user/sven/linux.git/kernel/livepatch/core.c, branch v4.2.4

Merge branches 'for-4.1/upstream-fixes', 'for-4.2/kaslr' and 'for-4.2/upstream' into for-linus

2015-06-22T14:26:56Z

livepatch: add module locking around kallsyms calls

2015-06-02T20:54:38Z

The list of loaded modules is walked through in module_kallsyms_on_each_symbol (called by kallsyms_on_each_symbol). The module_mutex lock should be acquired to prevent potential corruptions in the list. This was uncovered with new lockdep asserts in module code introduced by the commit 0be964be0d45 ("module: Sanitize RCU usage and locking") in recent next- trees. Signed-off-by: Miroslav Benes Acked-by: Josh Poimboeuf Cc: stable@vger.kernel.org Signed-off-by: Jiri Kosina

livepatch: annotate klp_init() with __init

2015-05-25T15:16:45Z

module_init() function should be marked __init. [jkosina@suse.cz: remove overly verbose changelog] Signed-off-by: Minfei Huang Acked-by: Josh Poimboeuf Signed-off-by: Jiri Kosina

livepatch: introduce patch/func-walking helpers

2015-05-19T21:58:43Z

klp_for_each_object and klp_for_each_func are now used all over the code. One need not think what is the proper condition to check in the for loop now. Signed-off-by: Jiri Slaby Acked-by: Josh Poimboeuf Signed-off-by: Jiri Kosina

livepatch: make kobject in klp_object statically allocated

2015-05-19T21:56:41Z

Make kobj variable (of type struct kobject) statically allocated in klp_object structure. It will allow us to move in the func-object-patch hierarchy through kobject links. The only reason to have it dynamic was to not have empty release callback in the code. However we have empty callbacks for function and patch in the code now, so it is no longer valid and the advantage of static allocation is clear. Signed-off-by: Miroslav Benes Signed-off-by: Jiri Slaby Acked-by: Josh Poimboeuf Signed-off-by: Jiri Kosina

livepatch: Prevent patch inconsistencies if the coming module notifier fails

2015-05-18T09:08:44Z

The previous patches can be applied, once the corresponding module is loaded. In general, the patch will do relocation (if necessary) and obtain/verify function address before we start to enable patch. There are three different situations in which the coming module notifier can fail: 1) relocations are not applied for some reason. In this case kallsyms for module symbol is not called at all. The patch is not applied to the module. If the user disable and enable patch again, there is possible bug in klp_enable_func. If the user specified func->old_addr for some function in the module (and he shouldn't do that, but nevertheless) our warning would not catch it, ftrace will reject to register the handler because of wrong address or will register the handler for wrong address. 2) relocations are applied successfully, but kallsyms lookup fails. In this case func->old_addr can be correct for all previous lookups, 0 for current failed one, and "unspecified" for the rest. If we undergo the same scenario as in 1, the behaviour differs for three cases, but the patch is not enabled anyway. 3) the object is initialized, but klp_enable_object fails in the notifier due to possible ftrace error. Since it is improbable that ftrace would heal itself in the future, we would get those errors everytime the patch is enabled. In order to fix above situations, we can make obj->mod to NULL, if the coming modified notifier fails. Signed-off-by: Minfei Huang Acked-by: Josh Poimboeuf Reviewed-by: Miroslav Benes Signed-off-by: Jiri Kosina

livepatch: match return value to function signature

2015-05-11T21:43:52Z

klp_initialized() should return bool but is actually returning struct kobject * - convert it to a boolean explicitly. Signed-off-by: Nicholas Mc Guire Reviewed-by: Jiri Slaby Signed-off-by: Jiri Kosina

livepatch: x86: make kASLR logic more accurate

2015-04-29T14:51:33Z

We give up old_addr hint from the coming patch module in cases when kernel load base has been randomized (as in such case, the coming module has no idea about the exact randomization offset). We are currently too pessimistic, and give up immediately as soon as CONFIG_RANDOMIZE_BASE is set; this doesn't however directly imply that the load base has actually been randomized. There are config options that disable kASLR (such as hibernation), user could have disabled kaslr on kernel command-line, etc. The loader propagates the information whether kernel has been randomized through bootparams. This allows us to have the condition more accurate. On top of that, it seems unnecessary to give up old_addr hints even if randomization is active. The relocation offset can be computed using kaslr_ofsset(), and therefore old_addr can be adjusted accordingly. Acked-by: Josh Poimboeuf Signed-off-by: Jiri Kosina

Merge branch 'for-4.1/core-noarch' into for-linus

2015-04-13T21:57:20Z

livepatch: Fix subtle race with coming and going modules

2015-03-17T09:31:54Z

There is a notifier that handles live patches for coming and going modules. It takes klp_mutex lock to avoid races with coming and going patches but it does not keep the lock all the time. Therefore the following races are possible: 1. The notifier is called sometime in STATE_MODULE_COMING. The module is visible by find_module() in this state all the time. It means that new patch can be registered and enabled even before the notifier is called. It might create wrong order of stacked patches, see below for an example. 2. New patch could still see the module in the GOING state even after the notifier has been called. It will try to initialize the related object structures but the module could disappear at any time. There will stay mess in the structures. It might even cause an invalid memory access. This patch solves the problem by adding a boolean variable into struct module. The value is true after the coming and before the going handler is called. New patches need to be applied when the value is true and they need to ignore the module when the value is false. Note that we need to know state of all modules on the system. The races are related to new patches. Therefore we do not know what modules will get patched. Also note that we could not simply ignore going modules. The code from the module could be called even in the GOING state until mod->exit() finishes. If we start supporting patches with semantic changes between function calls, we need to apply new patches to any still usable code. See below for an example. Finally note that the patch solves only the situation when a new patch is registered. There are no such problems when the patch is being removed. It does not matter who disable the patch first, whether the normal disable_patch() or the module notifier. There is nothing to do once the patch is disabled. Alternative solutions: ====================== + reject new patches when a patched module is coming or going; this is ugly + wait with adding new patch until the module leaves the COMING and GOING states; this might be dangerous and complicated; we would need to release kgr_lock in the middle of the patch registration to avoid a deadlock with the coming and going handlers; also we might need a waitqueue for each module which seems to be even bigger overhead than the boolean + stop modules from entering COMING and GOING states; wait until modules leave these states when they are already there; looks complicated; we would need to ignore the module that asked to stop the others to avoid a deadlock; also it is unclear what to do when two modules asked to stop others and both are in COMING state (situation when two new patches are applied) + always register/enable new patches and fix up the potential mess (registered patches order) in klp_module_init(); this is nasty and prone to regressions in the future development + add another MODULE_STATE where the kallsyms are visible but the module is not used yet; this looks too complex; the module states are checked on "many" locations Example of patch stacking breakage: =================================== The notifier could _not_ _simply_ ignore already initialized module objects. For example, let's have three patches (P1, P2, P3) for functions a() and b() where a() is from vmcore and b() is from a module M. Something like: a() b() P1 a1() b1() P2 a2() b2() P3 a3() b3(3) If you load the module M after all patches are registered and enabled. The ftrace ops for function a() and b() has listed the functions in this order: ops_a->func_stack -> list(a3,a2,a1) ops_b->func_stack -> list(b3,b2,b1) , so the pointer to b3() is the first and will be used. Then you might have the following scenario. Let's start with state when patches P1 and P2 are registered and enabled but the module M is not loaded. Then ftrace ops for b() does not exist. Then we get into the following race: CPU0 CPU1 load_module(M) complete_formation() mod->state = MODULE_STATE_COMING; mutex_unlock(&module_mutex); klp_register_patch(P3); klp_enable_patch(P3); # STATE 1 klp_module_notify(M) klp_module_notify_coming(P1); klp_module_notify_coming(P2); klp_module_notify_coming(P3); # STATE 2 The ftrace ops for a() and b() then looks: STATE1: ops_a->func_stack -> list(a3,a2,a1); ops_b->func_stack -> list(b3); STATE2: ops_a->func_stack -> list(a3,a2,a1); ops_b->func_stack -> list(b2,b1,b3); therefore, b2() is used for the module but a3() is used for vmcore because they were the last added. Example of the race with going modules: ======================================= CPU0 CPU1 delete_module() #SYSCALL try_stop_module() mod->state = MODULE_STATE_GOING; mutex_unlock(&module_mutex); klp_register_patch() klp_enable_patch() #save place to switch universe b() # from module that is going a() # from core (patched) mod->exit(); Note that the function b() can be called until we call mod->exit(). If we do not apply patch against b() because it is in MODULE_STATE_GOING, it will call patched a() with modified semantic and things might get wrong. [jpoimboe@redhat.com: use one boolean instead of two] Signed-off-by: Petr Mladek Acked-by: Josh Poimboeuf Acked-by: Rusty Russell Signed-off-by: Jiri Kosina