Documentation/livepatch/cumulative-patches.rst

   1 ===================================
   2 Atomic Replace & Cumulative Patches
   3 ===================================
   4
   5 There might be dependencies between livepatches. If multiple patches need
   6 to do different changes to the same function(s) then we need to define
   7 an order in which the patches will be installed. And function implementations
   8 from any newer livepatch must be done on top of the older ones.
   9
  10 This might become a maintenance nightmare. Especially when more patches
  11 modified the same function in different ways.
  12
  13 An elegant solution comes with the feature called "Atomic Replace". It allows
  14 creation of so called "Cumulative Patches". They include all wanted changes
  15 from all older livepatches and completely replace them in one transition.
  16
  17 Usage
  18 -----
  19
  20 The atomic replace can be enabled by setting "replace" flag in struct klp_patch,
  21 for example::
  22
  23         static struct klp_patch patch = {
  24                 .mod = THIS_MODULE,
  25                 .objs = objs,
  26                 .replace = true,
  27         };
  28
  29 All processes are then migrated to use the code only from the new patch.
  30 Once the transition is finished, all older patches are automatically
  31 disabled.
  32
  33 Ftrace handlers are transparently removed from functions that are no
  34 longer modified by the new cumulative patch.
  35
  36 As a result, the livepatch authors might maintain sources only for one
  37 cumulative patch. It helps to keep the patch consistent while adding or
  38 removing various fixes or features.
  39
  40 Users could keep only the last patch installed on the system after
  41 the transition to has finished. It helps to clearly see what code is
  42 actually in use. Also the livepatch might then be seen as a "normal"
  43 module that modifies the kernel behavior. The only difference is that
  44 it can be updated at runtime without breaking its functionality.
  45
  46
  47 Features
  48 --------
  49
  50 The atomic replace allows:
  51
  52   - Atomically revert some functions in a previous patch while
  53     upgrading other functions.
  54
  55   - Remove eventual performance impact caused by core redirection
  56     for functions that are no longer patched.
  57
  58   - Decrease user confusion about dependencies between livepatches.
  59
  60
  61 Limitations:
  62 ------------
  63
  64   - Once the operation finishes, there is no straightforward way
  65     to reverse it and restore the replaced patches atomically.
  66
  67     A good practice is to set .replace flag in any released livepatch.
  68     Then re-adding an older livepatch is equivalent to downgrading
  69     to that patch. This is safe as long as the livepatches do _not_ do
  70     extra modifications in (un)patching callbacks or in the module_init()
  71     or module_exit() functions, see below.
  72
  73     Also note that the replaced patch can be removed and loaded again
  74     only when the transition was not forced.
  75
  76
  77   - Only the (un)patching callbacks from the _new_ cumulative livepatch are
  78     executed. Any callbacks from the replaced patches are ignored.
  79
  80     In other words, the cumulative patch is responsible for doing any actions
  81     that are necessary to properly replace any older patch.
  82
  83     As a result, it might be dangerous to replace newer cumulative patches by
  84     older ones. The old livepatches might not provide the necessary callbacks.
  85
  86     This might be seen as a limitation in some scenarios. But it makes life
  87     easier in many others. Only the new cumulative livepatch knows what
  88     fixes/features are added/removed and what special actions are necessary
  89     for a smooth transition.
  90
  91     In any case, it would be a nightmare to think about the order of
  92     the various callbacks and their interactions if the callbacks from all
  93     enabled patches were called.
  94
  95
  96   - There is no special handling of shadow variables. Livepatch authors
  97     must create their own rules how to pass them from one cumulative
  98     patch to the other. Especially that they should not blindly remove
  99     them in module_exit() functions.
 100
 101     A good practice might be to remove shadow variables in the post-unpatch
 102     callback. It is called only when the livepatch is properly disabled.