/* * Copyright 2000, International Business Machines Corporation and others. * All Rights Reserved. * * This software has been released under the terms of the IBM Public * License. For details, see the LICENSE file in the top-level source * directory or online at http://www.openafs.org/dl/license10.html * * Portions Copyright (c) 2003 Apple Computer, Inc. * Portions Copyright (c) 2006 Sine Nomine Associates */ /* System: VICE-TWO Module: partition.c Institution: The Information Technology Center, Carnegie-Mellon University */ #include #include #include #include #ifdef AFS_NT40_ENV #include #include #include #else #if AFS_HAVE_STATVFS || AFS_HAVE_STATVFS64 #include #endif /* AFS_HAVE_STATVFS */ #if defined(AFS_DARWIN_ENV) || defined(AFS_XBSD_ENV) #include #endif #if !defined(AFS_SGI_ENV) #ifdef AFS_OSF_ENV #include #include #else /* AFS_OSF_ENV */ #ifdef AFS_VFSINCL_ENV #define VFS #ifdef AFS_SUN5_ENV #include #else #if defined(AFS_DARWIN_ENV) || defined(AFS_XBSD_ENV) #include #include #else #include #endif #endif #else /* AFS_VFSINCL_ENV */ #if !defined(AFS_AIX_ENV) && !defined(AFS_LINUX22_ENV) && !defined(AFS_DARWIN_ENV) && !defined(AFS_XBSD_ENV) #include #endif #endif /* AFS_VFSINCL_ENV */ #endif /* AFS_OSF_ENV */ #include #ifdef AFS_AIX_ENV #include #include #else #ifdef AFS_HPUX_ENV #include #include #else #if defined(AFS_SUN_ENV) #include #ifndef AFS_SUN5_ENV #include #endif #endif #ifdef AFS_SUN5_ENV #include #include #else #ifdef AFS_LINUX22_ENV #include #include #else #include #endif #endif #endif #endif #endif /* AFS_SGI_ENV */ #endif /* AFS_NT40_ENV */ #if defined(AFS_SGI_ENV) #include #include #endif #include #ifdef AFS_PTHREAD_ENV # include #endif #include #include #include "nfs.h" #include #include "lock.h" #include "lwp.h" #include #include "ihandle.h" #include "common.h" #ifdef AFS_NAMEI_ENV #ifdef AFS_NT40_ENV #include "ntops.h" #else #include "namei_ops.h" #endif /* AFS_NT40_ENV */ #endif /* AFS_NAMEI_ENV */ #include "vnode.h" #include "volume.h" #include "partition.h" #if defined(AFS_HPUX_ENV) #include #endif /* defined(AFS_HPUX_ENV) */ #ifdef AFS_AIX42_ENV #include #endif #ifdef AFS_NT40_ENV extern int VValidVPTEntry(struct vptab *vptp); #endif int aixlow_water = 8; /* default 8% */ struct DiskPartition64 *DiskPartitionList; #ifdef AFS_DEMAND_ATTACH_FS /* file to lock to conceptually "lock" the vol headers on a partition */ #define AFS_PARTLOCK_FILE ".volheaders.lock" #define AFS_VOLUMELOCK_FILE ".volume.lock" static struct DiskPartition64 *DiskPartitionTable[VOLMAXPARTS+1]; static struct DiskPartition64 * VLookupPartition_r(char * path); static void AddPartitionToTable_r(struct DiskPartition64 *); #endif /* AFS_DEMAND_ATTACH_FS */ #ifdef AFS_SGI_XFS_IOPS_ENV /* Verify that the on disk XFS inodes on the partition are large enough to * hold the AFS attribute. Returns -1 if the attribute can't be set or is * too small to fit in the inode. Returns 0 if the attribute does fit in * the XFS inode. */ #include static int VerifyXFSInodeSize(char *part, char *fstype) { afs_xfs_attr_t junk; int length = SIZEOF_XFS_ATTR_T; int fd = 0; int code = -1; struct fsxattr fsx; if (strcmp("xfs", fstype)) return 0; if (attr_set(part, AFS_XFS_ATTR, &junk, length, ATTR_ROOT) == 0) { if (((fd = open(part, O_RDONLY, 0)) != -1) && (fcntl(fd, F_FSGETXATTRA, &fsx) == 0)) { if (fsx.fsx_nextents) { Log("Partition %s: XFS inodes too small, exiting.\n", part); Log("Run xfs_size_check utility and remake partitions.\n"); } else code = 0; } if (fd > 0) close(fd); (void)attr_remove(part, AFS_XFS_ATTR, ATTR_ROOT); } return code; } #endif /* AFS_SGI_XFS_IOPS_ENV */ int VInitPartitionPackage(void) { #ifdef AFS_DEMAND_ATTACH_FS memset(&DiskPartitionTable, 0, sizeof(DiskPartitionTable)); #endif /* AFS_DEMAND_ATTACH_FS */ return 0; } static void VInitPartition_r(char *path, char *devname, Device dev) { struct DiskPartition64 *dp, *op; dp = malloc(sizeof(struct DiskPartition64)); /* Add it to the end, to preserve order when we print statistics */ for (op = DiskPartitionList; op; op = op->next) { if (!op->next) break; } if (op) op->next = dp; else DiskPartitionList = dp; dp->next = 0; dp->name = strdup(path); dp->index = volutil_GetPartitionID(path); #if defined(AFS_NAMEI_ENV) && !defined(AFS_NT40_ENV) /* Create a lockfile for the partition, of the form /vicepa/Lock/vicepa */ dp->devName = malloc(2 * strlen(path) + 6); strcpy(dp->devName, path); strcat(dp->devName, OS_DIRSEP); strcat(dp->devName, "Lock"); mkdir(dp->devName, 0700); strcat(dp->devName, path); close(afs_open(dp->devName, O_RDWR | O_CREAT, 0600)); dp->device = dp->index; #else dp->devName = strdup(devname); dp->device = dev; #endif dp->lock_fd = INVALID_FD; dp->flags = 0; dp->f_files = 1; /* just a default value */ #if defined(AFS_NAMEI_ENV) && !defined(AFS_NT40_ENV) if (programType == fileServer) (void)namei_ViceREADME(VPartitionPath(dp)); #endif VSetPartitionDiskUsage_r(dp); #ifdef AFS_DEMAND_ATTACH_FS AddPartitionToTable_r(dp); queue_Init(&dp->vol_list.head); CV_INIT(&dp->vol_list.cv, "vol list", CV_DEFAULT, 0); dp->vol_list.len = 0; dp->vol_list.busy = 0; { char lockpath[MAXPATHLEN+1]; snprintf(lockpath, MAXPATHLEN, "%s/" AFS_PARTLOCK_FILE, dp->name); lockpath[MAXPATHLEN] = '\0'; VLockFileInit(&dp->headerLockFile, lockpath); snprintf(lockpath, MAXPATHLEN, "%s/" AFS_VOLUMELOCK_FILE, dp->name); lockpath[MAXPATHLEN] = '\0'; VLockFileInit(&dp->volLockFile, lockpath); } VDiskLockInit(&dp->headerLock, &dp->headerLockFile, 1); #endif /* AFS_DEMAND_ATTACH_FS */ } static void VInitPartition(char *path, char *devname, Device dev) { VOL_LOCK; VInitPartition_r(path, devname, dev); VOL_UNLOCK; } #ifndef AFS_NT40_ENV /* VAttachPartitions() finds the vice partitions on this server. Calls * VCheckPartition() to do some basic checks on the partition. If the partition * is a valid vice partition, VCheckPartition will add it to the DiskPartition * list. * Returns the number of errors returned by VCheckPartition. An error in * VCheckPartition means that partition is a valid vice partition but the * fileserver should not start because of the error found on that partition. * * AFS_NAMEI_ENV * No specific user space file system checks, since we don't know what * is being used for vice partitions. * * Use partition name as devname. */ static int VCheckPartition(char *part, char *devname, int logging) { struct afs_stat_st status; #if !defined(AFS_LINUX20_ENV) && !defined(AFS_NT40_ENV) char AFSIDatPath[MAXPATHLEN]; #endif /* Only keep track of "/vicepx" partitions since it can get hairy * when NFS mounts are involved.. */ if (strncmp(part, VICE_PARTITION_PREFIX, VICE_PREFIX_SIZE)) { return 0; } if (afs_stat(part, &status) < 0) { Log("VInitVnodes: Couldn't find file system %s; ignored\n", part); return 0; } if (logging) { Log("This program is compiled without AFS_NAMEI_ENV, and " "partition %s is mounted with the 'logging' option. " "Using the inode fileserver backend with 'logging' UFS " "partitions causes volume corruption, so please either " "mount the partition without logging, or use the namei " "fileserver backend. Aborting...\n", part); return -1; } #ifndef AFS_AIX32_ENV if (programType == fileServer) { char salvpath[MAXPATHLEN]; strcpy(salvpath, part); strcat(salvpath, "/FORCESALVAGE"); if (afs_stat(salvpath, &status) == 0) { Log("VInitVnodes: Found %s; aborting\n", salvpath); return -1; } } #endif #if !defined(AFS_LINUX20_ENV) && !defined(AFS_NT40_ENV) strcpy(AFSIDatPath, part); strcat(AFSIDatPath, "/AFSIDat"); #ifdef AFS_NAMEI_ENV if (afs_stat(AFSIDatPath, &status) < 0) { DIR *dirp; struct dirent *dp; dirp = opendir(part); opr_Assert(dirp); while ((dp = readdir(dirp))) { if (dp->d_name[0] == 'V') { Log("This program is compiled with AFS_NAMEI_ENV, but partition %s seems to contain volumes which don't use the namei-interface; aborting\n", part); closedir(dirp); return -1; } } closedir(dirp); } #else /* AFS_NAMEI_ENV */ if (afs_stat(AFSIDatPath, &status) == 0) { Log("This program is compiled without AFS_NAMEI_ENV, but partition %s seems to contain volumes which use the namei-interface; aborting\n", part); return -1; } #ifdef AFS_SGI_XFS_IOPS_ENV if (VerifyXFSInodeSize(part, status.st_fstype) < 0) return -1; #endif #endif /* AFS_NAMEI_ENV */ #endif /* !AFS_LINUX20_ENV && !AFS_NT40_ENV */ #if defined(AFS_DUX40_ENV) && !defined(AFS_NAMEI_ENV) if (status.st_ino != ROOTINO) { Log("%s is not a mounted file system; ignored.\n", part); return 0; } #endif VInitPartition(part, devname, status.st_dev); return 0; } /* VIsAlwaysAttach() checks whether a /vicepX directory should always be * attached (return value 1), or only attached when it is a separately * mounted partition (return value 0). For non-NAMEI environments, it * always returns 0. * * *awouldattach will be set to 1 if the given path at least looks like a vice * partition (that is, if we return 0, the only thing preventing this partition * from being attached is the existence of the AlwaysAttach file), or to 0 * otherwise. *awouldattach is set regardless of whether or not the partition * should always be attached or not. */ static int VIsAlwaysAttach(char *part, int *awouldattach) { #ifdef AFS_NAMEI_ENV struct afs_stat_st st; char checkfile[256]; int ret; #endif /* AFS_NAMEI_ENV */ if (awouldattach) { *awouldattach = 0; } #ifdef AFS_NAMEI_ENV if (strncmp(part, VICE_PARTITION_PREFIX, VICE_PREFIX_SIZE)) return 0; if (awouldattach) { *awouldattach = 1; } strncpy(checkfile, part, 100); strcat(checkfile, OS_DIRSEP); strcat(checkfile, VICE_ALWAYSATTACH_FILE); ret = afs_stat(checkfile, &st); return (ret < 0) ? 0 : 1; #else /* AFS_NAMEI_ENV */ return 0; #endif /* AFS_NAMEI_ENV */ } /* VIsNeverAttach() checks whether a /vicepX directory should never be * attached (return value 1), or follow the normal mounting logic. The * Always Attach flag may override the NeverAttach flag. */ static int VIsNeverAttach(char *part) { struct afs_stat_st st; char checkfile[256]; int ret; if (strncmp(part, VICE_PARTITION_PREFIX, VICE_PREFIX_SIZE)) return 0; strncpy(checkfile, part, 100); strcat(checkfile, OS_DIRSEP); strcat(checkfile, VICE_NEVERATTACH_FILE); ret = afs_stat(checkfile, &st); return (ret < 0) ? 0 : 1; } /* VAttachPartitions2() looks for and attaches /vicepX partitions * where a special file (VICE_ALWAYSATTACH_FILE) exists. This is * used to attach /vicepX directories which aren't on dedicated * partitions, in the NAMEI fileserver. */ static void VAttachPartitions2(void) { #ifdef AFS_NAMEI_ENV DIR *dirp; struct dirent *de; char pname[32]; int wouldattach; dirp = opendir(OS_DIRSEP); while ((de = readdir(dirp))) { strcpy(pname, OS_DIRSEP); strncat(pname, de->d_name, 20); pname[sizeof(pname) - 1] = '\0'; /* Only keep track of "/vicepx" partitions since automounter * may hose us */ if (VIsAlwaysAttach(pname, &wouldattach)) { VCheckPartition(pname, "", 0); } else { struct afs_stat_st st; if (wouldattach && VGetPartition(pname, 0) == NULL && afs_stat(pname, &st) == 0 && S_ISDIR(st.st_mode)) { /* This is a /vicep* dir, and it has not been attached as a * partition. This probably means that this is a /vicep* dir * that is not a separate partition, so just give a notice so * admins are not confused as to why their /vicep* dirs are not * being attached. * * It is possible that the dir _is_ a separate partition and we * failed to attach it earlier, making this message a bit * confusing. But that should be rare, and an error message * about the failure will already be logged right before this, * so it should be clear enough. */ Log("VAttachPartitions: not attaching %s; either it is not a " "separate partition, or it failed to attach (create the " "file %s/" VICE_ALWAYSATTACH_FILE " to force attachment)\n", pname, pname); } } } closedir(dirp); #endif /* AFS_NAMEI_ENV */ } #endif /* AFS_NT40_ENV */ #ifdef AFS_SUN5_ENV int VAttachPartitions(void) { int errors = 0; struct mnttab mnt; FILE *mntfile; if (!(mntfile = afs_fopen(MNTTAB, "r"))) { Log("Can't open %s\n", MNTTAB); perror(MNTTAB); exit(-1); } while (!getmntent(mntfile, &mnt)) { int logging = 0; /* Ignore non ufs or non read/write partitions */ /* but allow zfs too if we're in the NAMEI environment */ if ( #ifdef AFS_NAMEI_ENV (((strcmp(mnt.mnt_fstype, "ufs") && strcmp(mnt.mnt_fstype, "zfs")))) #else (strcmp(mnt.mnt_fstype, "ufs") != 0) #endif || (strncmp(mnt.mnt_mntopts, "ro,ignore", 9) == 0)) continue; /* Skip this Partition? */ if (VIsNeverAttach(mnt.mnt_mountp)) continue; /* If we're going to always attach this partition, do it later. */ if (VIsAlwaysAttach(mnt.mnt_mountp, NULL)) continue; #ifndef AFS_NAMEI_ENV if (hasmntopt(&mnt, "logging") != NULL) { logging = 1; } #endif /* !AFS_NAMEI_ENV */ if (VCheckPartition(mnt.mnt_mountp, mnt.mnt_special, logging) < 0) errors++; } (void)fclose(mntfile); /* Process the always-attach partitions, if any. */ VAttachPartitions2(); return errors; } #endif /* AFS_SUN5_ENV */ #if defined(AFS_SGI_ENV) || (defined(AFS_SUN_ENV) && !defined(AFS_SUN5_ENV)) || defined(AFS_HPUX_ENV) int VAttachPartitions(void) { int errors = 0; FILE *mfd; struct mntent *mntent; if ((mfd = setmntent(MOUNTED, "r")) == NULL) { Log("Problems in getting mount entries(setmntent)\n"); exit(-1); } while (mntent = getmntent(mfd)) { if (!hasmntopt(mntent, MNTOPT_RW)) continue; /* Skip this Partition? */ if (VIsNeverAttach(mntent->mnt_dir)) continue; /* If we're going to always attach this partition, do it later. */ if (VIsAlwaysAttach(mntent->mnt_dir, NULL)) continue; if (VCheckPartition(mntent->mnt_dir, mntent->mnt_fsname, 0) < 0) errors++; } endmntent(mfd); /* Process the always-attach partitions, if any. */ VAttachPartitions2(); return errors; } #endif #ifdef AFS_AIX_ENV /* * (This function was grabbed from df.c) */ int getmount(struct vmount **vmountpp) { int size; struct vmount *vm; int nmounts; /* set initial size of mntctl buffer to a MAGIC NUMBER */ size = BUFSIZ; /* try the operation until ok or a fatal error */ while (1) { if ((vm = malloc(size)) == NULL) { /* failed getting memory for mount status buf */ perror("FATAL ERROR: get_stat malloc failed\n"); exit(-1); } /* * perform the QUERY mntctl - if it returns > 0, that is the * number of vmount structures in the buffer. If it returns * -1, an error occured. If it returned 0, then look in * first word of buffer for needed size. */ if ((nmounts = mntctl(MCTL_QUERY, size, (caddr_t) vm)) > 0) { /* OK, got it, now return */ *vmountpp = vm; return (nmounts); } else if (nmounts == 0) { /* the buffer wasn't big enough .... */ /* .... get required buffer size */ size = *(int *)vm; free(vm); } else { /* some other kind of error occurred */ free(vm); return (-1); } } } int VAttachPartitions(void) { int errors = 0; int nmounts; struct vmount *vmountp; if ((nmounts = getmount(&vmountp)) <= 0) { Log("Problems in getting # of mount entries(getmount)\n"); exit(-1); } for (; nmounts; nmounts--, vmountp = (struct vmount *)((int)vmountp + vmountp->vmt_length)) { char *part = vmt2dataptr(vmountp, VMT_STUB); if (vmountp->vmt_flags & (MNT_READONLY | MNT_REMOVABLE | MNT_REMOTE)) continue; /* Ignore any "special" partitions */ #ifdef AFS_AIX42_ENV #ifndef AFS_NAMEI_ENV { struct superblock fs; /* The Log statements are non-sequiters in the SalvageLog and don't * even appear in the VolserLog, so restrict them to the FileLog. */ if (ReadSuper(&fs, vmt2dataptr(vmountp, VMT_OBJECT)) < 0) { if (programType == fileServer) Log("Can't read superblock for %s, ignoring it.\n", part); continue; } if (IsBigFilesFileSystem(&fs)) { if (programType == fileServer) Log("%s is a big files filesystem, ignoring it.\n", part); continue; } } #endif #endif /* Skip this Partition? */ if (VIsNeverAttach(part)) continue; /* If we're going to always attach this partition, do it later. */ if (VIsAlwaysAttach(part, NULL)) continue; if (VCheckPartition(part, vmt2dataptr(vmountp, VMT_OBJECT), 0) < 0) errors++; } /* Process the always-attach partitions, if any. */ VAttachPartitions2(); return errors; } #endif #if defined(AFS_DUX40_ENV) || defined(AFS_DARWIN_ENV) || defined(AFS_XBSD_ENV) int VAttachPartitions(void) { int errors = 0; struct fstab *fsent; if (setfsent() < 0) { Log("Error listing filesystems.\n"); exit(-1); } while ((fsent = getfsent())) { if (strcmp(fsent->fs_type, "rw") != 0) continue; /* Skip this Partition? */ if (VIsNeverAttach(fsent->fs_file)) continue; /* If we're going to always attach this partition, do it later. */ if (VIsAlwaysAttach(fsent->fs_file, NULL)) continue; if (VCheckPartition(fsent->fs_file, fsent->fs_spec, 0) < 0) errors++; } endfsent(); /* Process the always-attach partitions, if any. */ VAttachPartitions2(); return errors; } #endif #ifdef AFS_NT40_ENV /* VValidVPTEntry * * validate names in vptab. * * Return value: * 1 valid entry * 0 invalid entry */ static int VValidVPTEntry(struct vptab *vpe) { int len = strlen(vpe->vp_name); int i; if (len < VICE_PREFIX_SIZE + 1 || len > VICE_PREFIX_SIZE + 2) return 0; if (strncmp(vpe->vp_name, VICE_PARTITION_PREFIX, VICE_PREFIX_SIZE)) return 0; for (i = VICE_PREFIX_SIZE; i < len; i++) { if (vpe->vp_name[i] < 'a' || vpe->vp_name[i] > 'z') { Log("Invalid partition name %s in registry, ignoring it.\n", vpe->vp_name); return 0; } } if (len == VICE_PREFIX_SIZE + 2) { i = (int)(vpe->vp_name[VICE_PREFIX_SIZE] - 'a') * 26 + (int)(vpe->vp_name[VICE_PREFIX_SIZE + 1] - 'a'); if (i > 255) { Log("Invalid partition name %s in registry, ignoring it.\n", vpe->vp_name); return 0; } } len = strlen(vpe->vp_dev); if (len != 2 || vpe->vp_dev[1] != ':' || vpe->vp_dev[0] < 'A' || vpe->vp_dev[0] > 'Z') { Log("Invalid device name %s in registry, ignoring it.\n", vpe->vp_dev); return 0; } return 1; } static int VCheckPartition(char *partName) { char volRoot[4]; char volFsType[64]; DWORD dwDummy; int err; /* partName is presumed to be of the form "X:" */ (void)sprintf(volRoot, "%c:\\", *partName); if (!GetVolumeInformation(volRoot, /* volume root directory */ NULL, /* volume name buffer */ 0, /* volume name size */ NULL, /* volume serial number */ &dwDummy, /* max component length */ &dwDummy, /* file system flags */ volFsType, /* file system name */ sizeof(volFsType))) { err = GetLastError(); Log("VCheckPartition: Failed to get partition information for %s, ignoring it.\n", partName); return -1; } if (strcmp(volFsType, "NTFS")) { Log("VCheckPartition: Partition %s is not an NTFS partition, ignoring it.\n", partName); return -1; } return 0; } int VAttachPartitions(void) { struct DiskPartition64 *partP, *prevP, *nextP; struct vpt_iter iter; struct vptab entry; if (vpt_Start(&iter) < 0) { Log("No partitions to attach.\n"); return 0; } while (0 == vpt_NextEntry(&iter, &entry)) { if (!VValidVPTEntry(&entry)) { continue; } /* This test for duplicates relies on the fact that the method * of storing the partition names in the NT registry means the same * partition name will never appear twice in the list. */ for (partP = DiskPartitionList; partP; partP = partP->next) { if (*partP->devName == *entry.vp_dev) { Log("Same drive (%s) used for both partition %s and partition %s, ignoring both.\n", entry.vp_dev, partP->name, entry.vp_name); partP->flags = PART_DUPLICATE; break; /* Only one entry will ever be in this list. */ } } if (partP) continue; /* found a duplicate */ if (VCheckPartition(entry.vp_dev) < 0) continue; /* This test allows for manually inserting the FORCESALVAGE flag * and thereby invoking the salvager. scandisk obviously won't be * doing this for us. */ if (programType == fileServer) { struct afs_stat_st status; char salvpath[MAXPATHLEN]; strcpy(salvpath, entry.vp_dev); strcat(salvpath, "\\FORCESALVAGE"); if (afs_stat(salvpath, &status) == 0) { Log("VAttachPartitions: Found %s; aborting\n", salvpath); exit(1); } } VInitPartition(entry.vp_name, entry.vp_dev, *entry.vp_dev - 'A'); } vpt_Finish(&iter); /* Run through partition list and clear out the dupes. */ prevP = nextP = NULL; for (partP = DiskPartitionList; partP; partP = nextP) { nextP = partP->next; if (partP->flags == PART_DUPLICATE) { if (prevP) prevP->next = partP->next; else DiskPartitionList = partP->next; free(partP); } else prevP = partP; } return 0; } #endif #ifdef AFS_LINUX22_ENV int VAttachPartitions(void) { int errors = 0; FILE *mfd; struct mntent *mntent; if ((mfd = setmntent("/proc/mounts", "r")) == NULL) { if ((mfd = setmntent("/etc/mtab", "r")) == NULL) { Log("Problems in getting mount entries(setmntent)\n"); exit(-1); } } while ((mntent = getmntent(mfd))) { /* Skip this Partition? */ if (VIsNeverAttach(mntent->mnt_dir)) continue; /* If we're going to always attach this partition, do it later. */ if (VIsAlwaysAttach(mntent->mnt_dir, NULL)) continue; if (VCheckPartition(mntent->mnt_dir, mntent->mnt_fsname, 0) < 0) errors++; } endmntent(mfd); /* Process the always-attach partitions, if any. */ VAttachPartitions2(); return errors; } #endif /* AFS_LINUX22_ENV */ /* This routine is to be called whenever the actual name of the partition * is required. The canonical name is still in part->name. */ char * VPartitionPath(struct DiskPartition64 *part) { #ifdef AFS_NT40_ENV return part->devName; #else return part->name; #endif } /* get partition structure, abortp tells us if we should abort on failure */ struct DiskPartition64 * VGetPartition_r(char *name, int abortp) { struct DiskPartition64 *dp; #ifdef AFS_DEMAND_ATTACH_FS dp = VLookupPartition_r(name); #else /* AFS_DEMAND_ATTACH_FS */ for (dp = DiskPartitionList; dp; dp = dp->next) { if (strcmp(dp->name, name) == 0) break; } #endif /* AFS_DEMAND_ATTACH_FS */ if (abortp) opr_Assert(dp != NULL); return dp; } struct DiskPartition64 * VGetPartition(char *name, int abortp) { struct DiskPartition64 *retVal; VOL_LOCK; retVal = VGetPartition_r(name, abortp); VOL_UNLOCK; return retVal; } #ifdef AFS_NT40_ENV void VSetPartitionDiskUsage_r(struct DiskPartition64 *dp) { ULARGE_INTEGER free_user, total, free_total; int ufree, tot, tfree; if (!GetDiskFreeSpaceEx (VPartitionPath(dp), &free_user, &total, &free_total)) { printf("Failed to get disk space info for %s, error = %d\n", dp->name, GetLastError()); return; } /* Convert to 1K units. */ ufree = (int)Int64ShraMod32(free_user.QuadPart, 10); tot = (int)Int64ShraMod32(total.QuadPart, 10); tfree = (int)Int64ShraMod32(free_total.QuadPart, 10); dp->minFree = tfree - ufree; /* only used in VPrintDiskStats_r */ dp->totalUsable = tot; dp->free = tfree; } #else void VSetPartitionDiskUsage_r(struct DiskPartition64 *dp) { int bsize, code; afs_int64 totalblks, free, used, availblks; int reserved; #ifdef afs_statvfs struct afs_statvfs statbuf; #else struct afs_statfs statbuf; #endif if (dp->flags & PART_DONTUPDATE) return; /* Note: we don't bother syncing because it's only an estimate, update * is syncing every 30 seconds anyway, we only have to keep the disk * approximately 10% from full--you just can't get the stuff in from * the net fast enough to worry */ #ifdef afs_statvfs code = afs_statvfs(dp->name, &statbuf); #else code = afs_statfs(dp->name, &statbuf); #endif if (code < 0) { Log("statfs of %s failed in VSetPartitionDiskUsage (errno = %d)\n", dp->name, errno); return; } if (statbuf.f_blocks == -1) { /* Undefined; skip stats.. */ Log("statfs of %s failed in VSetPartitionDiskUsage\n", dp->name); return; } totalblks = statbuf.f_blocks; free = statbuf.f_bfree; reserved = free - statbuf.f_bavail; #ifdef afs_statvfs bsize = statbuf.f_frsize; #else bsize = statbuf.f_bsize; #endif availblks = totalblks - reserved; dp->f_files = statbuf.f_files; /* max # of files in partition */ /* Now free and totalblks are in fragment units, but we want them in * 1K units. */ if (bsize >= 1024) { free *= (bsize / 1024); totalblks *= (bsize / 1024); availblks *= (bsize / 1024); reserved *= (bsize / 1024); } else { free /= (1024 / bsize); totalblks /= (1024 / bsize); availblks /= (1024 / bsize); reserved /= (1024 / bsize); } /* now compute remaining figures */ used = totalblks - free; dp->minFree = reserved; /* only used in VPrintDiskStats_r */ dp->totalUsable = availblks; dp->free = availblks - used; /* this is exactly f_bavail */ } #endif /* AFS_NT40_ENV */ void VSetPartitionDiskUsage(struct DiskPartition64 *dp) { VOL_LOCK; VSetPartitionDiskUsage_r(dp); VOL_UNLOCK; } void VResetDiskUsage_r(void) { struct DiskPartition64 *dp; for (dp = DiskPartitionList; dp; dp = dp->next) { VSetPartitionDiskUsage_r(dp); #ifndef AFS_PTHREAD_ENV IOMGR_Poll(); #endif /* !AFS_PTHREAD_ENV */ } } void VResetDiskUsage(void) { VOL_LOCK; VResetDiskUsage_r(); VOL_UNLOCK; } void VAdjustDiskUsage_r(Error * ec, Volume * vp, afs_sfsize_t blocks, afs_sfsize_t checkBlocks) { *ec = 0; /* why blocks instead of checkBlocks in the check below? Otherwise, any check * for less than BlocksSpare would skip the error-checking path, and we * could grow existing files forever, not just for another BlocksSpare * blocks. */ if (blocks > 0) { #ifdef AFS_AIX32_ENV afs_int32 rem, minavail; if ((rem = vp->partition->free - checkBlocks) < (minavail = (vp->partition-> totalUsable * aixlow_water) / 100)) #else if (vp->partition->free - checkBlocks < 0) #endif *ec = VDISKFULL; else if (V_maxquota(vp) && V_diskused(vp) + checkBlocks > V_maxquota(vp)) *ec = VOVERQUOTA; } vp->partition->free -= blocks; V_diskused(vp) += blocks; } void VAdjustDiskUsage(Error * ec, Volume * vp, afs_sfsize_t blocks, afs_sfsize_t checkBlocks) { VOL_LOCK; VAdjustDiskUsage_r(ec, vp, blocks, checkBlocks); VOL_UNLOCK; } int VDiskUsage_r(Volume * vp, afs_sfsize_t blocks) { if (blocks > 0) { #ifdef AFS_AIX32_ENV afs_int32 rem, minavail; if ((rem = vp->partition->free - blocks) < (minavail = (vp->partition-> totalUsable * aixlow_water) / 100)) #else if (vp->partition->free - blocks < 0) #endif return (VDISKFULL); } vp->partition->free -= blocks; return 0; } int VDiskUsage(Volume * vp, afs_sfsize_t blocks) { int retVal; VOL_LOCK; retVal = VDiskUsage_r(vp, blocks); VOL_UNLOCK; return retVal; } void VPrintDiskStats_r(void) { struct DiskPartition64 *dp; for (dp = DiskPartitionList; dp; dp = dp->next) { if (dp->free < 0) { Log("Partition %s: %lld " " available 1K blocks (minfree=%lld), " "overallocated by %lld blocks\n", dp->name, dp->totalUsable, dp->minFree, -dp->free); } else { Log("Partition %s: %lld" " available 1K blocks (minfree=%lld), " "%lld free blocks\n", dp->name, dp->totalUsable, dp->minFree, dp->free); } } } void VPrintDiskStats(void) { VOL_LOCK; VPrintDiskStats_r(); VOL_UNLOCK; } #ifdef AFS_NT40_ENV /* Need a separate lock file on NT, since NT only has mandatory file locks. */ #define LOCKFILE "LOCKFILE" void VLockPartition_r(char *name) { struct DiskPartition64 *dp = VGetPartition_r(name, 0); OVERLAPPED lap; if (!dp) return; if (dp->lock_fd == INVALID_FD) { char path[64]; int rc; (void)sprintf(path, "%s\\%s", VPartitionPath(dp), LOCKFILE); dp->lock_fd = (FD_t)CreateFile(path, GENERIC_WRITE, FILE_SHARE_READ | FILE_SHARE_WRITE, NULL, CREATE_ALWAYS, FILE_ATTRIBUTE_HIDDEN, NULL); opr_Assert(dp->lock_fd != INVALID_FD); memset(&lap, 0, sizeof(lap)); rc = LockFileEx((HANDLE) dp->lock_fd, LOCKFILE_EXCLUSIVE_LOCK, 0, 1, 0, &lap); opr_Assert(rc); } } void VUnlockPartition_r(char *name) { struct DiskPartition64 *dp = VGetPartition_r(name, 0); OVERLAPPED lap; if (!dp) return; /* no partition, will fail later */ memset(&lap, 0, sizeof(lap)); UnlockFileEx((HANDLE) dp->lock_fd, 0, 1, 0, &lap); CloseHandle((HANDLE) dp->lock_fd); dp->lock_fd = INVALID_FD; } #else /* AFS_NT40_ENV */ #if defined(AFS_HPUX_ENV) #define BITS_PER_CHAR (8) #define BITS(type) (sizeof(type) * BITS_PER_CHAR) #define LOCKRDONLY_OFFSET ((PRIV_LOCKRDONLY - 1) / BITS(int)) #endif /* defined(AFS_HPUX_ENV) */ void VLockPartition_r(char *name) { struct DiskPartition64 *dp = VGetPartition_r(name, 0); char *partitionName; int retries, code; struct timeval pausing; #if defined(AFS_HPUX_ENV) int lockfRtn; struct privgrp_map privGrpList[PRIV_MAXGRPS]; unsigned int *globalMask; int globalMaskIndex; #endif /* defined(AFS_HPUX_ENV) */ #if defined(AFS_DARWIN_ENV) char lockfile[MAXPATHLEN]; #endif /* defined(AFS_DARWIN_ENV) */ #ifdef AFS_NAMEI_ENV #ifdef AFS_AIX42_ENV char LockFileName[MAXPATHLEN + 1]; sprintf((char *)&LockFileName, "%s/AFSINODE_FSLock", name); partitionName = (char *)&LockFileName; #endif #endif if (!dp) return; /* no partition, will fail later */ if (dp->lock_fd != INVALID_FD) return; #if defined(AFS_SUN5_ENV) || defined(AFS_AIX41_ENV) #if !defined(AFS_AIX42_ENV) || !defined(AFS_NAMEI_ENV) partitionName = dp->devName; #endif code = O_RDWR; #elif defined(AFS_DARWIN_ENV) strlcpy((partitionName = lockfile), dp->name, sizeof(lockfile)); strlcat(lockfile, "/.lock.afs", sizeof(lockfile)); code = O_RDONLY | O_CREAT; #else partitionName = dp->name; code = O_RDONLY; #endif for (retries = 25; retries; retries--) { if (code & O_CREAT) dp->lock_fd = afs_open(partitionName, code, 0644); else dp->lock_fd = afs_open(partitionName, code); if (dp->lock_fd != INVALID_FD) break; if (errno == ENOENT) code |= O_CREAT; pausing.tv_sec = 0; pausing.tv_usec = 500000; select(0, NULL, NULL, NULL, &pausing); } opr_Assert(retries != 0); #if defined (AFS_HPUX_ENV) opr_Verify(getprivgrp(privGrpList) == 0); /* * In general, it will difficult and time-consuming ,if not impossible, * to try to find the privgroup to which this process belongs that has the * smallest membership, to minimise the security hole. So, we use the privgrp * to which everybody belongs. */ /* first, we have to find the global mask */ for (globalMaskIndex = 0; globalMaskIndex < PRIV_MAXGRPS; globalMaskIndex++) { if (privGrpList[globalMaskIndex].priv_groupno == PRIV_GLOBAL) { globalMask = &(privGrpList[globalMaskIndex].priv_mask[LOCKRDONLY_OFFSET]); break; } } if (((*globalMask) & privmask(PRIV_LOCKRDONLY)) == 0) { /* allow everybody to set a lock on a read-only file descriptor */ (*globalMask) |= privmask(PRIV_LOCKRDONLY); opr_Verify(setprivgrp(PRIV_GLOBAL, privGrpList[globalMaskIndex].priv_mask) == 0); lockfRtn = lockf(dp->lock_fd, F_LOCK, 0); /* remove the privilege granted to everybody to lock a read-only fd */ (*globalMask) &= ~(privmask(PRIV_LOCKRDONLY)); opr_Verify(setprivgrp(PRIV_GLOBAL, privGrpList[globalMaskIndex].priv_mask) == 0); } else { /* in this case, we should be able to do this with impunity, anyway */ lockfRtn = lockf(dp->lock_fd, F_LOCK, 0); } opr_Assert(lockfRtn != -1); #else #if defined(AFS_AIX_ENV) || defined(AFS_SUN5_ENV) opr_Verify(lockf(dp->lock_fd, F_LOCK, 0) != -1); #else opr_Verify(flock(dp->lock_fd, LOCK_EX) == 0); #endif /* defined(AFS_AIX_ENV) || defined(AFS_SUN5_ENV) */ #endif } void VUnlockPartition_r(char *name) { struct DiskPartition64 *dp = VGetPartition_r(name, 0); if (!dp) return; /* no partition, will fail later */ close(dp->lock_fd); dp->lock_fd = INVALID_FD; } #endif /* AFS_NT40_ENV */ void VLockPartition(char *name) { VOL_LOCK; VLockPartition_r(name); VOL_UNLOCK; } void VUnlockPartition(char *name) { VOL_LOCK; VUnlockPartition_r(name); VOL_UNLOCK; } #ifdef AFS_DEMAND_ATTACH_FS /* new-style partition locks; these are only to have some mutual exclusion * between the VGC scanner and volume utilies creating/altering vol headers */ /** * lock a partition's vol headers. * * @param[in] dp the partition to lock * @param[in] locktype READ_LOCK or WRITE_LOCK * * @return operation status * @retval 0 success */ int VPartHeaderLock(struct DiskPartition64 *dp, int locktype) { int code; /* block on acquiring the lock */ int nonblock = 0; code = VGetDiskLock(&dp->headerLock, locktype, nonblock); if (code) { Log("VPartHeaderLock: error %d locking partititon %s\n", code, VPartitionPath(dp)); } return code; } /** * unlock a partition's vol headers. * * @param[in] dp the partition to unlock * @param[in] locktype READ_LOCK or WRITE_LOCK */ void VPartHeaderUnlock(struct DiskPartition64 *dp, int locktype) { VReleaseDiskLock(&dp->headerLock, locktype); } /* XXX not sure this will work on AFS_NT40_ENV * needs to be tested! */ /** * lookup a disk partition object by its index number. * * @param[in] id partition index number * @param[in] abortp see abortp usage note below * * @return disk partition object * @retval NULL no such disk partition * * @note when abortp is non-zero, lookups which would return * NULL will result in an assertion failure * * @pre VOL_LOCK must be held * * @internal volume package internal use only */ struct DiskPartition64 * VGetPartitionById_r(afs_int32 id, int abortp) { struct DiskPartition64 *dp = NULL; if ((id >= 0) && (id <= VOLMAXPARTS)) { dp = DiskPartitionTable[id]; } if (abortp) { opr_Assert(dp != NULL); } return dp; } /** * lookup a disk partition object by its index number. * * @param[in] id partition index number * @param[in] abortp see abortp usage note below * * @return disk partition object * @retval NULL no such disk partition * * @note when abortp is non-zero, lookups which would return * NULL will result in an assertion failure */ struct DiskPartition64 * VGetPartitionById(afs_int32 id, int abortp) { struct DiskPartition64 * dp; VOL_LOCK; dp = VGetPartitionById_r(id, abortp); VOL_UNLOCK; return dp; } static struct DiskPartition64 * VLookupPartition_r(char * path) { afs_int32 id = volutil_GetPartitionID(path); if (id < 0 || id > VOLMAXPARTS) return NULL; return DiskPartitionTable[id]; } static void AddPartitionToTable_r(struct DiskPartition64 *dp) { opr_Assert(dp->index >= 0 && dp->index <= VOLMAXPARTS); DiskPartitionTable[dp->index] = dp; } #if 0 static void DeletePartitionFromTable_r(struct DiskPartition64 *dp) { opr_Assert(dp->index >= 0 && dp->index <= VOLMAXPARTS); DiskPartitionTable[dp->index] = NULL; } #endif #endif /* AFS_DEMAND_ATTACH_FS */