| Commit message (Expand) | Author | Age | Files | Lines |
|---|
| * | Fix a race between a diagnostic check and session close so test/thread "-S" r... | Michael Cahill | 2012-03-11 | 1 | -2/+3 |
| * | If forced page eviction fails, just give up and hope it works next time. | Michael Cahill | 2012-02-29 | 1 | -17/+14 |
| * | Additional re-structuring of the eviction code. The only real changes | Keith Bostic | 2012-02-27 | 1 | -44/+83 |
| * | Inline __rec_excl_int, it's tiny and it's caller isn't that much bigger. | Keith Bostic | 2012-02-27 | 1 | -31/+17 |
| * | If we have to clear session->excl_next in __wt_rec_evict, let's always | Keith Bostic | 2012-02-27 | 1 | -8/+8 |
| * | Group the exclusive-access functions together (no real change, just shifting | Keith Bostic | 2012-02-27 | 1 | -64/+64 |
| * | Simplify getting exclusive access to pages for eviction. | Michael Cahill | 2012-02-28 | 1 | -86/+37 |
| * | Allow eviction of pages marked split or empty. | Michael Cahill | 2012-02-28 | 1 | -30/+9 |
| * | Don't release references if an error occurs before we acquire references. | Keith Bostic | 2012-02-25 | 1 | -1/+1 |
| * | minor whitespace movement. | Keith Bostic | 2012-01-30 | 1 | -2/+2 |
| * | Free not only the WT_ADDR structure itself, but also the address cookie | Keith Bostic | 2012-01-30 | 1 | -3/+4 |
| * | Minor code shuffling to simplify detecting that a page was split, but | Keith Bostic | 2012-01-30 | 1 | -19/+35 |
| * | When we merge a split page into a parent page, we still have to discard | Keith Bostic | 2012-01-30 | 1 | -1/+17 |
| * | Add -L to the format test program, so you can configure verbose from the | Keith Bostic | 2012-01-27 | 1 | -8/+8 |
| * | Don't evict empty pages (fixing the problem where we would discard a | Keith Bostic | 2012-01-26 | 1 | -13/+11 |
| * | More eviction locking cleanups and fixes. | Michael Cahill | 2012-01-24 | 1 | -93/+40 |
| * | Some cleanup of the hazard code. In particular: | Michael Cahill | 2012-01-23 | 1 | -47/+40 |
| * | Update copyright notices to 2012. | Keith Bostic | 2012-01-22 | 1 | -3/+3 |
| * | I just made a change to the eviction server code to not put both a | Keith Bostic | 2012-01-20 | 1 | -35/+29 |
| * | Fix comments, I removed the requirement the tree be walked in any | Keith Bostic | 2012-01-20 | 1 | -13/+1 |
| * | There's no reason to pass __rec_excl_page() both the WT_REF and the | Keith Bostic | 2012-01-20 | 1 | -5/+8 |
| * | I got the last-page stuff wrong again, and I don't want to debug it | Keith Bostic | 2012-01-20 | 1 | -74/+32 |
| * | If we get an exclusive lock in __rec_excl_page() but still fail, clear the | Keith Bostic | 2012-01-20 | 1 | -0/+2 |
| * | Rename __wt_page_set_modified to __wt_page_modify_set, matching | Keith Bostic | 2012-01-15 | 1 | -1/+2 |
| * | whitespace, spelling | Keith Bostic | 2012-01-12 | 1 | -2/+2 |
| * | Figured it out: if application threads are evicting, we can encounter other | Keith Bostic | 2012-01-12 | 1 | -8/+3 |
| * | Last of the WT_ROW_REF/WT_COL_REF changes: rename the get-an-exclusive-lock | Keith Bostic | 2012-01-12 | 1 | -59/+75 |
| * | Essentially the same changes I just made to the page-discard routines, | Keith Bostic | 2012-01-12 | 1 | -140/+48 |
| * | Now that we've merged the WT_COL_REF/WT_ROW_REF structures into a single | Keith Bostic | 2012-01-12 | 1 | -65/+16 |
| * | Two fixes: first, if the first entry in the page we "review" returns that | Keith Bostic | 2012-01-12 | 1 | -2/+9 |
| * | Merge the WT_ROW_REF/WT_COL_REF structures into WT_REF: this wastes 4B | Keith Bostic | 2012-01-04 | 1 | -40/+38 |
| * | Rework verify to work in the new, block-manager world. | Keith Bostic | 2011-12-23 | 1 | -19/+61 |
| * | lint | Keith Bostic | 2011-12-20 | 1 | -2/+2 |
| * | Use a single mechanism (the tracking list) to schedule blocks to be | Keith Bostic | 2011-12-18 | 1 | -8/+2 |
| * | Clean up from the merge with the current tree. | Keith Bostic | 2011-12-18 | 1 | -3/+3 |
| * | Replace WT_OFF/WT_OFF_RECORD structures on row- and column-store internal | Keith Bostic | 2011-12-17 | 1 | -113/+110 |
| * | Update the connection-wide statistic when failing to get a hazard reference. | Michael Cahill | 2011-12-16 | 1 | -0/+2 |
| * | Maintain a count in the statistics of internal pages evicted during normal op... | Michael Cahill | 2011-12-16 | 1 | -0/+5 |
| * | Move the handling of WT_PAGE_FORCE_EVICT into bt_evict.c and simplify -- deal... | Michael Cahill | 2011-12-13 | 1 | -9/+0 |
| * | Back out the code that doesn't include split pages in the list of eviction ca... | Michael Cahill | 2011-12-13 | 1 | -3/+8 |
| * | Bug fix when tracking the last page locked during hazard checks. | Michael Cahill | 2011-12-12 | 1 | -24/+36 |
| * | Restructure reads to not use serialized functions, reduce the number of calls... | Michael Cahill | 2011-12-12 | 1 | -34/+19 |
| * | Check WT_PAGE_FORCE_EVICT in __wt_rec_evict. | Michael Cahill | 2011-12-12 | 1 | -8/+27 |
| * | Drop the message when eviction finds a subtree locked: that's normal with par... | Michael Cahill | 2011-12-12 | 1 | -6/+0 |
| * | Conflicting subtrees are expected with parallel eviction -- change an error m... | Michael Cahill | 2011-12-09 | 1 | -2/+2 |
| * | Overlapping subtrees are expected with parallel eviction. Use a spinlock to ... | Michael Cahill | 2011-12-09 | 1 | -36/+74 |
| * | Fix s_all warnings. | Michael Cahill | 2011-12-08 | 1 | -3/+3 |
| * | Parallelize eviction by having application threads evict pages while reads ar... | Michael Cahill | 2011-12-05 | 1 | -0/+4 |
| * | Handle errors during read (e.g., during verify). | Michael Cahill | 2011-12-07 | 1 | -0/+2 |
| * | Back out part of the last change: eviction has to return success on | Keith Bostic | 2011-12-06 | 1 | -12/+14 |
