2010, at 2:32 AM, oleg-gumenyuk wrote: Hello ALL! Historically our customer has got network where 2 DHCP servers are up and running for redundancy purpose. DHCPD config file is common for both servers. No issues if we boot down the wire WinPE, any Linux. We observe issue when gPXE is booted. I find it understandable that there are two as I have a similar situation at my work. As long as they don't try to assign the same IP to two different hosts, this is generally good. One way to prevent this issue is to configure them to hand out a different pool of IP addresses in the same subnet. Another is that the two must some how coordinate to prevent this issue (this coordination can be more complex and is DHCPD-specific). When gPXE requests IP address both DHCP servers send responces. The concequence of such behaviour is that gPXE lost network settings. We have localized problem to the following small scenario: This is a very nice small scenario. 1. boot to gPXE 1.0.1 on host A and enter to the interactive mode (Ctrl-B). 2. from another host B start pinging IP address which will be assigned to the host A How do you know that host A will be assigned a given IP? Is this a lease reservation configured in both servers? Is it a pre-existing unexpired lease that both servers are trying to use? 3. execute on Host A gPXE command: dhcp net0 4. ping from host B are successful. ...waiting a little bit about 5-7 secs Host A is not pingable. Does it drop just a few pings (even for something as extreme as 30 seconds) or does it never respond again? Have you ensured that there are no new hosts
Re: [gPXE] How to modify the code can not get DHCP information on the2nd
2010/11/6 sqwbwh sqw...@126.com: Gene Cumm thank you let I see reasons for this situation occurs. Want to solve this problem. How to do that? Forcing your DHCP server to always return option 17 is application specific. http://etherboot.org/wiki/howtos#how_to_use_advanced_dhcp_options http://etherboot.org/wiki/appnotes/pxelinux_tftp_to_http_migration#original_dhcpd_setup This person forcibly appended 3a and 3b to dhcp-parameter-request-list. This is what you'll need to do. I think it's a single byte in hex so 11 should be your option 17 decimal. -- -Gene At 2010-11-07 02:49:03,Gene Cumm gene.c...@gmail.com wrote: 2010/11/6 sqwbwh sqw...@126.com: Miller, Shao use Embedded script #!gpxe set use-cached 1 GPXE not have access to 017 settings How can either use the set use-cached 1 Also get 017 information? Your native PXE client probably did not request it and/or ignored it when storing the cached DHCP packets. If you do a packet capture (very easy in a hosted VMware environment), I'd guess you probably won't see option 17 in the response from the server. It is possible to configure some DHCP servers to always return certain options. However, when you don't set use-cached 1, the DHCP server is probably sending it because gPXE is probably requesting it. info: Dhcp server set up in the 017 options iscsi: 192.168.0.222:: 3260:: test I set it in an embedded script: #!gpxe set use-cached 1 autoboot please see 0171.jpg -0172.jpg root-path is nothing 0172.jpg is add script : #!gpxe set use-cached 1 config - normal.jpg is no script At 2010-11-05 21:13:33,Miller, Shao shao.mil...@yrdsb.edu.on.ca wrote: From: gpxe-boun...@etherboot.org [mailto:gpxe-boun...@etherboot.org] on behalf of sqwbwh Sent: Friday, November 05, 2010 09:05 To: GPXE Mailing List Subject: [gPXE] How to modify the code can not get DHCP information on the2nd How to modify the code can not get DHCP information on the 2nd?(GPXE only used to boot the system from ISCSI) Is to remove this step. net0: 00:0 c:: 29:08:11:35 on UNDI-PCI02: 01.0 (OPEN) [Link: up, TX: 0 TXE: 0 TX: 0 rxe: 0] DHCP (net0 00:0 c: 29:08:11: d5 ) ok net0: 192.168.0.202/255.255.255.0 This can speed up boot speed. Normal GPXE run to start from the boot system tftp process takes about 8 seconds or so If removed from the second time for DHCP information will only need a very quick 2.5 seconds. - You most likely want the 'use-cached' option. It is DHCP gPXE encapsulated option number 178. Or, you can set it in an embedded script: #!gpxe set use-cached 1 autoboot - Shao Miller ___ gPXE mailing list gPXE@etherboot.org http://etherboot.org/mailman/listinfo/gpxe
Re: [gPXE] How to modify the code can not get DHCP information on the2nd
2010/11/7 sqwbwh sqw...@126.com: Gene Cumm Forcing your DHCP server to always return option 17 is application specific. MSDHCP always return option 17 is application specific What practices can explain it in detail? -2003 http://technet.microsoft.com/en-us/library/cc778368%28WS.10%29.aspx -2008 http://technet.microsoft.com/en-us/library/cc755282.aspx Or find the equivalent page in your preferred language. Unfortunately, I have no idea if it's possible to force MS-DHCP to do this. You could put the option into a gPXE script and embed it into your gPXE binary. -- -Gene At 2010-11-07 19:26:31,Gene Cumm gene.c...@gmail.com wrote: 2010/11/6 sqwbwh sqw...@126.com: Gene Cumm thank you let I see reasons for this situation occurs. Want to solve this problem. How to do that? Forcing your DHCP server to always return option 17 is application specific. http://etherboot.org/wiki/howtos#how_to_use_advanced_dhcp_options http://etherboot.org/wiki/appnotes/pxelinux_tftp_to_http_migration#original_dhcpd_setup This person forcibly appended 3a and 3b to dhcp-parameter-request-list. This is what you'll need to do. I think it's a single byte in hex so 11 should be your option 17 decimal. -- -Gene At 2010-11-07 02:49:03,Gene Cumm gene.c...@gmail.com wrote: 2010/11/6 sqwbwh sqw...@126.com: Miller, Shao use Embedded script #!gpxe set use-cached 1 GPXE not have access to 017 settings How can either use the set use-cached 1 Also get 017 information? Your native PXE client probably did not request it and/or ignored it when storing the cached DHCP packets. If you do a packet capture (very easy in a hosted VMware environment), I'd guess you probably won't see option 17 in the response from the server. It is possible to configure some DHCP servers to always return certain options. However, when you don't set use-cached 1, the DHCP server is probably sending it because gPXE is probably requesting it. info: Dhcp server set up in the 017 options iscsi: 192.168.0.222:: 3260:: test I set it in an embedded script: #!gpxe set use-cached 1 autoboot please see 0171.jpg -0172.jpg root-path is nothing 0172.jpg is add script : #!gpxe set use-cached 1 config - normal.jpg is no script At 2010-11-05 21:13:33,Miller, Shao shao.mil...@yrdsb.edu.on.ca wrote: From: gpxe-boun...@etherboot.org [mailto:gpxe-boun...@etherboot.org] on behalf of sqwbwh Sent: Friday, November 05, 2010 09:05 To: GPXE Mailing List Subject: [gPXE] How to modify the code can not get DHCP information on the2nd How to modify the code can not get DHCP information on the 2nd?(GPXE only used to boot the system from ISCSI) Is to remove this step. net0: 00:0 c:: 29:08:11:35 on UNDI-PCI02: 01.0 (OPEN) [Link: up, TX: 0 TXE: 0 TX: 0 rxe: 0] DHCP (net0 00:0 c: 29:08:11: d5 ) ok net0: 192.168.0.202/255.255.255.0 This can speed up boot speed. Normal GPXE run to start from the boot system tftp process takes about 8 seconds or so If removed from the second time for DHCP information will only need a very quick 2.5 seconds. - You most likely want the 'use-cached' option. It is DHCP gPXE encapsulated option number 178. Or, you can set it in an embedded script: #!gpxe set use-cached 1 autoboot - Shao Miller ___ gPXE mailing list gPXE@etherboot.org http://etherboot.org/mailman/listinfo/gpxe
Re: [gPXE] At compile time how to add Embedded Script:
2010/11/6 sqwbwh sqw...@126.com: At compile time how to add Embedded Script: Such as add Embedded Script : #! Gpxe dhcp net0 autoboot Should be written where? I believe autoboot is a special command that tries to look to the network for everything and already does dhcp net0 (See also http://etherboot.org/wiki/commandline#other ). Try the # (optional) Customize image configuration options: (Customize button) and you get an input box Embedded Script:. Embedded Script sleep This command failed yet? I believe TIME_CMD is disabled by default. (Again Customize). Where can I find detailed information on the Embedded Script http://etherboot.org/wiki/commandline -- -Gene ___ gPXE mailing list gPXE@etherboot.org http://etherboot.org/mailman/listinfo/gpxe
Re: [gPXE] Fw:Re:Re: At compile time how to add Embedded Script:
2010/11/6 sqwbwh sqw...@126.com Thank you for forwarding back to the list. Forwarding messages From: sqwbwh sqw...@126.com Date: 2010-11-06 17:45:13 To: Gene Cumm gene.c...@gmail.com Subject: Re:Re: [gPXE] At compile time how to add Embedded Script: Gene Cumm I believe TIME_CMD is disabled by default. (Again Customize). http://etherboot.org/wiki/commandline -- I see,thank you #! Gpxe Dhcp net0 Autoboot This is just an example If it is used to compile the source code do?(add Embedded Script,use make ) Not rom-o-matic.net make [target] EMBEDDED_IMAGE=[path_to_script] TIME_CMD=1 This is what I've used (of course replacing [target] with something like bin/ne2k_isa.dsk). -sqwbwh At 2010-11-06 17:31:50,Gene Cumm gene.c...@gmail.com wrote: 2010/11/6 sqwbwh sqw...@126.com: At compile time how to add Embedded Script: Such as add Embedded Script : #! Gpxe dhcp net0 autoboot Should be written where? I believe autoboot is a special command that tries to look to the network for everything and already does dhcp net0 (See also http://etherboot.org/wiki/commandline#other ). Try the # (optional) Customize image configuration options: (Customize button) and you get an input box Embedded Script:. Embedded Script sleep This command failed yet? I believe TIME_CMD is disabled by default. (Again Customize). Where can I find detailed information on the Embedded Script http://etherboot.org/wiki/commandline -- -Gene -- -Gene No one ever says, 'I can't read that ASCII(plain text) e-mail you sent me.' ___ gPXE mailing list gPXE@etherboot.org http://etherboot.org/mailman/listinfo/gpxe
Re: [gPXE] Fw:Re:Re: Fw:Re:Re: At compile time how to add Embedded Script:
2010/11/6 sqwbwh sqw...@126.com: Over the last few years, I've gotten in the habit of always hitting reply all, even on personal messages, with the different lists I'm on. Forwarding messages From: sqwbwh sqw...@126.com Date: 2010-11-06 18:14:08 To: Gene Cumm gene.c...@gmail.com Subject: Re:Re: [gPXE] Fw:Re:Re: At compile time how to add Embedded Script: Gene Cumm #! Gpxe Dhcp net0 Autoboot This is just an example If it is used to compile the source code do?(add Embedded Script,use make ) Not rom-o-matic.net make [target] EMBEDDED_IMAGE=[path_to_script] TIME_CMD=1 This is what I've used (of course replacing [target] with something like bin/ne2k_isa.dsk). - Here I do not understand == http://rom-o-matic.net/gpxe/gpxe-1.0.1/contrib/rom-o-matic/build.php at Embedded Script: add: #!gpxe dhcp net0 autoboot To achieve the above functions How to modify the source code and compile If you're looking to modify the source and compile it yourself, put the contents of your script into a text file then look at what I did for my make line. This is one that I've actually used: make bin/ne2k_isa.dsk EMBEDDED_IMAGE=../../bochs-vnet.gpxe TIME_CMD=1 I made a script, bochs-vnet.gpxe, just for a special purpose. -sqwbwh At 2010-11-06 17:55:59,Gene Cumm gene.c...@gmail.com wrote: 2010/11/6 sqwbwh sqw...@126.com Thank you for forwarding back to the list. Forwarding messages From: sqwbwh sqw...@126.com Date: 2010-11-06 17:45:13 To: Gene Cumm gene.c...@gmail.com Subject: Re:Re: [gPXE] At compile time how to add Embedded Script: Gene Cumm I believe TIME_CMD is disabled by default. (Again Customize). http://etherboot.org/wiki/commandline -- I see,thank you #! Gpxe Dhcp net0 Autoboot This is just an example If it is used to compile the source code do?(add Embedded Script,use make ) Not rom-o-matic.net make [target] EMBEDDED_IMAGE=[path_to_script] TIME_CMD=1 This is what I've used (of course replacing [target] with something like bin/ne2k_isa.dsk). -sqwbwh At 2010-11-06 17:31:50,Gene Cumm gene.c...@gmail.com wrote: 2010/11/6 sqwbwh sqw...@126.com: At compile time how to add Embedded Script: Such as add Embedded Script : #! Gpxe dhcp net0 autoboot Should be written where? I believe autoboot is a special command that tries to look to the network for everything and already does dhcp net0 (See also http://etherboot.org/wiki/commandline#other ). Try the # (optional) Customize image configuration options: (Customize button) and you get an input box Embedded Script:. Embedded Script sleep This command failed yet? I believe TIME_CMD is disabled by default. (Again Customize). Where can I find detailed information on the Embedded Script http://etherboot.org/wiki/commandline -- -Gene No one ever says, 'I can't read that ASCII(plain text) e-mail you sent me.' -- -Gene ___ gPXE mailing list gPXE@etherboot.org http://etherboot.org/mailman/listinfo/gpxe
Re: [gPXE] Fw:Re:Re: Fw:Re:Re: At compile time how to add Embedded Script:
2010/11/6 sqwbwh sqw...@126.com Gene Cumm I see, thank you very much~:) -sqwbwh I'm glad I could help. At 2010-11-06 18:24:23,Gene Cumm gene.c...@gmail.com wrote: 2010/11/6 sqwbwh sqw...@126.com: Over the last few years, I've gotten in the habit of always hitting reply all, even on personal messages, with the different lists I'm on. Forwarding messages From: sqwbwh sqw...@126.com Date: 2010-11-06 18:14:08 To: Gene Cumm gene.c...@gmail.com Subject: Re:Re: [gPXE] Fw:Re:Re: At compile time how to add Embedded Script: Gene Cumm #! Gpxe Dhcp net0 Autoboot This is just an example If it is used to compile the source code do?(add Embedded Script,use make ) Not rom-o-matic.net make [target] EMBEDDED_IMAGE=[path_to_script] TIME_CMD=1 This is what I've used (of course replacing [target] with something like bin/ne2k_isa.dsk). - Here I do not understand == http://rom-o-matic.net/gpxe/gpxe-1.0.1/contrib/rom-o-matic/build.php at Embedded Script: add: #!gpxe dhcp net0 autoboot To achieve the above functions How to modify the source code and compile If you're looking to modify the source and compile it yourself, put the contents of your script into a text file then look at what I did for my make line. This is one that I've actually used: make bin/ne2k_isa.dsk EMBEDDED_IMAGE=../../bochs-vnet.gpxe TIME_CMD=1 I made a script, bochs-vnet.gpxe, just for a special purpose. -sqwbwh At 2010-11-06 17:55:59,Gene Cumm gene.c...@gmail.com wrote: 2010/11/6 sqwbwh sqw...@126.com Thank you for forwarding back to the list. Forwarding messages From: sqwbwh sqw...@126.com Date: 2010-11-06 17:45:13 To: Gene Cumm gene.c...@gmail.com Subject: Re:Re: [gPXE] At compile time how to add Embedded Script: Gene Cumm I believe TIME_CMD is disabled by default. (Again Customize). http://etherboot.org/wiki/commandline -- I see,thank you #! Gpxe Dhcp net0 Autoboot This is just an example If it is used to compile the source code do?(add Embedded Script,use make ) Not rom-o-matic.net make [target] EMBEDDED_IMAGE=[path_to_script] TIME_CMD=1 This is what I've used (of course replacing [target] with something like bin/ne2k_isa.dsk). -sqwbwh At 2010-11-06 17:31:50,Gene Cumm gene.c...@gmail.com wrote: 2010/11/6 sqwbwh sqw...@126.com: At compile time how to add Embedded Script: Such as add Embedded Script : #! Gpxe dhcp net0 autoboot Should be written where? I believe autoboot is a special command that tries to look to the network for everything and already does dhcp net0 (See also http://etherboot.org/wiki/commandline#other ). Try the # (optional) Customize image configuration options: (Customize button) and you get an input box Embedded Script:. Embedded Script sleep This command failed yet? I believe TIME_CMD is disabled by default. (Again Customize). Where can I find detailed information on the Embedded Script http://etherboot.org/wiki/commandline -- -Gene No one ever says, 'I can't read that ASCII(plain text) e-mail you sent me.' -- -Gene -- -Gene No one ever says, 'I can't read that ASCII(plain text) e-mail you sent me.' ___ gPXE mailing list gPXE@etherboot.org http://etherboot.org/mailman/listinfo/gpxe
Re: [gPXE] How to modify the code can not get DHCP information on the2nd
2010/11/6 sqwbwh sqw...@126.com: Miller, Shao use Embedded script #!gpxe set use-cached 1 GPXE not have access to 017 settings How can either use the set use-cached 1 Also get 017 information? Your native PXE client probably did not request it and/or ignored it when storing the cached DHCP packets. If you do a packet capture (very easy in a hosted VMware environment), I'd guess you probably won't see option 17 in the response from the server. It is possible to configure some DHCP servers to always return certain options. However, when you don't set use-cached 1, the DHCP server is probably sending it because gPXE is probably requesting it. info: Dhcp server set up in the 017 options iscsi: 192.168.0.222:: 3260:: test I set it in an embedded script: #!gpxe set use-cached 1 autoboot please see 0171.jpg -0172.jpg root-path is nothing 0172.jpg is add script : #!gpxe set use-cached 1 config - normal.jpg is no script At 2010-11-05 21:13:33,Miller, Shao shao.mil...@yrdsb.edu.on.ca wrote: From: gpxe-boun...@etherboot.org [mailto:gpxe-boun...@etherboot.org] on behalf of sqwbwh Sent: Friday, November 05, 2010 09:05 To: GPXE Mailing List Subject: [gPXE] How to modify the code can not get DHCP information on the2nd How to modify the code can not get DHCP information on the 2nd?(GPXE only used to boot the system from ISCSI) Is to remove this step. net0: 00:0 c:: 29:08:11:35 on UNDI-PCI02: 01.0 (OPEN) [Link: up, TX: 0 TXE: 0 TX: 0 rxe: 0] DHCP (net0 00:0 c: 29:08:11: d5 ) ok net0: 192.168.0.202/255.255.255.0 This can speed up boot speed. Normal GPXE run to start from the boot system tftp process takes about 8 seconds or so If removed from the second time for DHCP information will only need a very quick 2.5 seconds. - You most likely want the 'use-cached' option. It is DHCP gPXE encapsulated option number 178. Or, you can set it in an embedded script: #!gpxe set use-cached 1 autoboot - Shao Miller -- -Gene ___ gPXE mailing list gPXE@etherboot.org http://etherboot.org/mailman/listinfo/gpxe
Re: [gPXE] How to modify the code can not get DHCP information on the 2nd
2010/11/5 sqwbwh sqw...@126.com: How to modify the code can not get DHCP information on the 2nd?(GPXE only used to boot the system from ISCSI) Is to remove this step. net0: 00:0 c:: 29:08:11:35 on UNDI-PCI02: 01.0 (OPEN) [Link: up, TX: 0 TXE: 0 TX: 0 rxe: 0] DHCP (net0 00:0 c: 29:08:11: d5 ) ok net0: 192.168.0.202/255.255.255.0 This can speed up boot speed. Normal GPXE run to start from the boot system tftp process takes about 8 seconds or so If removed from the second time for DHCP information will only need a very quick 2.5 seconds. What file and version are you using? It sounds as if you're using undionly.kpxe or a .pxe file. undionly.kkpxe will skip this second DHCP request but certain hardware/firmware has issues with this, assuming you trust the stability of your OEM's PXE/UNDI stack. -- -Gene ___ gPXE mailing list gPXE@etherboot.org http://etherboot.org/mailman/listinfo/gpxe
Re: [gPXE] Pushing GPXE script to HTTP server
On Thu, Nov 4, 2010 at 02:09, htru...@tnhh.net wrote: Today I had a chance to test and implement David's suggestions. It worked well for what I wanted if I do single boot, however I couldn't get the multi-boot menu to work. So my gpxe.conf looks like this: #!gpxe chain vesamenu.c32 multibootmenu.conf I would have to assume that this is the chainloaded script (chainload http://host/script; from a previous message in thread). The station tries to switch graphics mode for a second and reboots. I guess that's because (vesa)menu.c32 of Syslinux couldn't find multibootmenu.conf (or whatever file it was supposed to read as its argument). I placed multibootmenu.conf and (vesa)menu.c32 in the same directory as gpxe.conf (which is hard-coded to the ROM). I wonder if there are tricks to make (vesa)menu.c32 to be able to read multibootmenu.conf. I used vesamenu.c32 from syslinux 4.01. I would recommend against using vesamenu.c32 from v4.00+ with gPXE's native COMBOOT API. It will probably guarantee errors. Use vesamenu.c32 from 3.86. There were some drastic changes from 3.xx to 4.xx. By the way, you also can't load COM32 modules from versions prior to 4.00 with 4.00+ (generates a user-visible error). If you want/need to use vesamenu.c32 from 4.00+, you should load pxelinux.0 from the same package/package set first. I believe there's an example either on the gPXE wiki or PXELINUX wiki page of how to set options such that PXELINUX will grab a specific config file and set a specific cwd (current working directory; including protocol and host). This should be easy to put into your second script (eliminating the need to re-burn your ROMs). -- -Gene ___ gPXE mailing list gPXE@etherboot.org http://etherboot.org/mailman/listinfo/gpxe
Re: [gPXE] [PATCH] Make gpxe report syslinux version 2.12
On Fri, Oct 22, 2010 at 13:05, Jeremy Eder je...@redhat.com wrote: Hi Gene, On Fri, 2010-10-22 at 13:00 -0400, Casey Dahlin wrote: On Tue, Oct 12, 2010 at 07:37:09PM -0400, Gene Cumm wrote: On Tue, Oct 12, 2010 at 16:29, Casey Dahlin cdah...@redhat.com wrote: Some versions of mboot.c32 shipped by VMWare fail because of a version check (done with INT 22h with AX set to 0001h). gPXE appears to use an empty version number (0.0). This patch makes it return 2.12 Which VMware product(s)/version(s) (ESXi, perhaps?) with what version of mboot.c32? VMware ESXi 4.0.1...not sure what version of mboot.c32 actually. Probably the same as the VMware build of mboot.c32 from the ESXi Installable CD, I'd guess. Yes, ESXi, I'll have to relay the version question (John, if you're reading, feel free to chime in). Have you considered using gpxelinux.0 (at gpxe/gpxelinux.0) and any needed modules (like com32/modules/mboot.c32) from the current Syslinux package? gpxelinux.0 in Syslinux 4.02 is at gPXE 1.0.0. Hmm, does that version include a fix for this ? Not sure what this suggestion is meant to test. Yes and no. If you use gpxelinux.0, mboot.c32 will be talking with PXELINUX and not the emulated API in gPXE. However, if you are not using gPXE on top of the hardware's PXE client and using it instead of any hardware-derived client, this would need a little more work (roughly gPXE to load gpxelinux.0 to run mboot.c32). Either way, the installer is looking for any version. It doesn't have to be 2.12. Can that field be populated automatically by a macro ? Why version 2.12? Is that the exact minimum version your mboot.c32 wants? Yes, correct...exact minimum...as I mentioned, anything higher would work just fine as well I would think that it should match the API implemented in gPXE or be zero. I'm not sure if HPA is on this list or not and what his opinion/preference is. As of Syslinux-4.00, there was a big change in the API and COM32 format. -- -Gene ___ gPXE mailing list gPXE@etherboot.org http://etherboot.org/mailman/listinfo/gpxe
Re: [gPXE] Network Driver for NVIDIA
On Wed, Oct 13, 2010 at 17:15, Amir Marashi geekmicroc...@gmail.com wrote: Hi, Is there any NIC driver for the NVIDIA with the following information, I tried rom o matic but I could not find this specific driver. DEV ID: 0x0ab0 VEN ID: 0x10de Thank you in advance. http://etherboot.org/wiki/faq/drivers The IRC bot gpxebot, when asked gpxebot: lspci 10de:0ab0 said forcedeth -- -Gene ___ gPXE mailing list gPXE@etherboot.org http://etherboot.org/mailman/listinfo/gpxe
[gPXE] Invitation to connect on LinkedIn
LinkedIn Gene Cumm requested to add you as a connection on LinkedIn: -- Marty, I'd like to add you to my professional network on LinkedIn. - Gene Accept invitation from Gene Cumm http://www.linkedin.com/e/-1fbmf0-gf4emta1-u/XFq7-jooHGef7nsMdOpYLSrobnb3Q7/blk/I893194438_3/pmpxnSRJrSdvj4R5fnhv9ClRsDgZp6lQs6lzoQ5AomZIpn8_cRYUcPgQej4Pejx9bTdLrQBPskBfbP4Udz4RczgSdP4LrCBxbOYWrSlI/EML_comm_afe/ View invitation from Gene Cumm http://www.linkedin.com/e/-1fbmf0-gf4emta1-u/XFq7-jooHGef7nsMdOpYLSrobnb3Q7/blk/I893194438_3/0PnPwPd3gVcjcVe4ALqnpPbOYWrSlI/svi/ -- DID YOU KNOW you can be the first to know when a trusted member of your network changes jobs? With Network Updates on your LinkedIn home page, you'll be notified as members of your network change their current position. Be the first to know and reach out! http://www.linkedin.com/ -- (c) 2010, LinkedIn Corporation___ gPXE mailing list gPXE@etherboot.org http://etherboot.org/mailman/listinfo/gpxe
Re: [gPXE] Invitation to connect on LinkedIn
On Sun, Oct 10, 2010 at 19:53, Marty Connor m...@etherboot.org wrote: Hi Gene, Surely you didn't mean to send this request to the entire gPXE mailing list. Correct. I should have been more careful when adding connections to accounts recognized by my address book. I apologize for this mistake. Since this seems to have happened a number of times, can anyone suggest (reasonably graeceful) ways to keep this from happening? We use GNU Mailman to manage our mailing lists. Two ideas. 1) Marty, by the looks of it, you have a LinkedIn account and have gpxe@etherboot.org listed as the primary email. Changing this email would of course change this behavior. 2) I'm not sure if Mailman can do it but filtering based on the inbound IP address, DNS name or other distinguisher to direct traffic from LinkedIn to be moderated or moved directly to your personal account. Thanks, / Marty / -- -Gene No one ever says, 'I can't read that ASCII(plain text) e-mail you sent me.' On 10/10/10 5:06 PM, Gene Cumm via LinkedIn wrote: LinkedIn Gene Cumm requested to add you as a connection on LinkedIn: -- Marty, I'd like to add you to my professional network on LinkedIn. - Gene ___ gPXE mailing list gPXE@etherboot.org http://etherboot.org/mailman/listinfo/gpxe
Re: [gPXE] Oddball results
On Sun, Oct 3, 2010 at 18:01, Dustin Eward emptythemagaz...@gmail.com wrote: Dell Dimension E521? On the US support site, I see 1.1.11 (released on 2007-09-04) is the latest. I don't see anything in the notes for this machine but sometimes they update the PXE code without any notes. I am using the version, the PXE in it is 100% fail. left the Phoenix BIOS boot-up identifier. It shows for only a split second, but I could probably figure out the REAL motherboard MFR and part/model number if I rebooted a lot... I'll have to get back to you because I can't do that right this moment. I've seen a few systems in the last 2+ years that do this (showing signs of the underlying BIOS) but not many. Yeah, and DELL missing that one... They usually go out of their way to hide/cripple everything. It was one of their rare AMD machines from a few years ago, I'm guessing they just weren't putting much effort into it. experiment, but I got impatient. From what I've seen, I have the impression that Intel has generally been very good for both PXE and Linux. I got that impression, so I spent the money instead of continuing to hammer on this embedded BCM4401. The gPXE b44 .usb image off of ROM-o-Matic doesn't even recognize it. Neitehr do the UNDI/UNDIONLY. But, the 'all drivers' loads the UNDI and it pulls DHCP... Meh, bought the Intel cards. I have a Dell laptop with a BCM4401 LoM (14e4:170c 1028:0188). It's an Inspiron 6000 with an Intel Pentium M running BIOS A09 (2005-09-28; latest). This machine appears to be fine with a PXE boot (at least with PXELINUX). This is the versioning information I was looking for off of yours: Broadcom UNDI, PE-2.1 (build 082) v2.0.4 Copyright (C) 2000-2003 Broadcom Corporation Copyright (C) 1997-2000 Intel Corporation All rights reserved. Intel Base-Code, PXE-2.1 (build 082j) Copyright (C) 1997-2000 Intel Corporation -- -Gene ___ gPXE mailing list gPXE@etherboot.org http://etherboot.org/mailman/listinfo/gpxe
Re: [gPXE] Oddball results
On Sat, Oct 2, 2010 at 22:59, Dustin Eward emptythemagaz...@gmail.com wrote: I have the following device embedded in my motherboard. BCM4401-B0 14e4:170c The currently installed Broadcom Proprietary PXE attempts DHCP, but always times out. The gPXE, when using the 'all drivers' image, works flawlessly. If only I could flash 600K+ to EEPROM. I haven't had personal experience with this particular LoM (LAN on Motherboard) but all of the LoMs I've dealt with are given an option ROM by the BIOS. Upgrading the BIOS may update the Broadcom PXE code. I've also seen where connecting to a manage switch with STP enabled but not set to a quick mode like rSTP or portfast causes timeouts. For reference, could you provide all of the versioning information for the PXE code used by the LoM, the make/model and BIOS version of the machine? -- -Gene ___ gPXE mailing list gPXE@etherboot.org http://etherboot.org/mailman/listinfo/gpxe
Re: [gPXE] Oddball results
On Sun, Oct 3, 2010 at 17:22, Dustin Eward emptythemagaz...@gmail.com wrote: It is a Phoenix BIOS, which I see no modification method for in the wiki. It's a Dell e521 machine, which is deceptively not Dell-ified. They even Dell Dimension E521? On the US support site, I see 1.1.11 (released on 2007-09-04) is the latest. I don't see anything in the notes for this machine but sometimes they update the PXE code without any notes. left the Phoenix BIOS boot-up identifier. It shows for only a split second, but I could probably figure out the REAL motherboard MFR and part/model number if I rebooted a lot... I'll have to get back to you because I can't do that right this moment. I've seen a few systems in the last 2+ years that do this (showing signs of the underlying BIOS) but not many. I should add, by 'works flawlessly, I mean it identifies the NIC and pulls DHCP. The rest, well I've got a few PCI and a PCIe Ethernet Adapters on the way... 2 are Intel, one is a super cheap generic. I was trying to avoid spending $85 just to experiment, but I got impatient. From what I've seen, I have the impression that Intel has generally been very good for both PXE and Linux. On Sun, Oct 3, 2010 at 6:47 AM, Gene Cumm gene.c...@gmail.com wrote: On Sat, Oct 2, 2010 at 22:59, Dustin Eward emptythemagaz...@gmail.com wrote: I have the following device embedded in my motherboard. BCM4401-B0 14e4:170c The currently installed Broadcom Proprietary PXE attempts DHCP, but always times out. The gPXE, when using the 'all drivers' image, works flawlessly. If only I could flash 600K+ to EEPROM. I haven't had personal experience with this particular LoM (LAN on Motherboard) but all of the LoMs I've dealt with are given an option ROM by the BIOS. Upgrading the BIOS may update the Broadcom PXE code. I've also seen where connecting to a manage switch with STP enabled but not set to a quick mode like rSTP or portfast causes timeouts. For reference, could you provide all of the versioning information for the PXE code used by the LoM, the make/model and BIOS version of the machine? -- -Gene ___ gPXE mailing list gPXE@etherboot.org http://etherboot.org/mailman/listinfo/gpxe