Anaconda-devel June 2004

anaconda-devel@lists.stg.fedoraproject.org

25 participants
35 discussions

FC2 and dual-boot Windows
by Taylor, ForrestX 15 Jul '04

15 Jul '04

I am working on getting my FC2 build released, but I am worried about the problem with dual-booting Windows that I have seen. Is there any way to get the legacy partitioning info for the installer without any manual intervention? Are there any updates for the kernel/parted/anaconda that would remedy this problem? I've seen solutions that use the legacy EDD information to change the heads to 255/240, although I haven't been able to determine which disks to perform this on. I assume that int13_dev80 is only for device 0x80, thus the discussion about disks other than 0x80. Is Red Hat working on a fix/respin? Here is what I have so far in the %pre (untested): ------------- modprobe edd # It seems to be only the heads that are incorrect H=$(echo "`cat /sys/firmware/edd/int13_dev80/legacy_max_head` + 1" | bc -l) S=$(cat /sys/firmware/edd/int13_dev80/legacy_sectors_per_track) C=$(echo "scale=0;`cat /sys/firmware/edd/int13_dev80/sectors` / $H / $S" | bc -l) # I'm not sure how to determine which drives are present (maybe # /tmp/hda vs. /tmp/sda?? echo "bios_cyl:$C bios_head:$H bios_sect:$S" > /proc/ide/hda/settings ------------- Does anyone see any problems with this--or any better solutions? I still haven't figured out how to catch the exception where the system disks are SCSI and the data disks are IDE (/ on SCSI). Thanks, Forrest

2 6

anaconda-help
by Patrick Devine 02 Jul '04

02 Jul '04

Has anyone managed to get anaconda-help to build behind a firewall? I always get this lovely error message: xmlto --skip-validation -o C html -m onlinehelp.xsl help-screens-C.xml Attempt to load network entity http://docbook.sourceforge.net/release/xsl/current/html/chunk.xsl warning: failed to load external entity "http://docbook.sourceforge.net/release/xsl/current/html/chunk.xsl" compilation error: file /tmp/xmlto-xsl.GerB3E line 4 element import xsl:import : unable to load http://docbook.sourceforge.net/release/xsl/current/html/chunk.xsl cp: cannot stat `*.htm*': No such file or directory make[1]: *** [C-html] Error 1 make[1]: Leaving directory `/usr/src/redhat/BUILD/anaconda-help-9.1/help' make: *** [subdirs] Error 1 error: Bad exit status from /var/tmp/rpm-tmp.37092 (%build) It seems really stupid to make your build reliant on pulling style sheets from a network server. --Patrick.

3 5

RAID and LVM
by Paul Nasrat 30 Jun '04

30 Jun '04

I'm testing a fix I've done related to RAID device solidification for regressions with RAID/LVM interaction, hit a stacktrace. Removed fix (*hugs* RHupdates), retested and same issue occurs. I kickstart (config attached), step through to partitioning screen I double clicked on LVM Volume Groups > vg and get the below stack. Traceback (most recent call last): File "/usr/lib/anaconda/iw/partition_gui.py", line 938, in treeActivateCb self.editCb() File "/usr/lib/anaconda/iw/partition_gui.py", line 1045, in editCb self.editLVMVolumeGroup(request) File "/usr/lib/anaconda/iw/partition_gui.py", line 1128, in editLVMVolumeGroup origvgrequest, isNew) File "/usr/lib/anaconda/iw/lvm_dialog_gui.py", line 987, in __init__ (self.lvmlist, sw) = self.createAllowedLvmPartitionsList(self.availlvmparts, self.origvgrequest.physicalVolumes, self.partitions, origvgrequest.getPreExisting()) File "/usr/lib/anaconda/iw/lvm_dialog_gui.py", line 315, in createAllowedLvmPartitionsList partname = "md%d" % (request.raidminor,) TypeError: int argument required I believe the issue is related to the fact that if we are Auto/None for the raidminor here and we're not testing for that. Carrying on the install without trying to get the edit screen works happily. I can catch the TypeError (or explicitly test), but the later getRequestByDeviceName("Auto") is going to fail. The requests don't get minors until we do toEntry(). This is making me think do we need to look at how we currently track minors in partRequests. Paul

2 2

RE: EDD disksigs for disks other than 0x80
by Rezwanul_Kabir＠Dell.com 30 Jun '04

30 Jun '04

> -----Original Message----- > From: Jeremy Katz [mailto:katzj@redhat.com] > Sent: Tuesday, June 29, 2004 2:34 PM > To: Kabir, Rezwanul > Cc: Domsch, Matt; anaconda-devel-list(a)redhat.com > Subject: RE: EDD disksigs for disks other than 0x80 > > > On Fri, 2004-06-25 at 14:53 -0500, Rezwanul_Kabir(a)Dell.com wrote: > > One of the reasons ,I used environment variables is to > access them in > > consistent manner between loader and anaconda. Otherwise I > would have to > > touch more codes to pass them to anaconda in some way, e.g. > as command > > line arguments.. > > There shouldn't be a need to pass information around, though. Whether > or not edd is being used should be pretty obvious based on whether or > not they ask for something in terms of BIOS disks. I must be missing something here...I understood the part(described in your next email) that you want to populate a struct of some sort ( a hash table perhaps) with the mapping between BIOS and real device names. But as I understood the loader/anaconda code, you actually interprete some of the kickstart directives (e.g. harddrive --partition=<part#> --dir=/) in loader and some (e.g. part / --size=<size> --ondisk=<disk> ) in anacoda. If I'm correct, then you'll need to know the BIOS disk mappings both in loader and in anaconda. How do you propose to propagate this information from loader to anaconda..? --rez

2 1

RE: EDD disksigs for disks other than 0x80
by Rezwanul_Kabir＠Dell.com 30 Jun '04

30 Jun '04

> -----Original Message----- > From: Jeremy Katz [mailto:katzj@redhat.com] > Sent: Tuesday, June 29, 2004 3:11 PM > To: Kabir, Rezwanul > Cc: Domsch, Matt; anaconda-devel-list(a)redhat.com > Subject: RE: EDD disksigs for disks other than 0x80 > > > > diff -urNp old-anaconda-10/loader2/eddsupport.c > > anaconda-10.0/loader2/eddsupport.c > > --- old-anaconda-10/loader2/eddsupport.c 1969-12-31 > > 18:00:00.000000000 -0600 > > +++ anaconda-10.0/loader2/eddsupport.c 2004-06-29 > 11:24:47.632167320 > > -0600 > > @@ -0,0 +1,464 @@ > > +/* This is the top level function that creates the disk > list present in > > the > > + system, checks to see if unique signatures exist on the disks at > > offset 0x1b8. > > + If unique signature exists then it will find out the BIOS disk80 > > device by > > + reading in the MBR sig and compare it against the > signatures on the > > disks and > > + set the environment variable DISK80 to whatever the hd/sd > name it has. > > If > > + unique signatures don't exist on the disks, then it will > write unique > > + signatures on the disks and reboot */ > > Unfortunately, this is a bit of a non-starter for real inclusion. > Random reboots aren't at all guaranteed to get back into the > installation environment. There's actually a pretty good chance that > they won't. So, I think that instead, the proper course is if a bios > disk is specified that we can't determine, we just have to > pop up a nice > error dialog. > > anaconda can't get into the business of being responsible for your > machine being set up properly to begin with (although it's conceivable > to add some stuff to make it easier to add unique sigs if > you're in the > case where you don't have them; that shouldn't be the default behavior > though) I understand your concerns here. The above actually works better in a more controlled environment.Okay, I'll remove the part where it tries to write signatures and reboot... --rez

1 0

RE: EDD disksigs for disks other than 0x80
by Rezwanul_Kabir＠Dell.com 29 Jun '04

29 Jun '04

As per Jermey's recommendations, I have done the followings: 1) Using kudzu to create the disk list instead of parsing the /proc/partitions 2) As a consequence of the above, regexp is no longer needed, so gotten rid of it. I still need a consensus on a proper naming conventions for BIOS disks and partitions. How about identifying BIOS disks as "bd[a-z]" and partitions as "bd[a-z][1-9][0-9]*" So, for example kickstart directives would be: harddrive --partition=bda2 --dir=/ part / --size=1646 --ondisk=bdb etc. --rez Here's the latest patch.... diff -urNp old-anaconda-10/anaconda anaconda-10.0/anaconda --- old-anaconda-10/anaconda 2004-05-07 11:22:38.000000000 -0600 +++ anaconda-10.0/anaconda 2004-06-29 10:38:13.000000000 -0600 @@ -483,11 +483,23 @@ for n in args: elif (str == '--kickstart'): from kickstart import Kickstart ksfile = arg - instClass = Kickstart(ksfile, flags.serial) + + ## Code inserted by Rez ### + print 'Rez code here' + if os.environ.has_key("EDD"): + print 'The DISK80 var is : '+os.environ["DISK80"] + time.sleep(10) + import eddparser + eddparser.Eddfy(ksfile) + print 'Eddfy over' + print 'Rez code ended' + time.sleep(20) + + instClass = Kickstart(ksfile, flags.serial) elif (str == '--lang'): - lang = arg + lang = arg elif (str == '--lowres'): - runres = '640x480' + runres = '640x480' elif (str == '-m' or str == '--method'): method = arg if method[0] == '@': diff -urNp old-anaconda-10/.cvsignore anaconda-10.0/.cvsignore --- old-anaconda-10/.cvsignore 1969-12-31 18:00:00.000000000 -0600 +++ anaconda-10.0/.cvsignore 2003-01-14 00:19:38.000000000 -0600 @@ -0,0 +1,5 @@ +*.pyc +anaconda*.tar.gz +anaconda*.tar.bz2 +lang-names +mini-wm diff -urNp old-anaconda-10/eddparser.py anaconda-10.0/eddparser.py --- old-anaconda-10/eddparser.py 1969-12-31 18:00:00.000000000 -0600 +++ anaconda-10.0/eddparser.py 2004-06-29 10:27:48.000000000 -0600 @@ -0,0 +1,120 @@ +import time +import sys +import os +import re +import string +import getopt + +module_name="eddparser" +class EddClass: + def __init__(self,ksFileName): + self.kickStartFile=ksFileName + + def run(self): + print self.kickStartFile + fh_old=open(self.kickStartFile,"r") + newFileName=self.kickStartFile+".tmp" + fh_new=open(newFileName,"a+") + self.parseCommands(fh_old,fh_new) + os.rename(newFileName,self.kickStartFile) + fh_old.close() + + def foundMatch(self,ksLine,regCompiledList): + tokenList=[] + match=None + for regexp in regCompiledList: + match=regexp.match(ksLine) + if match != None: + tokenList=tokenList+string.split(ksLine) + break + return (match,tokenList) + + def handleDrive(self,(m,argString)): + print 'Handler for drive' + disk="" + part="" + options,extra=getopt.getopt(argString,'',['drive=','partition=','dir=']) + if extra: + raise ValueError, "Error in handleDrive: Unsupported format in hardrive command" + else: +# print 'Valid Options found' + + for (opt,val) in options: + if opt=='--drive': + disk=val + elif opt=='--partition': + part=val + else: + pass + + tempString=string.join(argString," ") + if len(disk)>0: + newOption="--partition="+os.environ[string.upper(disk)]+part + newString="harddrive "+string.replace(tempString,"--drive="+disk+" "+"--partition="+part,newOption)+"\n" + else: + newString="harddrive "+tempString + return newString + + + + + def handlePart(self,(m,argString)): + print 'Handle partition' + newOption=os.environ[string.upper(m.group(1))] + tempString=string.join(argString," ") + newString="part "+string.replace(tempString,m.group(1),os.environ[string.upper(m.group(1 ))])+"\n" + return newString + + + def parseCommands(self,oldFile,newFile): + print 'In parse' + cmdHandlers = { "harddrive":self.handleDrive , + "part":self.handlePart ,} + + commandList=[r"^harddrive[ \t]+--drive.*",r"^part[ \t].*(disk8.).*"] + + compiledList=[] + for item in commandList: + compiledList.append(re.compile(item)) + for line in oldFile.readlines(): + (matchObject,args)=self.foundMatch(line,compiledList) + if matchObject: + newLine=cmdHandlers[args[0]]((matchObject,args[1:])) + print newLine + newFile.write(newLine) + else: + newFile.write(line) + + + +def Eddfy(ksFile): + try: + + if os.environ.has_key("EDD") and (os.environ["EDD"] == "1"): + EddInstance=EddClass(ksFile) + EddInstance.run() + else: + print 'eddon option not set' + except ValueError,e: + print "Invalid Value in module :"+ module_name + print e.args + sys.exit(1) + except getopt.GetoptError,e: + print "Error with options in module:"+ module_name + print e.args + sys.exit(1) + except IOError,e: + print "IO error in module: " + module_name + print e.args + sys.exit(1) + except Exception,e: + print "Unknown Exception occured in module: "+ module_name + print e.args + sys.exit(1) + +#time.sleep(10) +#sys.exit(0) +#print 'Program started' +#Eddfy(sys.argv[1]) +#print 'The programe ended ' + diff -urNp old-anaconda-10/loader2/eddsupport.c anaconda-10.0/loader2/eddsupport.c --- old-anaconda-10/loader2/eddsupport.c 1969-12-31 18:00:00.000000000 -0600 +++ anaconda-10.0/loader2/eddsupport.c 2004-06-29 11:24:47.632167320 -0600 @@ -0,0 +1,464 @@ +#include "eddsupport.h" + +/*extern int errno; */ + + +char* strupper(char * orig) { + +char *current_ptr; + + if ( orig) { + for (current_ptr=orig;*current_ptr;(*current_ptr)=toupper(*current_ptr),curr ent_ptr++); + } + return(orig); +} + + +/* This is the top level function that creates the disk list present in the + system, checks to see if unique signatures exist on the disks at offset 0x1b8. + If unique signature exists then it will find out the BIOS disk80 device by + reading in the MBR sig and compare it against the signatures on the disks and + set the environment variable DISK80 to whatever the hd/sd name it has. If + unique signatures don't exist on the disks, then it will write unique + signatures on the disks and reboot */ + + +void prepareForEdd() { + +struct device ** disk_list=NULL; +__u32 mbr_sig; +status local_status; + + DPRINT("Going into createDiskList \n"); + disk_list=createDiskList(); + if(!disk_list){ + logMessage("Error: Could not create disk list, exiting"); + sleep(10); + exit(1); + } + + #ifdef LOCAL_DEBUG + printDiskList(disk_list); + #endif + + DPRINT("Trying to see if unique sig exists \n"); + + local_status=uniqueSignatureExists(disk_list); + switch(local_status){ + + case ERROR: + logMessage("Error: Could not determine if unique sig exists"); + sleep(10); + exit(1); + + case SUCCESS: + logMessage( "Unique sig exists \n"); + logMessage( "Reading MBR sig \n"); + local_status=readMbrSig(&mbr_sig); + if(local_status!=SUCCESS) { + logMessage("Problem reading mbr signature\n"); + sleep(10); + exit(1); + } + DPRINT("The hex mbr signature is : %x \n",mbr_sig); + + local_status=tagBiosLabel(disk_list,mbr_sig); + if (local_status!=SUCCESS){ + logMessage("Could not find BIOS boot device \n"); + sleep(10); + exit(1); + } + + logMessage( "Woo hooh , Found BIOS dev80 \n"); + sleep(10); + return; + + case FAILURE: + logMessage( "Unique sig does not exists \n"); + local_status=writeUniqueSignatures(disk_list); + if (local_status!=SUCCESS){ + logMessage("Problem writing disk signatures \n"); + sleep(10); + exit(1); + } + logMessage("Done writing unique signatures, rebooting\n"); + sleep(10); + reboot(RB_AUTOBOOT); + + } /* end switch */ + +} /* end prepareForEdd */ + + +struct device ** createDiskList(){ + +struct device ** disk_list_head=NULL; + + + disk_list_head=probeDevices(CLASS_HD,BUS_SCSI|BUS_IDE,PROBE_ALL); + + return(disk_list_head); + +} + + +status readMbrSig(__u32 * int_sig){ + +FILE* fh; + + fh=fopen(SIG_FILE,"r"); + if(fh==NULL){ + logMessage("Error opening mbr_signature file"); + return ERROR; + } + fseek(fh,0,SEEK_SET); + if(fscanf(fh,"%x",int_sig)!=1){ + logMessage("Error reading mbr_signature from /sys filesystem"); + fclose(fh); + return ERROR; + } + + fclose(fh); + return SUCCESS; + + } + + +status readDiskSig(char *devname,char* filename, __u32 *disk_signature) { + +int fd, bytes_read,offset; + + + if (devMakeInode(devname,filename)){ + logMessage("Could not create device node"); + sleep(20); + return ERROR; + } + fd=open(filename,O_RDONLY); + if ( fd < 0 ) { + logMessage("Error opening file : %s\n ",filename); + sleep(20); + unlink(filename); + return ERROR; + } + offset=lseek (fd,OFFSET,SEEK_SET); + if ( offset < 0 ){ + close(fd); + logMessage("Error with seek operation %s : ",filename); + sleep(20); + unlink(filename); + return ERROR; + } + + DPRINT( "reading disk sig\n"); + bytes_read=read(fd,disk_signature,sizeof(__u32)); + if ( bytes_read < sizeof(__u32)){ + close(fd); + if ( bytes_read==-1 ){ + logMessage("Error reading file : %s ",filename); + sleep(10); + } + else { + logMessage("Byte size doesn't match, read only %d bytes \n",bytes_read); + sleep(10); + } + + unlink(filename); + return ERROR; + } + + unlink(filename); + + return SUCCESS; + + +} + +status uniqueSignatureExists(struct device ** list_head) +{ + + __u32 current_device_signature,list_head_signature; + char list_head_device_name[9], current_device_name[9]; + struct device **current_list_head, **current_device_list; + status local_status; + + + memset(list_head_device_name,0,9); + memset(current_device_name,0,9); + + + for (current_list_head=list_head;(*current_list_head)->next!=NULL;current_li st_head++) { + + CREATE_DEVICE_NAME(list_head_device_name,(*current_list_head)->device); + DPRINT( "Current list head device name is : %s\n",list_head_device_name); + + local_status=readDiskSig((*current_list_head)->device,list_head_device_n ame,&list_head_signature); + + if(local_status==ERROR) { + logMessage("Error: Reading disk sig"); + sleep(10); + return ERROR; + } + + DPRINT( "The list head disk signature is : %x, IMPORTANT \n", list_head_signature); + + for (current_device_list=current_list_head+1;(*current_device_list)!=NULL;cu rrent_device_list++) { + + CREATE_DEVICE_NAME(current_device_name,(*current_device_list)->device); + DPRINT( "Current device name is : %s\n",current_device_name); + local_status=readDiskSig((*current_device_list)->device,current_device_n ame,&current_device_signature); + + if(local_status==ERROR) { + logMessage("Error: Reading disk sig"); + sleep(10); + return ERROR; + } + + + DPRINT( "The current disk signature is : %x IMPORTANT\n", current_device_signature); + if(list_head_signature==current_device_signature) + return FAILURE; + + } /* end inner for */ + + }/* end outer for */ + + return SUCCESS; + + + } + + +status writeDiskSig(char *devname,char *filename, __u32 *signature) { + +int fd, bytes_written,offset; + + + if(devMakeInode(devname,filename)) { + logMessage("Error creating device node"); + sleep(20); + return ERROR; + } + + fd=open(filename,O_WRONLY); + if ( fd < 0 ) { + logMessage("Error opening file : %s ",filename); + sleep(10); + unlink(filename); + return ERROR; + } + + offset=lseek (fd,OFFSET,SEEK_SET); + if ( offset < 0 ){ + close(fd); + logMessage("Error with seek operation : %s ",filename); + unlink(filename); + return ERROR; + } + + if((bytes_written=write(fd,signature,sizeof(__u32))) != 4){ + close(fd); + logMessage("Error writing to file: %s \n",filename); + unlink(filename); + return ERROR; + } + unlink(filename); + return SUCCESS; + +} + + + + +status writeUniqueSignatures(struct device ** list_head) { + +__u32 disk_sig, tmp_sig; +struct device **current_device_list; +char current_device_name[9]; +memset(current_device_name,0,9); +status local_status; + + for (current_device_list=list_head,disk_sig=START_SIG;(*current_device_list) !=NULL;current_device_list++,disk_sig++) { + DPRINT("Inside write loop"); + CREATE_DEVICE_NAME(current_device_name,(*current_device_list)->device); + DPRINT("Wriitng on device: %s ",current_device_name); + + local_status=writeDiskSig((*current_device_list)->device,current_device_ name,&disk_sig); + if (local_status != SUCCESS) + return local_status; + DPRINT("Written sig successfully"); + readDiskSig((*current_device_list)->device,current_device_name,&tmp_sig) ; + DPRINT("The sig is : %x",tmp_sig); + + + } + + return SUCCESS; +} + + + +status tagBiosLabel(struct device ** list_head,__u32 mbr_signature) { + + __u32 current_device_signature; + char current_device_name[9]; + struct device **current_device_list=NULL; + status local_status; + + memset(current_device_name,0,9); + + DPRINT("Inside tag bios label \n"); + + for (current_device_list=list_head;(*current_device_list)!=NULL;current_devi ce_list++) { + CREATE_DEVICE_NAME(current_device_name,(*current_device_list)->device); + DPRINT( "Current device name is : %s\n",current_device_name); + + local_status=readDiskSig((*current_device_list)->device,current_device_n ame,&current_device_signature); + + if(local_status==ERROR) { logMessage("Error: Reading disk sig"); + sleep(10); + return ERROR; + } + + + DPRINT( "The current disk signature is : %x \n", current_device_signature); + + if(mbr_signature==current_device_signature){ + + logMessage( "Found MBR signature in device : %s\n",current_device_name); + setenv("DISK80",(*current_device_list)->device,1); + DPRINT("The DISK80 env var is : %s \n",getenv("DISK80")); + return SUCCESS; + } + + } + + return FAILURE; + +} + +/* This function converts the "harddrive" command in ks.cfg that EDD supports + into the regular format that the redhat code supports */ + +char* parseHDCommand(char* drv, char* prt) { + +char * part; + + if (drv) { + + DPRINT("Parsing EDD options of kickstart harddrive command"); + logMessage ( "The EDD options are drive: %s, and partition: %s \n",drv,prt); + + if ( ! prt) { + logMessage( " EDD requires both drive and partition options \n"); + sleep(20); + exit(1); + } + + part=(char*)malloc((strlen(drv)+strlen(prt)+1)*sizeof(char)); + memset(part,0,strlen(drv)+strlen(prt)+1); + strncat(strncpy(part,getenv(strupper(drv)),strlen(drv)),prt,strlen(prt)) ; + DPRINT("The EDD partition is %s \n",part); + return part; + + } + return prt; + +} + +/* This routine parses the boot option, ks=..., for EDD option. If it finds +EDD options it will parse it and return a string that kickstartFromHD (in +hdinstall.c) understands..Otherwise it will perform an Identity operation on +the input string */ + + +char* parseEddHDCommand(char* tempstr) { + +char *p, *q = NULL, *r=NULL, *tmpksdev=NULL; +char *eddtmpstr=NULL; + + + if(strstr(tempstr,"disk8")) { + logMessage("EDD option found as boot option\n"); + p=strchr(tempstr,':'); + r=strchr(p+1,':'); + q=strchr(r+1,':'); + + if(!p || !q || !r) { + logMessage("EDD Format of command line is ks=hd:device:/path/to/ks.cfg"); + free(tempstr); + exit(1); + } + + *q='\0'; + *r='\0'; + + tmpksdev=strdup(getenv(strupper(p+1))); + DPRINT("kickstart device is %s\n",tmpksdev); + *(p+1)='\0'; + + + eddtmpstr=(char*)calloc((strlen(tempstr)+strlen(tmpksdev)+strlen(r+1)+st rlen(q+1)+2),sizeof(char)); + strcpy(eddtmpstr,tempstr); + DPRINT("The new tmp str is :%s \n",eddtmpstr); + strncat(eddtmpstr,tmpksdev,strlen(tmpksdev)); + DPRINT("partially built kickstart string is %s\n",eddtmpstr); + strncat(eddtmpstr,r+1,strlen(r+1)); + DPRINT("partially built kickstart string is %s\n",eddtmpstr); + strncat(eddtmpstr,":",1); + DPRINT("partially built kickstart string is %s\n",eddtmpstr); + strncat(eddtmpstr,q+1,strlen(q+1)); + free(tempstr); + free(tmpksdev); + free(r+1); + free(q+1); + p=q=r=NULL; + DPRINT("Completely built kickstart string is %s\n",eddtmpstr); + DPRINT("Final EDD tmpstr is %s",eddtmpstr); + return eddtmpstr; + + } + + return tempstr; +} + + +void printDiskList(struct device ** list_head) { + +struct device **current_disk; + +logMessage("The disk list with kudzu is as follows:\n"); +sleep(10); +for (current_disk=list_head;(*current_disk);current_disk++) { + + logMessage("%s\n",(*current_disk)->device); + +} + +} + +void printFile(char *filename) { + + + struct stat stat_buff; + char *buff; + int fd; + + fd=open(filename,O_RDONLY); + fstat(fd,&stat_buff); + + if(!stat_buff.st_size){ + DPRINT("The file size is zero\n"); + sleep(20); + } + else { + buff=malloc(stat_buff.st_size + 1); + read(fd,buff,stat_buff.st_size); + buff[stat_buff.st_size]='\0'; + DPRINT("The file is : \n %s \n",buff); + sleep(20); + + } + + +} diff -urNp old-anaconda-10/loader2/eddsupport.h anaconda-10.0/loader2/eddsupport.h --- old-anaconda-10/loader2/eddsupport.h 1969-12-31 18:00:00.000000000 -0600 +++ anaconda-10.0/loader2/eddsupport.h 2004-06-29 09:53:18.000000000 -0600 @@ -0,0 +1,53 @@ +#ifndef EDDSUPPORT_H +#define EDDSUPPORT_H +#include <stdio.h> +#include <string.h> +#include <unistd.h> +#include <fcntl.h> +#include <ctype.h> +#include <linux/types.h> +#include <sys/stat.h> +#include <sys/reboot.h> +#include <stdlib.h> +#include <kudzu/kudzu.h> +#include "log.h" +#include "../isys/isys.h" + + +#define LOCAL_DEBUG + +#define BUFF_SIZE 4 +#define OFFSET 0x1b8 +#define SIG_FILE "/sys/firmware/edd/int13_dev80/mbr_signature" +#define START_SIG 0x80 +#define CREATE_DEVICE_NAME(fullname,suffix) strcat(strcpy(fullname,"/tmp/"),\ +suffix) + +#ifdef LOCAL_DEBUG +#define DPRINT(s...) do { \ + logMessage(s);\ + sleep(5);\ + } while(0) +#else +#define DPRINT(s...) +#endif + + + +typedef enum {SUCCESS,FAILURE,ERROR} status; +struct device ** createDiskList(); +status readDiskSig(char*,char*, __u32*); +status writeDiskSig(char*,char*,__u32*); +status uniqueSignatureExists(struct device ** ); +status writeUniqueSignatures(struct device ** ); +status readMbrSig(__u32 *); +status tagBiosLabel(struct device **,__u32 ); +void prepareForEdd(); +char *strupper( char *orig); +char* parseHDCommand(char* drv, char* prt); +char* parseEddHDCommand(char* tempstr); +void printDiskList(struct device ** ); +void printFile(char *); +#endif + + diff -urNp old-anaconda-10/loader2/hdinstall.c anaconda-10.0/loader2/hdinstall.c --- old-anaconda-10/loader2/hdinstall.c 2003-12-16 10:51:01.000000000 -0600 +++ anaconda-10.0/loader2/hdinstall.c 2004-06-29 10:12:31.000000000 -0600 @@ -36,7 +36,7 @@ #include "modules.h" #include "method.h" #include "mediacheck.h" - +#include "eddsupport.h" #include "../isys/imount.h" #include "../isys/isys.h" @@ -583,14 +583,19 @@ char * mountHardDrive(struct installMeth void setKickstartHD(struct loaderData_s * loaderData, int argc, char ** argv, int * flagsPtr) { - char *partition, *dir; + char *drive=NULL, *partition, *dir; poptContext optCon; int rc; + + /*Rez code here */ struct poptOption ksHDOptions[] = { + {"drive", '\0', POPT_ARG_STRING, &drive, 0 }, { "partition", '\0', POPT_ARG_STRING, &partition, 0 }, { "dir", '\0', POPT_ARG_STRING, &dir, 0 }, { 0, 0, 0, 0, 0 } }; + + /*Rez code ends here */ logMessage("kickstartFromHD"); optCon = poptGetContext(NULL, argc, (const char **) argv, ksHDOptions, 0); @@ -604,6 +609,20 @@ void setKickstartHD(struct loaderData_s return; } +/*Rez code starts */ +/* If EDD option is set,parse the "hardrive" command in ks.cfg to find out the +actual partition to mount.The function parseHDCommand will take drive and +partition option from the "harddrive" command in ks.cfg and return the partitionas required by the rest of the code.If the "drive" option is not available it will just become and Identity operation. */ + + + if (getenv("EDD")) + partition=parseHDCommand(drive,partition); + + DPRINT("EDD code to find the harddrive and partitions to start hard drive based installation is done"); + +/* Rez code ends */ + + loaderData->method = strdup("hd"); loaderData->methodData = calloc(sizeof(struct hdInstallData *), 1); if (partition) @@ -623,6 +642,18 @@ int kickstartFromHD(char *kssrc, int fla /* format is ks=hd:[device]:/path/to/ks.cfg */ /* split of pieces */ tmpstr = strdup(kssrc); + + /*Rez code starts here */ +/* If EDD support is set , then parse the boot option ks=hd:drive:partition:/path/to/ks.cfg if available and return the format as recognized by the rest of +the code. If the new option is not available then the function just performs +Identity operation on the string tmpstr */ + + if (getenv("EDD")) + tmpstr=parseEddHDCommand(tmpstr); + +/*Rez code ended here */ + + p = strchr(tmpstr, ':'); if (p) q = strchr(p+1, ':'); @@ -642,6 +673,12 @@ int kickstartFromHD(char *kssrc, int fla ksdev = p+1; kspath = q+1; +/*Rez code here */ + DPRINT("The ksdev is : %s ",ksdev); + DPRINT("The kspath is : %s ",kspath); +/*Rez code ended */ + + logMessage("Loading ks from device %s on path %s", ksdev, kspath); if ((rc=getKickstartFromBlockDevice(ksdev, kspath))) { if (rc == 3) { diff -urNp old-anaconda-10/loader2/loader.c anaconda-10.0/loader2/loader.c --- old-anaconda-10/loader2/loader.c 2004-05-05 13:00:21.000000000 -0600 +++ anaconda-10.0/loader2/loader.c 2004-06-29 10:08:46.000000000 -0600 @@ -74,6 +74,7 @@ #include "net.h" #include "telnetd.h" +#include "eddsupport.h" #include "selinux.h" #include "../isys/imount.h" @@ -554,6 +555,17 @@ static int parseCmdLineFlags(int flags, flags &= ~LOADER_FLAGS_SELINUX; else if (!strncasecmp(argv[i], "selinux", 7)) flags |= LOADER_FLAGS_SELINUX; + + /* Code inserted here by Rez */ + + else if(!strncasecmp(argv[i],"noedd",5)) { + printf("Unsetting EDD env variable \n"); + unsetenv("EDD"); + printf("unsetting done"); + sleep(10); + } + + else if (numExtraArgs < (MAX_EXTRA_ARGS - 1)) { /* go through and append args we just want to pass on to */ /* the anaconda script, but don't want to represent as a */ @@ -1068,6 +1080,8 @@ int main(int argc, char ** argv) { struct loaderData_s loaderData; + char *env_val; + char * cmdLine = NULL; char * ksFile = NULL; int testing = 0; @@ -1144,6 +1158,17 @@ int main(int argc, char ** argv) { memset(&loaderData, 0, sizeof(loaderData)); + /* Assume that EDD support is turned on */ + + setenv("EDD","1",1); + env_val=getenv("EDD"); + printf("The EDD val is: %s \n",env_val); + sleep(10); + + + + + extraArgs[0] = NULL; flags = parseCmdLineFlags(flags, &loaderData, cmdLine); @@ -1229,6 +1254,23 @@ int main(int argc, char ** argv) { getDDFromSource(&loaderData, loaderData.ddsrc, flags); } + +/* Rez code here for edd. If EDD support is on, load the edd module and find the BIOS disk80 device*/ + + + if (getenv("EDD")) { + logMessage("new routine"); + logMessage("Found edd option"); + logMessage("Loading edd module"); + sleep(10); + mlLoadModuleSet("edd", modLoaded, modDeps,modInfo, flags); + logMessage("Done loading edd"); + sleep(5); + prepareForEdd(); + } + + + /* JKFIXME: loaderData->ksFile is set to the arg from the command line, * and then getKickstartFile() changes it and sets FL_KICKSTART. * kind of weird. */ diff -urNp old-anaconda-10/loader2/Makefile anaconda-10.0/loader2/Makefile --- old-anaconda-10/loader2/Makefile 2004-02-23 14:53:41.000000000 -0600 +++ anaconda-10.0/loader2/Makefile 2004-06-28 15:11:32.000000000 -0600 @@ -16,7 +16,7 @@ GUNZIP = -lz BINS = loader HWOBJS = pcmcia.o usb.o firewire.o hardware.o -METHOBJS = method.o cdinstall.o hdinstall.o nfsinstall.o urlinstall.o +METHOBJS = method.o cdinstall.o hdinstall.o nfsinstall.o urlinstall.o eddsupport.o OBJS = log.o moduleinfo.o loadermisc.o modules.o moduledeps.o windows.o \ lang.o kbd.o modstubs.o driverdisk.o selinux.o \ md5.o mediacheck.o kickstart.o driverselect.o \ > -----Original Message----- > From: Kabir, Rezwanul > Sent: Friday, June 25, 2004 2:54 PM > To: 'Jeremy Katz'; Domsch, Matt > Cc: anaconda-devel-list(a)redhat.com > Subject: RE: EDD disksigs for disks other than 0x80 > > > Sorry to chime in a little late. I wasn't following the issue > tracker.. > Anyway, I wanted to comment on Jeremy's comments about the > style of the coding. > > In order to facilitate other's understanding, I'm copying and > pasting jeremy's comments.. > > >Hrmmm, I'm not sure that I like the approach being taken > here. Before getting into > > details, a couple of process related things: > >* Having the whole thing as a unified diff makes things far > simpler to review. The simple >way to do this is to copy > files before you modify them to file.foo and then run > `gendiff >anaconda .foo`. For new files, touch newfile.foo > and then it'll get included inthe diff as >well. > >* Instead of copying and pasting functions and commenting > old versions out, please just put >your changes inline. It > makes the diff much more difficult to parse otherwise. > > Your points well taken. I'll stick to them from now on.. > > >1) Stylistically, this doesn't fit in properly with > anaconda. We don't pass around state > >via environmental variables, instead, the flags structure is > used with all of the defines > >in loader.h. Then, parsing of /proc/cmdline can occur in one place > > (loader.c:parseCmdLineFlags()) > > > One of the reasons ,I used environment variables is to access > them in consistent manner between loader and anaconda. > Otherwise I would have to touch more codes to pass them to > anaconda in some way, e.g. as command line arguments.. > > >2) Doing replacements of things inside the kickstart config > is going to be fragile. I'd > >really prefer to avoid that. Two possible approaches for > doing so would be > > a) Decide to use something like biosdisk0 -> biosdisk7 to > correspond to bios disks 0x80 > >-0x87 and just then use that for corresponding to the disk > and then use biosdisk0p1 for the >first partition. (biosdisk > or bd or something that doesn't actually exist, not > completely >sold on which syntax) > > b) Instead of using the existing directives, add a new > one of bd: instead of places > >where hd: gets used for defining a hard drive. Then, > --onbiosdisk or similar with > >partitioning commands. This is probably the better > approach, but I haven't thought through >it as completely > > My idea here was to decouple the "EDD" stuff as much as > possible from the rest of the code.This would allow > development/modification/experimentation with "EDD" code > independently from the rest of the bulk. With this in mind, > to parse boot options or ks.cfg, I introduced some hooks into > the loader/anaconda code (parseEDDHDCommand,EDDfy etc.)that > would behave as wrapper/adapter. They allow "current" > options to pass through without any scrutiny and will convert > any new "EDD options" into their corresponding "current" > format.The back end, which makes sense of this options > remains the same... > > > >Why have the noedd flag -- edd will only get utilized if > people request it, especially if > >we're not rewriting the ks.cfg. Does loading the module > cause an oops in some cases? > > I used it basically to do two things 1) load the edd.o > module 2) activate hooks in loader/anaconda that deals with > EDD specific options.. > > >5) regexp are likely to increase the size of the loader > greatly -- are they really needed? > > Otherwise, I probably needed to write a tiny state > machine. I went for the easier way :-) > > > >6) Parsing of /proc/partitions isn't recommended (your > current parsing will miss cciss and >other Weird (tm) devices > for example). Instead, getting a list of drives from kudzu is > > better as that's then completely consistent > > Point well taken.. I'll see if I can use kudzu instead... > > > >7) With the way you're currently doing things, you're not > guaranteed to have the device > > node when you need it -- we don't create the full set of > devices in /dev until we get to > > the second stage and the loader always makes device nodes > in /tmp as needed using > > devMakeInode() from isys > > Aahah.. if you look into the readDiskSig/writeDiskSig in > eddsupport.c you'll see that I'm indeed using the > devMakeInode interface.. > I did some preliminary testing before sending you the > code..So the code works at least with some simple configs... > > Thanks.. > --rez > > > > > > -----Original Message----- > > From: anaconda-devel-list-bounces(a)redhat.com > > [mailto:anaconda-devel-list-bounces@redhat.com]On Behalf Of > > Jeremy Katz > > Sent: Friday, June 25, 2004 9:54 AM > > To: Domsch, Matt > > Cc: anaconda-devel-list(a)redhat.com > > Subject: Re: EDD disksigs for disks other than 0x80 > > > > > > On Thu, 2004-06-24 at 15:40 -0500, Matt Domsch wrote: > > > Patch below against 2.6.7-bk(current) gives mbr_signature > for first > > > six BIOS disks. > > > > Seems to work, at least for two (all that I have handy on a > > test machine > > right now). Unfortunately, the mbr_signature on both of those disks > > (and the sig on the first from before rebooting into the > newer kernel) > > is 0x00000000. So I guess these don't have any guarantee of > > uniqueness? > > > > Jeremy > > > > > > _______________________________________________ > > Anaconda-devel-list mailing list > > Anaconda-devel-list(a)redhat.com > > https://www.redhat.com/mailman/listinfo/anaconda-devel-list > > >

2 1

RE: EDD disksigs for disks other than 0x80
by Rezwanul_Kabir＠Dell.com 29 Jun '04

29 Jun '04

Sorry to chime in a little late. I wasn't following the issue tracker.. Anyway, I wanted to comment on Jeremy's comments about the style of the coding. In order to facilitate other's understanding, I'm copying and pasting jeremy's comments.. >Hrmmm, I'm not sure that I like the approach being taken here. Before getting into > details, a couple of process related things: >* Having the whole thing as a unified diff makes things far simpler to review. The simple >way to do this is to copy files before you modify them to file.foo and then run `gendiff >anaconda .foo`. For new files, touch newfile.foo and then it'll get included inthe diff as >well. >* Instead of copying and pasting functions and commenting old versions out, please just put >your changes inline. It makes the diff much more difficult to parse otherwise. Your points well taken. I'll stick to them from now on.. >1) Stylistically, this doesn't fit in properly with anaconda. We don't pass around state >via environmental variables, instead, the flags structure is used with all of the defines >in loader.h. Then, parsing of /proc/cmdline can occur in one place > (loader.c:parseCmdLineFlags()) One of the reasons ,I used environment variables is to access them in consistent manner between loader and anaconda. Otherwise I would have to touch more codes to pass them to anaconda in some way, e.g. as command line arguments.. >2) Doing replacements of things inside the kickstart config is going to be fragile. I'd >really prefer to avoid that. Two possible approaches for doing so would be > a) Decide to use something like biosdisk0 -> biosdisk7 to correspond to bios disks 0x80 >-0x87 and just then use that for corresponding to the disk and then use biosdisk0p1 for the >first partition. (biosdisk or bd or something that doesn't actually exist, not completely >sold on which syntax) > b) Instead of using the existing directives, add a new one of bd: instead of places >where hd: gets used for defining a hard drive. Then, --onbiosdisk or similar with >partitioning commands. This is probably the better approach, but I haven't thought through >it as completely My idea here was to decouple the "EDD" stuff as much as possible from the rest of the code.This would allow development/modification/experimentation with "EDD" code independently from the rest of the bulk. With this in mind, to parse boot options or ks.cfg, I introduced some hooks into the loader/anaconda code (parseEDDHDCommand,EDDfy etc.)that would behave as wrapper/adapter. They allow "current" options to pass through without any scrutiny and will convert any new "EDD options" into their corresponding "current" format.The back end, which makes sense of this options remains the same... >Why have the noedd flag -- edd will only get utilized if people request it, especially if >we're not rewriting the ks.cfg. Does loading the module cause an oops in some cases? I used it basically to do two things 1) load the edd.o module 2) activate hooks in loader/anaconda that deals with EDD specific options.. >5) regexp are likely to increase the size of the loader greatly -- are they really needed? Otherwise, I probably needed to write a tiny state machine. I went for the easier way :-) >6) Parsing of /proc/partitions isn't recommended (your current parsing will miss cciss and >other Weird (tm) devices for example). Instead, getting a list of drives from kudzu is > better as that's then completely consistent Point well taken.. I'll see if I can use kudzu instead... >7) With the way you're currently doing things, you're not guaranteed to have the device > node when you need it -- we don't create the full set of devices in /dev until we get to > the second stage and the loader always makes device nodes in /tmp as needed using > devMakeInode() from isys Aahah.. if you look into the readDiskSig/writeDiskSig in eddsupport.c you'll see that I'm indeed using the devMakeInode interface.. I did some preliminary testing before sending you the code..So the code works at least with some simple configs... Thanks.. --rez > -----Original Message----- > From: anaconda-devel-list-bounces(a)redhat.com > [mailto:anaconda-devel-list-bounces@redhat.com]On Behalf Of > Jeremy Katz > Sent: Friday, June 25, 2004 9:54 AM > To: Domsch, Matt > Cc: anaconda-devel-list(a)redhat.com > Subject: Re: EDD disksigs for disks other than 0x80 > > > On Thu, 2004-06-24 at 15:40 -0500, Matt Domsch wrote: > > Patch below against 2.6.7-bk(current) gives mbr_signature for first > > six BIOS disks. > > Seems to work, at least for two (all that I have handy on a > test machine > right now). Unfortunately, the mbr_signature on both of those disks > (and the sig on the first from before rebooting into the newer kernel) > is 0x00000000. So I guess these don't have any guarantee of > uniqueness? > > Jeremy > > > _______________________________________________ > Anaconda-devel-list mailing list > Anaconda-devel-list(a)redhat.com > https://www.redhat.com/mailman/listinfo/anaconda-devel-list >

2 1

Problem hitting enter at final reboot screen
by Eric V. Smith 29 Jun '04

29 Jun '04

This is with FC2. I'm having a problem with pressing "Enter" at the final dialog (labeled "Complete") that asks you to reboot. This is after doing a PXE kickstart installation. The entire install is automated, this is the first and only screen that asks for any user input. I realize I can make this screen go away with a ks.cfg entry, but I need it because I need to catch the box before it reboots again. Anyway, the problem is that there is a blinking cursor just to the right of the reboot "button". If I press Enter, nothing happens, except the blinking cursor moves down the screen. Only when I type space and then press Enter does the button actually get "pressed", and the box reboots. I can type other chars and then Enter and nothing happens, the space is definitely significant. I assume something I can't see is asking for input. Any characters I type before I hit Enter are echoed to the screen. I run a "ps" and I don't see anything out of the ordinary. I can switch to other virtual terminals and see them while the Complete dialog is being displayed. The F2 console interaction works correctly both before and after the Complete dialog is displayed. One possibility is that anaconda or kickstart is objecting to some rpm that I haven't installed. I edit comps.xml pretty heavily to remove unused packages. I did a similar thing with RH8 (different comps.xml, of course), and it worked fine. Everything seems to work okay, it's just annoying and I'm worried I might have broken something I haven't noticed yet. Any ideas how I can tell what is going on, or what program is actually asking for input? Thanks in advance. Eric.

2 1

Couldn't find packages while running getfullcomps.py
by Manilal K M 29 Jun '04

29 Jun '04

Hello Friends, I am trying to customize Fedora core 1 for i386 Architecture. When I run /usr/share/comps-extras/getfullcomps.py comps.xml /home/fedora i386 > /dev/null I got the error as follows: CRITICAL ERROR: Unable to find package prctl CRITICAL ERROR: Unable to find package efibootmgr CRITICAL ERROR: Unable to find package elilo CRITICAL ERROR: Unable to find package ppc64-utils CRITICAL ERROR: Unable to find package s390utils CRITICAL ERROR: Unable to find package yaboot I searched for these packages in the 3 CDs but I can't find it. My doubt can I ignore this error or should I download these and include it. I am also unaware about what the packages are meant for? Please help me regards Manilal -- ______________________________________________ Check out the latest SMS services @ http://www.linuxmail.org This allows you to send and receive SMS through your mailbox. Powered by Outblaze

3 2

Patch review.
by Paul Nasrat 29 Jun '04

29 Jun '04

This fixes https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=125006 Tested locally using kickstart, shouldn't have any undesirable impact. Paul

1 0

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

Anaconda-devel June 2004