Permalink
Switch branches/tags
Commits on Aug 24, 2011
  1. Test stevedore/LRU/Nuke when the body allocation fails

    bsdphk committed with Tollef Fog Heen Aug 22, 2011
Commits on Aug 23, 2011
  1. SMA->trim() did not return the trimmed space to the pool causing

    bsdphk committed with Tollef Fog Heen Aug 22, 2011
    the SMA to run out of space eventually.
  2. Document 3.0 changes

    Tollef Fog Heen committed Aug 19, 2011
  3. Lock panic buffer before mangling it

    Tollef Fog Heen committed Aug 19, 2011
    If we were panic-ing in multiple threads at the same time, the panic
    buffer would be partially overwritten.  Prevent this with a mutex
  4. An assert to make sure...

    bsdphk committed with Tollef Fog Heen Aug 22, 2011
  5. Adjust c0004[45] limits to work on 32 bit

    Tollef Fog Heen committed Aug 23, 2011
  6. Allow larger bodies in varnishtest

    bsdphk committed with Tollef Fog Heen Aug 22, 2011
  7. The law of unconsidered consequences strikes again:

    bsdphk committed with Tollef Fog Heen Aug 22, 2011
    When we pushed the object allocation into the stevedores for
    -spersistent, we did not add LRU eviction to that allocation
    path.
    
    Then we added the Transient storage as a fallback for objects
    we could not allocate, and they all went there.
    
    Change the way object allocation works as follows:
    
            If VCL set a stevedore hint and it is valid, we stick
            with it, and LRU that stevedore attempting to make space.
    
            If no valid hint is given, try all stevedores in turn,
            then LRU one of them to make space.
    
    Fixes	#953
  8. Properly clean up if we bail out of cnt_error because we cannot allocate

    bsdphk committed with Tollef Fog Heen Aug 22, 2011
    an object.
    
    Fixes	#985
  9. Add a new paramter "nuke_limit" which controls how many objects we

    bsdphk committed with Tollef Fog Heen Aug 22, 2011
    are willing to evict per storage allocation attempt.
    
    Previously this was hardcoded at 50, make it 10 instead.
  10. Catch up with Tollefs Vmod_Varnish_ABI change.

    bsdphk committed with Tollef Fog Heen Aug 22, 2011
  11. Update sphinx Makefile to contain the right set of files

    Tollef Fog Heen committed Aug 23, 2011
Commits on Aug 19, 2011
  1. Check the ABI of VMODs.

    Tollef Fog Heen committed Aug 19, 2011
    The ABI we give vmods consist of the Varnish version number and the
    git commit ID meaning we can break ABI at will.
    
    Output a warning if we can't determine git commit ID
Commits on Aug 17, 2011
  1. If the backend used chunked encoding and sent junk after the gzip

    bsdphk committed with Tollef Fog Heen Aug 17, 2011
    data, the thread would go into a spin.
    
    Fixes	#942
  2. added () to EXAMPLES for std.log

    Per Buer committed with Tollef Fog Heen Aug 17, 2011
  3. Add a "sendhex" verb

    bsdphk committed with Tollef Fog Heen Aug 17, 2011
  4. std.log missing ()

    Per Buer committed with Tollef Fog Heen Aug 17, 2011
  5. hit_for_pass in vcl_fetch

    Per Buer committed with Tollef Fog Heen Aug 17, 2011
  6. Cleare do_stream on all esi objects, including included objects.

    bsdphk committed with Tollef Fog Heen Aug 17, 2011
    Found & Fixed by:	Martin
    
    Fixes	#978
  7. restart in vcl_deliver{} would crash in vcl_fetch{} due to missing cl…

    bsdphk committed with Tollef Fog Heen Aug 17, 2011
    …eanup.
    
    Found & Fixed by:	Martin
    
    Fixes	#979
  8. We can not use a lenght based response when

    bsdphk committed with Tollef Fog Heen Aug 17, 2011
    	we transform (gzip/gunzip)
    	we stream and backend didn't send c-l.
    
    Fixes	#980
    
    Testcase by:	Martin
  9. Implement a consistent retry policy in the random/client/hash director:

    bsdphk committed with Tollef Fog Heen Aug 15, 2011
    If the first (policy-chosen) backend fails to get us a connection,
    retry a random backend (still according to their weight) until
    retries are exhausted.
    
    Kristian sent a proof of concept patch, I just cleaned it up and made
    it compile.
    
    Thanks to:	Kristian
    
    Fixes	#977
  10. My recent 'discover correct -spersistent mmap address' fix broke

    bsdphk committed with Tollef Fog Heen Aug 15, 2011
    32 bit systems.
    
    Spotted by:	Kristian
  11. the basics on persistent storage.

    Per Buer committed with Tollef Fog Heen Aug 15, 2011
    Explained the semantics quickly so people understand why not all objects survive a crash
  12. document the transient stevedore

    Tollef Fog Heen committed Aug 17, 2011
    more on transient.
    didn't get it rigth the first time.
    Thanks to scoof for pointing it out.
    (cherry picked from commit 6a59dac
    and 0b23a0f)
  13. more on bans, the lurker

    Per Buer committed with Tollef Fog Heen Aug 15, 2011
  14. Add emphasis to explain when bans are checked against

    Per Buer committed with Tollef Fog Heen Aug 15, 2011
  15. Add (disabled) test-case for #977

    KristianLyng committed with Tollef Fog Heen Aug 15, 2011
  16. Prompted by HTTPbis and FlexeLint in unison, go over the nhttp and

    bsdphk committed with Tollef Fog Heen Aug 15, 2011
    response code stuff and fully incorporate the fact that they are both
    16 unsigned now.
    
    No functional changes.
  17. We always allocate an object entirely inside one stevedore now.

    bsdphk committed with Tollef Fog Heen Aug 15, 2011
  18. Constify

    bsdphk committed with Tollef Fog Heen Aug 15, 2011
  19. Try to read the silo signature to find the correct address to map

    bsdphk committed with Tollef Fog Heen Aug 11, 2011
    the silo into VM.  If this fails or we get garbage, the silo will
    be cleared.
    
    Fixes	#962
  20. ifdef the OFOF development tool

    bsdphk committed with Tollef Fog Heen Aug 15, 2011
  21. Clamp rather than overflow on child indexes when we get to the

    bsdphk committed with Tollef Fog Heen Aug 17, 2011
    end of the UINT_MAX items we can support.
    
    Found by adding a lot of asserts and brute force testing, both
    of which I have left in.
    
    Fixes       #967
    
    May also be relevant to #827
    (cherry picked from commit 726d93f
    and a6ddafc)