| Age | Commit message (Collapse) | Author |
|---|
|
|
|
ok krw@ otto@
|
|
objection form the usual suspects
|
|
and speeds things up in a lot of cases.
Prompted by Amit Kulkarni; ok krw@ on a slighly diffrenent incarnation
|
|
when fsck is run against a disklabel UID. This allows a user to determine
which device is really being scanned.
ok krw@
|
|
|
|
large on very large filesystems; reported by Benny Lofgren; partly
from FreeBSD. ok deraadt@ beck@ millert@
|
|
blockcheck() since we try to stat() the UID. This means that we fail to
reload the mount if we have indeed cleaned the read-only root file system.
To avoid this, rerun blockcheck() if the real name is different to the
original device name, once we have opened the device with opendev(3).
ok krw@ thib@
|
|
ok jsing@ krw@
|
|
to operate with disklabel UIDs.
ok marco@ krw@ otto@
|
|
gets rid of #include <sys/dkio.h> in sys/ioctl.h and adds #include
<sys/dkio.h> to the places that actually want and use the disk
ioctls.
this became an issue when krw@'s X build failed when he was testing
a change to dkio.h.
tested by krw@
help from and ok miod@
|
|
Prompted by J.C. Robers; ok jmc@
|
|
|
|
files. If any information found in these documents is worthwhile and you
miss it, please make the time to work it into the manual pages (which people
actually do read).
ok guenther
|
|
instead, .It is required. Thus, move .Pp and text before the .Bl,
and remove the .Pp altogether where it is not needed.
Syntax errors found by mandoc(1), also required to fix the mandoc build;
feedback and ok jmc@, and sobrado@ also supports the direction.
|
|
(one case was a false positive, but one is a modified copy of the other)
|
|
unmaintainable). these days, people use source. these id's do not provide
any benefit, and do hurt the small install media
(the 33,000 line diff is essentially mechanical)
ok with the idea millert, ok dms
|
|
against forms of cg header corruption as experienced by todd@; ok
millert@
|
|
negative and cause SEGVs. Handle this the same as an out of range
blockno. ok jsg@ (also victim) pedro@ thib@
|
|
ok millert@
|
|
automatic symbol that has this information. PAGE_SIZE is simply not
portable, or even fixed on some systems.
ok otto
|
|
ok millert@
|
|
byte. Original diff by drahn@; twists by me; ok millert@ thib@
|
|
|
|
different locations than before. Actually, the disklabel does not
contain enough info to completely reconstruct the locations of all
alternate sb's. So use a hardcoded list of all possible 1st alternate
sb locations, but don't forget to verify against the label.
ok millert@ thib@
|
|
mickey; ok millert@
|
|
|
|
that gcc likes better
|
|
|
|
|
|
|
|
Fixes PR 5494; with pedro@; ok pedro@ millert@
|
|
|
|
|
|
free room in struct partition by packing fragment size and
fragments/block more tighlty and use the resulting space to make
offset and size 48 bits. For the disk part we use spare fields in
struct disklabel. Kernel converts in-mem copy of the on-disk
label if needed, disklabel(8) writes new version.
We are careful to only change fields not used by bootloaders.
Conception of basic scheme by deraadt.
ok deraadt@ krw@
|
|
|
|
partially initialized summary info struct. ok millert@ pedro@
|
|
does not use fs_cpg at all. From FreeBSD. OK otto@
|
|
architectures, CGSIZE() was returning something sligthly over one block.
The 'new' fsck would round this value up to a fragment boundary, and end
up trying to access memory beyond allocated space. From mickey@, okay
pedro@, millert@ and otto@.
|
|
now jmc does not use 8TB in /tmp anymore. ok pedro@
|
|
a ffs1 alternate sb at the primary ffs2 sb location. Use the same test
as the kernel does. ok millert@ pedro@
|
|
the ufs2 changes in FreeBSD by Kirk Mckusick.
|
|
sizes; requested by mickey@
|
|
accessing FFS1 fields, okay art@, quite some testing by ckuethe@, simon@
and thib@, thanks.
|
|
fragment sizes, so just make those conditions fatal errors and tell
the calling code to go look for alternate superblocks. ok millert@
|
|
|
|
locations based on the disklabel.
- unit bug: disklabel size is in sectors, but superblock
fs size is in fragments.
- The disklabel contains the maximum superblock size, not the actual
as used by newfs.
- problem is that mkfs reduces track/sector and secors/track to
make the cylindergroup fit in a fs block. In this case the values
in the disklabel do not predict what's in the superblock.
This fixes all these problems. It correctly predicts the location
of all superblocks on all my filesystems I have online (sized between
100MB and 1000GB). Also tested with filesystems with different
block and fragment sizes.
Written this with a goto to make the diff easier to read.
ok krw@ deraadt@
|
|
signal handlers, calling free() and not being reentrant. This commit
fixes to most importamt problem (^C on the "MARK FILESYSTEM CLEAN?"
question causes a double free), but actually SIGINT handling should
be completely rewritten. ok millert@ deraadt@
|
|
superblock might lead to crashes or other mishap. Now my fuzzed fs
images no longer crash fsck_ffs. ok mickey@ pedro@ millert@
|
|
Solves a case were fsck_ffs was causing a segv. If it didn't do that
it would have mangled the filesystem later, very probably.
Diff from FreeBSD; ok millert@ pedro@
|
