- Always look for labels in the first two sectors
- Update all existing labels, but only write new labels to netbsd mbr
partitions, and the first/second sector of disks that don't have an mbr.
Moving disks between systems still sucks bigtime!
- raw patition is either c or d, if d then c is reserved
- max partitions might be 8 or 16, nothing in the label gives the maximum
- endianness
the time pool_get() calls pool_catchup(), pp has been free'd but it is still
in the "entered" state. The chain pool_catchup() -> pool_allocator_alloc()
-> pool_reclaim() on pp fails because pp is still in the "entered" state.
Call pr_leave() before calling calling pool_catchup() to avoid this.
Thanks for the excellent analysis!
Be more flexible in what we accept as a valid LINTSTUB directive.
Don't abort on first error.
Separate LINTSTUB comments look ugly if the function/variable already
has a descriptive comment. People don't like to write ugly code.
Now one can write:
/*
* LINTSTUB: Func: type function(args)
* Some descriptive comment about the function.
*/
buffer when the passed nested buffer has no B_ERROR flag set but not all
was transfered for the nested iobuf extent.
Discussed on tech-kern and ok'd by Takashi
otherwise generate an UVM trap or will access random memory. This is due to
the dereference of vp->v_specmountpoint that is really
vp->v_specinfo->si_mountpoint. The field v_specinfo is multiplexed with
other structs in the vun union in struct vnode like struct socket.
The patch adds a sanity check for accessing the specinfo fields by only
allowing VBLK nodes to be passed. In theory also VCHR could be valid since
its also a special node though mounting is only done on VBLK so be strict.
Ok'd by yamt.
- for structure fields that are conditionally present,
make those fields always present.
- for functions which are conditionally inline, make them never inline.
- remove some other functions which are conditionally defined but
don't actually do anything anymore.
- make a lock-debugging function conditional on only LOCKDEBUG.
as discussed on tech-kern some time back.
for each fork-wait cycles.
- updatepri: factor out the code to decay estcpu so that it can be used
by scheduler_wait_hook.
- scheduler_fork_hook: record how much estcpu is inherited from
the parent process.
- scheduler_wait_hook: don't add back inherited estcpu to the parent.
otherwise, once the corresponding bit in the inode bitmap is cleared,
an unrelated inode with the same inode number can be allocated and
ufs_ihashget() picks a stale in-core vnode for it.
PR/32301 by Matthias Scheler.
- pool_allocator_alloc: drain ourselves as well,
so that pool_cache on us is drained as well.
- pool_cache_put_paddr: destruct objects if underlying pool is starved.
- pool_get: on kva starvation, wake up once a second and try again.
Fixes:
PR/32287: Processes hang in "mclpl"
PR/32330: shark kernel hangs under memory load.
- don't compare a scaled value with a unscaled value.
- actually, 7 times the loadfactor is necessary to decay p_estcpu enough,
even before the recent p_estcpu changes.
after the recent p_estcpu change, 8 times loadavg decay is needed.
- fix a comment to match with the recent reality.
It breaks the code that generates a default label (with partition 'a'
covering the entire volume - which is what everything expects) for disks that
don't have a NetBSD label nor an MBR, but do have a single filesytem covering
the entire volume.
yamt