view Documentation/keys-request-key.txt @ 897:329ea0ccb344

balloon: try harder to balloon up under memory pressure.

Currently if the balloon driver is unable to increase the guest's
reservation it assumes the failure was due to reaching its full
allocation, gives up on the ballooning operation and records the limit
it reached as the "hard limit". The driver will not try again until
the target is set again (even to the same value).

However it is possible that ballooning has in fact failed due to
memory pressure in the host and therefore it is desirable to keep
attempting to reach the target in case memory becomes available. The
most likely scenario is that some guests are ballooning down while
others are ballooning up and therefore there is temporary memory
pressure while things stabilise. You would not expect a well behaved
toolstack to ask a domain to balloon to more than its allocation nor
would you expect it to deliberately over-commit memory by setting
balloon targets which exceed the total host memory.

This patch drops the concept of a hard limit and causes the balloon
driver to retry increasing the reservation on a timer in the same
manner as when decreasing the reservation.

Also if we partially succeed in increasing the reservation
(i.e. receive less pages than we asked for) then we may as well keep
those pages rather than returning them to Xen.

Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
author Keir Fraser <keir.fraser@citrix.com>
date Fri Jun 05 14:01:20 2009 +0100 (2009-06-05)
parents 831230e53067
line source
1 ===================
3 ===================
5 The key request service is part of the key retention service (refer to
6 Documentation/keys.txt). This document explains more fully how the requesting
7 algorithm works.
9 The process starts by either the kernel requesting a service by calling
10 request_key*():
12 struct key *request_key(const struct key_type *type,
13 const char *description,
14 const char *callout_string);
16 or:
18 struct key *request_key_with_auxdata(const struct key_type *type,
19 const char *description,
20 const char *callout_string,
21 void *aux);
23 Or by userspace invoking the request_key system call:
25 key_serial_t request_key(const char *type,
26 const char *description,
27 const char *callout_info,
28 key_serial_t dest_keyring);
30 The main difference between the access points is that the in-kernel interface
31 does not need to link the key to a keyring to prevent it from being immediately
32 destroyed. The kernel interface returns a pointer directly to the key, and
33 it's up to the caller to destroy the key.
35 The request_key_with_auxdata() call is like the in-kernel request_key() call,
36 except that it permits auxiliary data to be passed to the upcaller (the default
37 is NULL). This is only useful for those key types that define their own upcall
38 mechanism rather than using /sbin/request-key.
40 The userspace interface links the key to a keyring associated with the process
41 to prevent the key from going away, and returns the serial number of the key to
42 the caller.
45 The following example assumes that the key types involved don't define their
46 own upcall mechanisms. If they do, then those should be substituted for the
47 forking and execution of /sbin/request-key.
50 ===========
52 ===========
54 A request proceeds in the following manner:
56 (1) Process A calls request_key() [the userspace syscall calls the kernel
57 interface].
59 (2) request_key() searches the process's subscribed keyrings to see if there's
60 a suitable key there. If there is, it returns the key. If there isn't,
61 and callout_info is not set, an error is returned. Otherwise the process
62 proceeds to the next step.
64 (3) request_key() sees that A doesn't have the desired key yet, so it creates
65 two things:
67 (a) An uninstantiated key U of requested type and description.
69 (b) An authorisation key V that refers to key U and notes that process A
70 is the context in which key U should be instantiated and secured, and
71 from which associated key requests may be satisfied.
73 (4) request_key() then forks and executes /sbin/request-key with a new session
74 keyring that contains a link to auth key V.
76 (5) /sbin/request-key assumes the authority associated with key U.
78 (6) /sbin/request-key execs an appropriate program to perform the actual
79 instantiation.
81 (7) The program may want to access another key from A's context (say a
82 Kerberos TGT key). It just requests the appropriate key, and the keyring
83 search notes that the session keyring has auth key V in its bottom level.
85 This will permit it to then search the keyrings of process A with the
86 UID, GID, groups and security info of process A as if it was process A,
87 and come up with key W.
89 (8) The program then does what it must to get the data with which to
90 instantiate key U, using key W as a reference (perhaps it contacts a
91 Kerberos server using the TGT) and then instantiates key U.
93 (9) Upon instantiating key U, auth key V is automatically revoked so that it
94 may not be used again.
96 (10) The program then exits 0 and request_key() deletes key V and returns key
97 U to the caller.
99 This also extends further. If key W (step 7 above) didn't exist, key W would
100 be created uninstantiated, another auth key (X) would be created (as per step
101 3) and another copy of /sbin/request-key spawned (as per step 4); but the
102 context specified by auth key X will still be process A, as it was in auth key
103 V.
105 This is because process A's keyrings can't simply be attached to
106 /sbin/request-key at the appropriate places because (a) execve will discard two
107 of them, and (b) it requires the same UID/GID/Groups all the way through.
110 ======================
112 ======================
114 Rather than instantiating a key, it is possible for the possessor of an
115 authorisation key to negatively instantiate a key that's under construction.
116 This is a short duration placeholder that causes any attempt at re-requesting
117 the key whilst it exists to fail with error ENOKEY.
119 This is provided to prevent excessive repeated spawning of /sbin/request-key
120 processes for a key that will never be obtainable.
122 Should the /sbin/request-key process exit anything other than 0 or die on a
123 signal, the key under construction will be automatically negatively
124 instantiated for a short amount of time.
127 ====================
129 ====================
131 A search of any particular keyring proceeds in the following fashion:
133 (1) When the key management code searches for a key (keyring_search_aux) it
134 firstly calls key_permission(SEARCH) on the keyring it's starting with,
135 if this denies permission, it doesn't search further.
137 (2) It considers all the non-keyring keys within that keyring and, if any key
138 matches the criteria specified, calls key_permission(SEARCH) on it to see
139 if the key is allowed to be found. If it is, that key is returned; if
140 not, the search continues, and the error code is retained if of higher
141 priority than the one currently set.
143 (3) It then considers all the keyring-type keys in the keyring it's currently
144 searching. It calls key_permission(SEARCH) on each keyring, and if this
145 grants permission, it recurses, executing steps (2) and (3) on that
146 keyring.
148 The process stops immediately a valid key is found with permission granted to
149 use it. Any error from a previous match attempt is discarded and the key is
150 returned.
152 When search_process_keyrings() is invoked, it performs the following searches
153 until one succeeds:
155 (1) If extant, the process's thread keyring is searched.
157 (2) If extant, the process's process keyring is searched.
159 (3) The process's session keyring is searched.
161 (4) If the process has assumed the authority associated with a request_key()
162 authorisation key then:
164 (a) If extant, the calling process's thread keyring is searched.
166 (b) If extant, the calling process's process keyring is searched.
168 (c) The calling process's session keyring is searched.
170 The moment one succeeds, all pending errors are discarded and the found key is
171 returned.
173 Only if all these fail does the whole thing fail with the highest priority
174 error. Note that several errors may have come from LSM.
176 The error priority is:
180 EACCES/EPERM are only returned on a direct search of a specific keyring where
181 the basal keyring does not grant Search permission.