Home
last modified time | relevance | path

Searched refs:we (Results 1 – 25 of 694) sorted by relevance

12345678910>>...28

/linux-3.4.99/Documentation/x86/
Dentry_64.txt49 so. If we mess that up even slightly, we crash.
51 So when we have a secondary entry, already in kernel mode, we *must
52 not* use SWAPGS blindly - nor must we forget doing a SWAPGS when it's
81 If we are at an interrupt or user-trap/gate-alike boundary then we can
83 whether SWAPGS was already done: if we see that we are a secondary
84 entry interrupting kernel mode execution, then we know that the GS
85 base has already been switched. If it says that we interrupted
86 user-space execution then we must do the SWAPGS.
88 But if we are in an NMI/MCE/DEBUG/whatever super-atomic entry context,
90 stack but before we executed SWAPGS, then the only safe way to check
[all …]
/linux-3.4.99/fs/jffs2/
DTODO5 reservations will necessarily be pessimistic. With this we could even
6 do shared writable mmap, if we can have a fs hook for do_wp_page() to
11 - checkpointing (do we need this? scan is quite fast)
25 nodes to a different one, we can separate clean nodes from those which
28 - Stop keeping name in-core with struct jffs2_full_dirent. If we keep the hash in
29 the full dirent, we only need to go to the flash in lookup() when we think we've
/linux-3.4.99/tools/perf/
Dbuiltin-timechart.c432 struct wake_event *we; in sched_wakeup() local
436 we = malloc(sizeof(struct wake_event)); in sched_wakeup()
437 if (!we) in sched_wakeup()
440 memset(we, 0, sizeof(struct wake_event)); in sched_wakeup()
441 we->time = timestamp; in sched_wakeup()
442 we->waker = pid; in sched_wakeup()
445 we->waker = -1; in sched_wakeup()
447 we->wakee = wake->pid; in sched_wakeup()
448 we->next = wake_events; in sched_wakeup()
449 wake_events = we; in sched_wakeup()
[all …]
/linux-3.4.99/arch/frv/kernel/
Dbreak.S83 # traps will be enabled, so we have to do this now
92 # determine whether we have stepped through into an exception
93 # - we need to take special action to suspend h/w single stepping if we've done
273 # do the bit we had to skip
305 # we single-stepped into an interrupt handler whilst interrupts were merely virtually disabled
334 # we stepped through into the virtual interrupt reenablement trap
336 # we also want to single step anyway, but after fixing up so that we get an event on the
381 # do the bit we had to skip
412 # we'll want to try the trap stub again
427 # we'll swap the real return address for one with a BREAK insn so that we can re-enable
[all …]
/linux-3.4.99/Documentation/
Drt-mutex-design.txt28 with the resource. This is a priority inversion. What we want to prevent
40 to release the lock, because for all we know, B is a CPU hog and will
61 for this document. Here we only discuss PI.
144 To show where two chains merge, we could add another process F and
154 Here we show both chains:
162 For PI to work, the processes at the right end of these chains (or we may
166 Also since a mutex may have more than one process blocked on it, we can
167 have multiple chains merge at mutexes. If we add another process G that is
238 the nesting of mutexes. Let's look at the example where we have 3 mutexes,
283 Now we add 4 processes that run each of these functions separately.
[all …]
Dprio_tree.txt26 tree cannot index them all, we build an overflow-sub-tree that indexes such
31 the same vm_pgoff (radix_index) and if necessary we build different
33 in figure we have 3 overflow-sub-trees corresponding to radix indices
74 Note that we use prio_tree_root->index_bits to optimize the height
76 set according to the maximum end_vm_pgoff mapped, we are sure that all
78 we only use the first prio_tree_root->index_bits as radix_index.
79 Whenever index_bits is increased in prio_tree_expand, we shuffle the tree
87 path which is not desirable. Hence, we do not optimize the height of the
94 same start_vm_pgoff but different end_vm_pgoffs. Therefore, we normally
99 However, in most of the common cases we do not need overflow-sub-trees,
Dlockstat.txt43 With these hooks we provide the following statistics:
47 wait time min - shortest (non-0) time we ever had to wait for a lock
48 max - longest time we ever had to wait for a lock
49 total - total time we spend waiting on this lock
51 acquisitions - number of times we took the lock
52 hold time min - shortest (non-0) time we ever held the lock
53 max - longest time we ever held the lock
128 points are the points we're contending with.
/linux-3.4.99/Documentation/filesystems/
Dxfs-delayed-logging-design.txt25 That is, if we have a sequence of changes A through to F, and the object was
26 written to disk after change D, we would see in the log the following series
91 relogging technique XFS uses is that we can be relogging changed objects
92 multiple times before they are committed to disk in the log buffers. If we
98 contains all the changes from the previous changes. In other words, we have one
100 wasting space. When we are doing repeated operations on the same set of
103 log would greatly reduce the amount of metadata we write to the log, and this
110 formatting the changes in a transaction to the log buffer. Hence we cannot avoid
113 Delayed logging is the name we've given to keeping and tracking transactional
163 changes to the log buffers, we need to ensure that the object we are formatting
[all …]
Dpath-lookup.txt49 the path given by the name's starting point (which we know in advance -- eg.
55 A parent, of course, must be a directory, and we must have appropriate
79 In order to lookup a dcache (parent, name) tuple, we take a hash on the tuple
81 in that bucket is then walked, and we do a full comparison of each entry
148 However, when inserting object 2 onto a new list, we end up with this:
161 Because we didn't wait for a grace period, there may be a concurrent lookup
182 As explained above, we would like to do path walking without taking locks or
188 than reloading from the dentry later on (otherwise we'd have interesting things
192 no non-atomic stores to shared data), and to recheck the seqcount when we are
194 Avoiding destructive or changing operations means we can easily unwind from
[all …]
Ddirectory-locking7 1) read access. Locking rules: caller locks directory we are accessing.
44 First of all, at any moment we have a partial ordering of the
50 attempts to acquire lock on B, A will remain the parent of B until we
56 renames will be blocked on filesystem lock and we don't start changing
57 the order until we had acquired all locks).
80 would have a contended child and we had assumed that no object is its
85 of its descendents is locked by cross-directory rename (otherwise we
88 to (2) the order hadn't changed since we had acquired filesystem lock.
89 But locking rules for cross-directory rename guarantee that we do not
102 we had acquired filesystem lock and rename() would fail with -ELOOP in that
[all …]
/linux-3.4.99/drivers/block/paride/
DTransition-notes9 ps_spinlock. C is always preceded by B, since we can't reach it
10 other than through B and we don't drop ps_spinlock between them.
14 A and each B is preceded by either A or C. Moments when we enter
37 * in ps_tq_int(): from the moment when we get ps_spinlock() to the
73 we would have to be called for the PIA that got ->claimed_cont
83 it is holding pd_lock. The only place within the area where we
87 we were acquiring the lock, (1) would be already false, since
89 If it was 0 before we tried to acquire pd_lock, (2) would be
96 (4) is done the same way - all places where we release pi_spinlock within
100 in the area, under pi_spinlock and we do not release it until after leaving
[all …]
/linux-3.4.99/Documentation/usb/
DWUSB-Design-overview.txt173 [*] The reason for RCs being also devices is that not only we can
201 So let's say we connect a dongle to the system: it is detected and
204 Now we have a real HWA device connected and
223 So assuming we have devices and we have agreed for a channel to connect
224 on (let's say 9), we put the new RC to beacon:
292 ID and tell the HC to use all that. Then we start it. This means the HC
318 the device. First we allocate a /fake port/ and assign an
319 unauthenticated address (128 to 255--what we really do is
323 So now we are in the reset path -- we know we have a non-yet enumerated
324 device with an unauthorized address; we ask user space to authenticate
[all …]
/linux-3.4.99/drivers/scsi/aic7xxx/
Daic79xx.seq85 * If we have completions stalled waiting for the qfreeze
109 * ENSELO is cleared by a SELDO, so we must test for SELDO
169 * Since this status did not consume a FIFO, we have to
170 * be a bit more dilligent in how we check for FIFOs pertaining
178 * count in the SCB. In this case, we allow the routine servicing
183 * we detect case 1, we will properly defer the post of the SCB
222 * bad SCSI status (currently only for underruns), we
223 * queue the SCB for normal completion. Otherwise, we
258 * If we have relatively few commands outstanding, don't
303 * one byte of lun information we support.
[all …]
Daic7xxx.seq52 * After starting the selection hardware, we check for reconnecting targets
54 * bus arbitration. The problem with this is that we must keep track of the
55 * SCB that we've already pulled from the QINFIFO and started the selection
56 * on just in case the reselection wins so that we can retry the selection at
104 * We have at least one queued SCB now and we don't have any
124 * before we completed the DMA operation. If it was,
211 /* The Target ID we were selected at */
239 * Watch ATN closely now as we pull in messages from the
285 * we've got a failed selection and must transition to bus
333 * Reselection has been initiated by a target. Make a note that we've been
[all …]
/linux-3.4.99/arch/cris/arch-v10/mm/
Dfault.c51 int miss, we, writeac; in handle_mmu_bus_fault() local
67 we = IO_EXTRACT(R_MMU_CAUSE, we_excp, cause); in handle_mmu_bus_fault()
71 regs->irp, address, miss, inv, we, acc, index, page_id)); in handle_mmu_bus_fault()
77 do_page_fault(address, regs, 1, we); in handle_mmu_bus_fault()
/linux-3.4.99/Documentation/power/
Dfreezing-of-tasks.txt17 we only consider hibernation, but the description also applies to suspend).
28 it loop until PF_FROZEN is cleared for it. Then, we say that the task is
75 - freezes all tasks (including kernel threads) because we can't freeze
79 - thaws only kernel threads; this is particularly useful if we need to do
81 userspace tasks, or if we want to postpone the thawing of userspace tasks
84 - thaws all tasks (including kernel threads) because we can't thaw userspace
95 IV. Why do we do that?
100 hibernation. At the moment we have no simple means of checkpointing
102 metadata on disks, we cannot bring them back to the state from before the
113 2. Next, to create the hibernation image we need to free a sufficient amount of
[all …]
/linux-3.4.99/Documentation/RCU/
Drculist_nulls.txt23 * reuse these object before the RCU grace period, we
26 if (obj->key != key) { // not the object we expected
87 * we need to make sure obj->key is updated before obj->next
98 Nothing special here, we can use a standard RCU hlist deletion.
112 With hlist_nulls we can avoid extra smp_rmb() in lockless_lookup()
115 For example, if we choose to store the slot number as the 'nulls'
116 end-of-list marker for each slot of the hash table, we can detect
120 is not the slot number, then we must restart the lookup at
135 if (obj->key != key) { // not the object we expected
142 * if the nulls value we got at the end of this lookup is
[all …]
/linux-3.4.99/Documentation/block/
Ddeadline-iosched.txt21 service time for a request. As we focus mainly on read latencies, this is
50 When we have to move requests from the io scheduler queue to the block
51 device dispatch queue, we always give a preference to reads. However, we
53 how many times we give preference to reads over writes. When that has been
54 done writes_starved number of times, we dispatch some writes based on the
69 that comes at basically 0 cost we leave that on. We simply disable the
/linux-3.4.99/drivers/scsi/aic7xxx_old/
Daic7xxx.seq43 * After starting the selection hardware, we check for reconnecting targets
45 * bus arbitration. The problem with this is that we must keep track of the
46 * SCB that we've already pulled from the QINFIFO and started the selection
47 * on just in case the reselection wins so that we can retry the selection at
86 * interrupts on the "background" channel. So, if we
110 * We have at least one queued SCB now and we don't have any
111 * SCBs in the list of SCBs awaiting selection. If we have
135 * Preset the residual fields in case we never go through a data phase.
136 * This isn't done by the host so we can avoid a DMA to clear these
236 * Reselection has been initiated by a target. Make a note that we've been
[all …]
/linux-3.4.99/Documentation/networking/
Dfib_trie.txt33 verify that they actually do match the key we are searching for.
68 fib_find_node(). Inserting a new node means we might have to run the
103 slower than the corresponding fib_hash function, as we have to walk the
120 trie, key segment by key segment, until we find a leaf. check_leaf() does
123 If we find a match, we are done.
125 If we don't find a match, we enter prefix matching mode. The prefix length,
127 and we backtrack upwards through the trie trying to find a longest matching
133 the child index until we find a match or the child index consists of nothing but
136 At this point we backtrack (t->stats.backtrack++) up the trie, continuing to
139 At this point we will repeatedly descend subtries to look for a match, and there
/linux-3.4.99/arch/cris/boot/rescue/
Dhead_v10.S71 ;; sector. Sector size is 65536 bytes in all flashes we use.
77 ;; That is not where we put our downloaded serial boot-code.
123 ;; Since etrax actually starts at address 2 when booting from flash, we
124 ;; put a nop (2 bytes) here first so we dont accidentally skip the di
135 ;; We put a longword of -1 here and if it is not -1, we jump using
136 ;; the value as jump target. Since we can always change 1's to 0's
157 ;; We need to setup the bus registers before we start using the DRAM
160 ;; we now should go through the checksum-table and check the listed
232 ;; (so we can flash LEDs, and so that DTR and others are set)
289 ;; check if we got something on the serial port
[all …]
/linux-3.4.99/arch/mn10300/mm/
Dcache-inv-by-tag.S123 # If we are in writeback mode we check the start and end alignments,
240 # approx every N steps we re-enable the cache and see if there are any
242 # we also break out if we've reached the end of the loop
263 # - we don't bother with delay NOPs as we'll have enough instructions
264 # before we disable interrupts again to give the interrupts a chance
/linux-3.4.99/scripts/genksyms/
Dlex.lex.c_shipped16 /* First, we deal with platform-specific or compiler-specific issues. */
115 * we want to instead treat it as an 8-bit unsigned char, hence the
121 * but we do it the disgusting crufty way forced on us by the ()-less
205 /* Whether we "own" the buffer - i.e., we know we created it,
212 * if we're using stdio for input, then we want to use getc()
213 * instead of fread(), to make sure we stop fetching input after
218 /* Whether we're considered to be at the beginning of a line.
227 /* Whether to try to fill the input buffer when we reach the
237 * then we mark the buffer as YY_EOF_PENDING, to indicate that we
242 * When we actually see the EOF, we change the status to "new"
[all …]
/linux-3.4.99/arch/cris/arch-v10/kernel/
Dentry.S88 ;; we cannot simply test $dccr, because that does not necessarily
89 ;; reflect what mode we'll return into.
111 ;; Since we dont really want to have two epilogues (one for system calls
112 ;; and one for interrupts) we push the contents of BRP instead of IRP in the
116 ;; Since we can't have system calls inside interrupts, it should not matter
119 ;; In r9 we have the wanted syscall number. Arguments come in r10,r11,r12,r13,mof,srp
200 ;; now we have a 4-word SBFS frame which we do not want to restore
201 ;; using RBF since it was not stacked with SBFS. instead we would like to
277 ;; we need to restore $r9 here to contain the wanted syscall, and
299 push $srp ; we keep the old/new PC on the stack
[all …]
/linux-3.4.99/firmware/
DREADME.AddingFirmware8 As we update those drivers to use request_firmware() and keep a clean
9 separation between code and firmware, we put the extracted firmware
19 we were never permitted to include in a GPL'd work, but which we _have_

12345678910>>...28