Fiasco.OC: introduce Cap_index (fixes #149, #112)
This commit introduces a Cap_index class for Fiasco.OC's capabilities.
A Cap_index is a combination of the global capability id, that is used by Genode
to correctly identify a kernel-object, and a corresponding entry in a
protection-domain's (kernel-)capability-space. The cap-indices are non-copyable,
unique objects, that are held in a Cap_map. The Cap_map is used to re-find
capabilities already present in the protection-domain, when a capability is
received via IPC. The retrieval of capabilities effectively fixes issue #112,
meaning the waste of capability-space entries.
Because Cap_index objects are non-copyable (their address indicates the position
in the capability-space of the pd), they are inappropriate to use as
Native_capability. Therefore, Native_capability is implemented as a reference
to Cap_index objects. This design seems to be a good pre-condition to implement
smart-pointers for entries in the capability-space, and thereby closing existing
leaks (please refer to issue #32).
Cap_index, Cap_map, and the allocator for Cap_index objects are designed in a way,
that it should be relatively easy to apply the same concept to NOVA also. By now,
these classes are located in the `base-foc` repository, but they intentionally
contain no Fiasco.OC specific elements.
The previously explained changes had extensive impact on the whole Fiasco.OC
platform implementation, due to various dependencies. The following things had to
be changed:
* The Thread object's startup and destruction routine is re-arranged, to
enable another thread (that calls the Thread destructor) gaining the
capability id of the thread's gate to remove it from the Cap_map, the
thread's UTCB had to be made available to the caller, because there
is the current location of that id. After having the UTCB available
in the Thread object for that reason, the whole thread bootstrapping
could be simplified.
* In the course of changing the Native_capability's semantic, a new Cap_mapping
class was introduced in core, that facilitates the establishment and
destruction of capability mappings between core and it's client's, especially
mappings related to Platform_thread and Platform_task, that are relevant to
task and thread creation and destruction. Thereby, the destruction of
threads had to be reworked, which effectively removed a bug (issue #149)
where some threads weren't destroyed properly.
* In the quick fix for issue #112, something similar to the Cap_map was
introduced available in all processes. Moreover, some kind of a capability
map already existed in core, to handle cap-session request properly. The
introduction of the Cap_map unified both structures, so that the
cap-session component code in core had to be reworked too.
* The platform initialization code had to be changed sligthly due to the
changes in Native_capability
* The vcpu initialization in the L4Linux support library had to be adapted
according to the already mentioned changes in the Thread object's bootstrap
code.
2012-03-15 11:41:24 +00:00
|
|
|
/*
|
|
|
|
* \brief Mapping of Genode's capability names to kernel capabilities.
|
|
|
|
* \author Stefan Kalkowski
|
|
|
|
* \date 2012-02-16
|
|
|
|
*
|
|
|
|
* Although kernels like Fiasco.OC and NOVA provide capability mechanisms
|
|
|
|
* to us, which should prevent the usage of global names, there is no
|
|
|
|
* efficient way to retrieve a capability a process owns, when it gets the
|
|
|
|
* same capability delivered again via IPC from another thread. But in some
|
|
|
|
* use-cases in Genode this is essential (e.g. parent getting a close-session
|
|
|
|
* request from a child). Moreover, we waste a lot of slots in the
|
|
|
|
* capability-space of the process for one and the same kernel-object.
|
|
|
|
* That's why we introduce a map of Genode's global capability names to the
|
|
|
|
* process-local addresses in the capability-space.
|
|
|
|
*
|
|
|
|
* TODO: should be moved to the generic part of the framework, and used by
|
|
|
|
* NOVA too.
|
|
|
|
*/
|
|
|
|
|
|
|
|
/*
|
2013-01-10 20:44:47 +00:00
|
|
|
* Copyright (C) 2012-2013 Genode Labs GmbH
|
Fiasco.OC: introduce Cap_index (fixes #149, #112)
This commit introduces a Cap_index class for Fiasco.OC's capabilities.
A Cap_index is a combination of the global capability id, that is used by Genode
to correctly identify a kernel-object, and a corresponding entry in a
protection-domain's (kernel-)capability-space. The cap-indices are non-copyable,
unique objects, that are held in a Cap_map. The Cap_map is used to re-find
capabilities already present in the protection-domain, when a capability is
received via IPC. The retrieval of capabilities effectively fixes issue #112,
meaning the waste of capability-space entries.
Because Cap_index objects are non-copyable (their address indicates the position
in the capability-space of the pd), they are inappropriate to use as
Native_capability. Therefore, Native_capability is implemented as a reference
to Cap_index objects. This design seems to be a good pre-condition to implement
smart-pointers for entries in the capability-space, and thereby closing existing
leaks (please refer to issue #32).
Cap_index, Cap_map, and the allocator for Cap_index objects are designed in a way,
that it should be relatively easy to apply the same concept to NOVA also. By now,
these classes are located in the `base-foc` repository, but they intentionally
contain no Fiasco.OC specific elements.
The previously explained changes had extensive impact on the whole Fiasco.OC
platform implementation, due to various dependencies. The following things had to
be changed:
* The Thread object's startup and destruction routine is re-arranged, to
enable another thread (that calls the Thread destructor) gaining the
capability id of the thread's gate to remove it from the Cap_map, the
thread's UTCB had to be made available to the caller, because there
is the current location of that id. After having the UTCB available
in the Thread object for that reason, the whole thread bootstrapping
could be simplified.
* In the course of changing the Native_capability's semantic, a new Cap_mapping
class was introduced in core, that facilitates the establishment and
destruction of capability mappings between core and it's client's, especially
mappings related to Platform_thread and Platform_task, that are relevant to
task and thread creation and destruction. Thereby, the destruction of
threads had to be reworked, which effectively removed a bug (issue #149)
where some threads weren't destroyed properly.
* In the quick fix for issue #112, something similar to the Cap_map was
introduced available in all processes. Moreover, some kind of a capability
map already existed in core, to handle cap-session request properly. The
introduction of the Cap_map unified both structures, so that the
cap-session component code in core had to be reworked too.
* The platform initialization code had to be changed sligthly due to the
changes in Native_capability
* The vcpu initialization in the L4Linux support library had to be adapted
according to the already mentioned changes in the Thread object's bootstrap
code.
2012-03-15 11:41:24 +00:00
|
|
|
*
|
|
|
|
* This file is part of the Genode OS framework, which is distributed
|
|
|
|
* under the terms of the GNU General Public License version 2.
|
|
|
|
*/
|
|
|
|
|
|
|
|
#ifndef _INCLUDE__BASE__CAP_MAP_H_
|
|
|
|
#define _INCLUDE__BASE__CAP_MAP_H_
|
|
|
|
|
|
|
|
/* Genode includes */
|
|
|
|
#include <base/exception.h>
|
|
|
|
#include <base/stdint.h>
|
|
|
|
#include <base/lock_guard.h>
|
|
|
|
#include <util/avl_tree.h>
|
|
|
|
#include <util/noncopyable.h>
|
|
|
|
#include <util/string.h>
|
|
|
|
|
|
|
|
namespace Genode
|
|
|
|
{
|
|
|
|
/**
|
|
|
|
* A Cap_index represents a single mapping of the global capability id
|
|
|
|
* to the address in the local capability space.
|
|
|
|
*
|
|
|
|
* The address of the Cap_index determines the location in the
|
|
|
|
* (platform-specific) capability space of the process. Therefore it
|
|
|
|
* shouldn't be copied around, but only referenced by
|
|
|
|
* e.g. Native_capability.
|
|
|
|
*/
|
|
|
|
class Cap_index : public Avl_node<Cap_index>,
|
|
|
|
Noncopyable
|
|
|
|
{
|
|
|
|
private:
|
|
|
|
|
|
|
|
enum { INVALID_ID = -1, UNUSED = 0 };
|
|
|
|
|
2012-04-17 13:09:27 +00:00
|
|
|
uint8_t _ref_cnt; /* reference counter */
|
|
|
|
uint16_t _id; /* global capability id */
|
Fiasco.OC: introduce Cap_index (fixes #149, #112)
This commit introduces a Cap_index class for Fiasco.OC's capabilities.
A Cap_index is a combination of the global capability id, that is used by Genode
to correctly identify a kernel-object, and a corresponding entry in a
protection-domain's (kernel-)capability-space. The cap-indices are non-copyable,
unique objects, that are held in a Cap_map. The Cap_map is used to re-find
capabilities already present in the protection-domain, when a capability is
received via IPC. The retrieval of capabilities effectively fixes issue #112,
meaning the waste of capability-space entries.
Because Cap_index objects are non-copyable (their address indicates the position
in the capability-space of the pd), they are inappropriate to use as
Native_capability. Therefore, Native_capability is implemented as a reference
to Cap_index objects. This design seems to be a good pre-condition to implement
smart-pointers for entries in the capability-space, and thereby closing existing
leaks (please refer to issue #32).
Cap_index, Cap_map, and the allocator for Cap_index objects are designed in a way,
that it should be relatively easy to apply the same concept to NOVA also. By now,
these classes are located in the `base-foc` repository, but they intentionally
contain no Fiasco.OC specific elements.
The previously explained changes had extensive impact on the whole Fiasco.OC
platform implementation, due to various dependencies. The following things had to
be changed:
* The Thread object's startup and destruction routine is re-arranged, to
enable another thread (that calls the Thread destructor) gaining the
capability id of the thread's gate to remove it from the Cap_map, the
thread's UTCB had to be made available to the caller, because there
is the current location of that id. After having the UTCB available
in the Thread object for that reason, the whole thread bootstrapping
could be simplified.
* In the course of changing the Native_capability's semantic, a new Cap_mapping
class was introduced in core, that facilitates the establishment and
destruction of capability mappings between core and it's client's, especially
mappings related to Platform_thread and Platform_task, that are relevant to
task and thread creation and destruction. Thereby, the destruction of
threads had to be reworked, which effectively removed a bug (issue #149)
where some threads weren't destroyed properly.
* In the quick fix for issue #112, something similar to the Cap_map was
introduced available in all processes. Moreover, some kind of a capability
map already existed in core, to handle cap-session request properly. The
introduction of the Cap_map unified both structures, so that the
cap-session component code in core had to be reworked too.
* The platform initialization code had to be changed sligthly due to the
changes in Native_capability
* The vcpu initialization in the L4Linux support library had to be adapted
according to the already mentioned changes in the Thread object's bootstrap
code.
2012-03-15 11:41:24 +00:00
|
|
|
|
|
|
|
public:
|
|
|
|
|
2012-04-17 13:09:27 +00:00
|
|
|
Cap_index() : _ref_cnt(0), _id(INVALID_ID) { }
|
Fiasco.OC: introduce Cap_index (fixes #149, #112)
This commit introduces a Cap_index class for Fiasco.OC's capabilities.
A Cap_index is a combination of the global capability id, that is used by Genode
to correctly identify a kernel-object, and a corresponding entry in a
protection-domain's (kernel-)capability-space. The cap-indices are non-copyable,
unique objects, that are held in a Cap_map. The Cap_map is used to re-find
capabilities already present in the protection-domain, when a capability is
received via IPC. The retrieval of capabilities effectively fixes issue #112,
meaning the waste of capability-space entries.
Because Cap_index objects are non-copyable (their address indicates the position
in the capability-space of the pd), they are inappropriate to use as
Native_capability. Therefore, Native_capability is implemented as a reference
to Cap_index objects. This design seems to be a good pre-condition to implement
smart-pointers for entries in the capability-space, and thereby closing existing
leaks (please refer to issue #32).
Cap_index, Cap_map, and the allocator for Cap_index objects are designed in a way,
that it should be relatively easy to apply the same concept to NOVA also. By now,
these classes are located in the `base-foc` repository, but they intentionally
contain no Fiasco.OC specific elements.
The previously explained changes had extensive impact on the whole Fiasco.OC
platform implementation, due to various dependencies. The following things had to
be changed:
* The Thread object's startup and destruction routine is re-arranged, to
enable another thread (that calls the Thread destructor) gaining the
capability id of the thread's gate to remove it from the Cap_map, the
thread's UTCB had to be made available to the caller, because there
is the current location of that id. After having the UTCB available
in the Thread object for that reason, the whole thread bootstrapping
could be simplified.
* In the course of changing the Native_capability's semantic, a new Cap_mapping
class was introduced in core, that facilitates the establishment and
destruction of capability mappings between core and it's client's, especially
mappings related to Platform_thread and Platform_task, that are relevant to
task and thread creation and destruction. Thereby, the destruction of
threads had to be reworked, which effectively removed a bug (issue #149)
where some threads weren't destroyed properly.
* In the quick fix for issue #112, something similar to the Cap_map was
introduced available in all processes. Moreover, some kind of a capability
map already existed in core, to handle cap-session request properly. The
introduction of the Cap_map unified both structures, so that the
cap-session component code in core had to be reworked too.
* The platform initialization code had to be changed sligthly due to the
changes in Native_capability
* The vcpu initialization in the L4Linux support library had to be adapted
according to the already mentioned changes in the Thread object's bootstrap
code.
2012-03-15 11:41:24 +00:00
|
|
|
|
|
|
|
bool valid() const { return _id != INVALID_ID; }
|
|
|
|
bool used() const { return _id != UNUSED; }
|
|
|
|
uint16_t id() const { return _id; }
|
|
|
|
void id(uint16_t id) { _id = id; }
|
2012-05-29 09:14:53 +00:00
|
|
|
uint8_t inc();
|
|
|
|
uint8_t dec();
|
Fiasco.OC: introduce Cap_index (fixes #149, #112)
This commit introduces a Cap_index class for Fiasco.OC's capabilities.
A Cap_index is a combination of the global capability id, that is used by Genode
to correctly identify a kernel-object, and a corresponding entry in a
protection-domain's (kernel-)capability-space. The cap-indices are non-copyable,
unique objects, that are held in a Cap_map. The Cap_map is used to re-find
capabilities already present in the protection-domain, when a capability is
received via IPC. The retrieval of capabilities effectively fixes issue #112,
meaning the waste of capability-space entries.
Because Cap_index objects are non-copyable (their address indicates the position
in the capability-space of the pd), they are inappropriate to use as
Native_capability. Therefore, Native_capability is implemented as a reference
to Cap_index objects. This design seems to be a good pre-condition to implement
smart-pointers for entries in the capability-space, and thereby closing existing
leaks (please refer to issue #32).
Cap_index, Cap_map, and the allocator for Cap_index objects are designed in a way,
that it should be relatively easy to apply the same concept to NOVA also. By now,
these classes are located in the `base-foc` repository, but they intentionally
contain no Fiasco.OC specific elements.
The previously explained changes had extensive impact on the whole Fiasco.OC
platform implementation, due to various dependencies. The following things had to
be changed:
* The Thread object's startup and destruction routine is re-arranged, to
enable another thread (that calls the Thread destructor) gaining the
capability id of the thread's gate to remove it from the Cap_map, the
thread's UTCB had to be made available to the caller, because there
is the current location of that id. After having the UTCB available
in the Thread object for that reason, the whole thread bootstrapping
could be simplified.
* In the course of changing the Native_capability's semantic, a new Cap_mapping
class was introduced in core, that facilitates the establishment and
destruction of capability mappings between core and it's client's, especially
mappings related to Platform_thread and Platform_task, that are relevant to
task and thread creation and destruction. Thereby, the destruction of
threads had to be reworked, which effectively removed a bug (issue #149)
where some threads weren't destroyed properly.
* In the quick fix for issue #112, something similar to the Cap_map was
introduced available in all processes. Moreover, some kind of a capability
map already existed in core, to handle cap-session request properly. The
introduction of the Cap_map unified both structures, so that the
cap-session component code in core had to be reworked too.
* The platform initialization code had to be changed sligthly due to the
changes in Native_capability
* The vcpu initialization in the L4Linux support library had to be adapted
according to the already mentioned changes in the Thread object's bootstrap
code.
2012-03-15 11:41:24 +00:00
|
|
|
addr_t kcap();
|
|
|
|
|
|
|
|
void* operator new (size_t size, Cap_index* idx) { return idx; }
|
|
|
|
void operator delete (void* idx) { memset(idx, 0, sizeof(Cap_index)); }
|
|
|
|
|
|
|
|
|
|
|
|
/************************
|
|
|
|
** Avl node interface **
|
|
|
|
************************/
|
|
|
|
|
|
|
|
bool higher(Cap_index *n);
|
|
|
|
Cap_index *find_by_id(uint16_t id);
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
* Allocator for Cap_index objects.
|
|
|
|
*
|
|
|
|
* This is just an interface, as the real allocator has to be
|
|
|
|
* implemented platform-specific.
|
|
|
|
*/
|
|
|
|
class Cap_index_allocator: Noncopyable
|
|
|
|
{
|
|
|
|
public:
|
|
|
|
|
|
|
|
class Index_out_of_bounds : public Exception { };
|
|
|
|
class Region_conflict : public Exception { };
|
|
|
|
|
|
|
|
virtual ~Cap_index_allocator() {}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* Allocate a range of Cap_index objects
|
|
|
|
*
|
|
|
|
* \param cnt number of objects to allocate
|
|
|
|
* \return pointer to first allocated object, or zero if
|
|
|
|
* out of entries
|
|
|
|
*/
|
Fiasco.OC: several capability ref-counter fixes.
This commit fixes several issues that were triggered e.g. by the
'noux_tool_chain' run-script (fix #208 in part). The following problems
are tackled:
* Don't reference count capability selectors within a task that are actually
controlled by core (all beneath 0x200000), because it's undecideable which
"version" of a capability selector we currently use, e.g. a thread gets
destroyed and a new one gets created immediately some other thread might
have a Native_capability pointing to the already destroyed thread's gate
capability-slot, that is now a new valid one (the one of the new thread)
* In core we cannot invalidate and remove a capability from the so called
Cap_map before each reference to it is destroyed, so don't do this in
Cap_session_component::free, but only reference-decrement within there,
the actual removal can only be done in Cap_map::remove. Because core also
has to invalidate a capability to be removed in all protection-domains
we have to implement a core specific Cap_map::remove method
* When a capability gets inserted into the Cap_map, and we detect an old
invalid entry with the dame id in the tree, don't just overmap that
invalid entry (as there exist remaining references to it), but just remove
it from the tree and allocate an new entry.
* Use the Cap_session_component interface to free a Pager_object when it
gets dissolved, as its also used for allocation
2012-08-29 13:52:18 +00:00
|
|
|
virtual Cap_index* alloc_range(size_t cnt) = 0;
|
Fiasco.OC: introduce Cap_index (fixes #149, #112)
This commit introduces a Cap_index class for Fiasco.OC's capabilities.
A Cap_index is a combination of the global capability id, that is used by Genode
to correctly identify a kernel-object, and a corresponding entry in a
protection-domain's (kernel-)capability-space. The cap-indices are non-copyable,
unique objects, that are held in a Cap_map. The Cap_map is used to re-find
capabilities already present in the protection-domain, when a capability is
received via IPC. The retrieval of capabilities effectively fixes issue #112,
meaning the waste of capability-space entries.
Because Cap_index objects are non-copyable (their address indicates the position
in the capability-space of the pd), they are inappropriate to use as
Native_capability. Therefore, Native_capability is implemented as a reference
to Cap_index objects. This design seems to be a good pre-condition to implement
smart-pointers for entries in the capability-space, and thereby closing existing
leaks (please refer to issue #32).
Cap_index, Cap_map, and the allocator for Cap_index objects are designed in a way,
that it should be relatively easy to apply the same concept to NOVA also. By now,
these classes are located in the `base-foc` repository, but they intentionally
contain no Fiasco.OC specific elements.
The previously explained changes had extensive impact on the whole Fiasco.OC
platform implementation, due to various dependencies. The following things had to
be changed:
* The Thread object's startup and destruction routine is re-arranged, to
enable another thread (that calls the Thread destructor) gaining the
capability id of the thread's gate to remove it from the Cap_map, the
thread's UTCB had to be made available to the caller, because there
is the current location of that id. After having the UTCB available
in the Thread object for that reason, the whole thread bootstrapping
could be simplified.
* In the course of changing the Native_capability's semantic, a new Cap_mapping
class was introduced in core, that facilitates the establishment and
destruction of capability mappings between core and it's client's, especially
mappings related to Platform_thread and Platform_task, that are relevant to
task and thread creation and destruction. Thereby, the destruction of
threads had to be reworked, which effectively removed a bug (issue #149)
where some threads weren't destroyed properly.
* In the quick fix for issue #112, something similar to the Cap_map was
introduced available in all processes. Moreover, some kind of a capability
map already existed in core, to handle cap-session request properly. The
introduction of the Cap_map unified both structures, so that the
cap-session component code in core had to be reworked too.
* The platform initialization code had to be changed sligthly due to the
changes in Native_capability
* The vcpu initialization in the L4Linux support library had to be adapted
according to the already mentioned changes in the Thread object's bootstrap
code.
2012-03-15 11:41:24 +00:00
|
|
|
|
|
|
|
/**
|
Fiasco.OC: several capability ref-counter fixes.
This commit fixes several issues that were triggered e.g. by the
'noux_tool_chain' run-script (fix #208 in part). The following problems
are tackled:
* Don't reference count capability selectors within a task that are actually
controlled by core (all beneath 0x200000), because it's undecideable which
"version" of a capability selector we currently use, e.g. a thread gets
destroyed and a new one gets created immediately some other thread might
have a Native_capability pointing to the already destroyed thread's gate
capability-slot, that is now a new valid one (the one of the new thread)
* In core we cannot invalidate and remove a capability from the so called
Cap_map before each reference to it is destroyed, so don't do this in
Cap_session_component::free, but only reference-decrement within there,
the actual removal can only be done in Cap_map::remove. Because core also
has to invalidate a capability to be removed in all protection-domains
we have to implement a core specific Cap_map::remove method
* When a capability gets inserted into the Cap_map, and we detect an old
invalid entry with the dame id in the tree, don't just overmap that
invalid entry (as there exist remaining references to it), but just remove
it from the tree and allocate an new entry.
* Use the Cap_session_component interface to free a Pager_object when it
gets dissolved, as its also used for allocation
2012-08-29 13:52:18 +00:00
|
|
|
* Allocate a Cap_index object at a specific
|
Fiasco.OC: introduce Cap_index (fixes #149, #112)
This commit introduces a Cap_index class for Fiasco.OC's capabilities.
A Cap_index is a combination of the global capability id, that is used by Genode
to correctly identify a kernel-object, and a corresponding entry in a
protection-domain's (kernel-)capability-space. The cap-indices are non-copyable,
unique objects, that are held in a Cap_map. The Cap_map is used to re-find
capabilities already present in the protection-domain, when a capability is
received via IPC. The retrieval of capabilities effectively fixes issue #112,
meaning the waste of capability-space entries.
Because Cap_index objects are non-copyable (their address indicates the position
in the capability-space of the pd), they are inappropriate to use as
Native_capability. Therefore, Native_capability is implemented as a reference
to Cap_index objects. This design seems to be a good pre-condition to implement
smart-pointers for entries in the capability-space, and thereby closing existing
leaks (please refer to issue #32).
Cap_index, Cap_map, and the allocator for Cap_index objects are designed in a way,
that it should be relatively easy to apply the same concept to NOVA also. By now,
these classes are located in the `base-foc` repository, but they intentionally
contain no Fiasco.OC specific elements.
The previously explained changes had extensive impact on the whole Fiasco.OC
platform implementation, due to various dependencies. The following things had to
be changed:
* The Thread object's startup and destruction routine is re-arranged, to
enable another thread (that calls the Thread destructor) gaining the
capability id of the thread's gate to remove it from the Cap_map, the
thread's UTCB had to be made available to the caller, because there
is the current location of that id. After having the UTCB available
in the Thread object for that reason, the whole thread bootstrapping
could be simplified.
* In the course of changing the Native_capability's semantic, a new Cap_mapping
class was introduced in core, that facilitates the establishment and
destruction of capability mappings between core and it's client's, especially
mappings related to Platform_thread and Platform_task, that are relevant to
task and thread creation and destruction. Thereby, the destruction of
threads had to be reworked, which effectively removed a bug (issue #149)
where some threads weren't destroyed properly.
* In the quick fix for issue #112, something similar to the Cap_map was
introduced available in all processes. Moreover, some kind of a capability
map already existed in core, to handle cap-session request properly. The
introduction of the Cap_map unified both structures, so that the
cap-session component code in core had to be reworked too.
* The platform initialization code had to be changed sligthly due to the
changes in Native_capability
* The vcpu initialization in the L4Linux support library had to be adapted
according to the already mentioned changes in the Thread object's bootstrap
code.
2012-03-15 11:41:24 +00:00
|
|
|
* point in the capability space
|
|
|
|
*
|
|
|
|
* \param kcap address in capability space
|
|
|
|
* \throw Index_out_of_bounds if address is out of scope
|
Fiasco.OC: several capability ref-counter fixes.
This commit fixes several issues that were triggered e.g. by the
'noux_tool_chain' run-script (fix #208 in part). The following problems
are tackled:
* Don't reference count capability selectors within a task that are actually
controlled by core (all beneath 0x200000), because it's undecideable which
"version" of a capability selector we currently use, e.g. a thread gets
destroyed and a new one gets created immediately some other thread might
have a Native_capability pointing to the already destroyed thread's gate
capability-slot, that is now a new valid one (the one of the new thread)
* In core we cannot invalidate and remove a capability from the so called
Cap_map before each reference to it is destroyed, so don't do this in
Cap_session_component::free, but only reference-decrement within there,
the actual removal can only be done in Cap_map::remove. Because core also
has to invalidate a capability to be removed in all protection-domains
we have to implement a core specific Cap_map::remove method
* When a capability gets inserted into the Cap_map, and we detect an old
invalid entry with the dame id in the tree, don't just overmap that
invalid entry (as there exist remaining references to it), but just remove
it from the tree and allocate an new entry.
* Use the Cap_session_component interface to free a Pager_object when it
gets dissolved, as its also used for allocation
2012-08-29 13:52:18 +00:00
|
|
|
* \return pointer to allocated object
|
Fiasco.OC: introduce Cap_index (fixes #149, #112)
This commit introduces a Cap_index class for Fiasco.OC's capabilities.
A Cap_index is a combination of the global capability id, that is used by Genode
to correctly identify a kernel-object, and a corresponding entry in a
protection-domain's (kernel-)capability-space. The cap-indices are non-copyable,
unique objects, that are held in a Cap_map. The Cap_map is used to re-find
capabilities already present in the protection-domain, when a capability is
received via IPC. The retrieval of capabilities effectively fixes issue #112,
meaning the waste of capability-space entries.
Because Cap_index objects are non-copyable (their address indicates the position
in the capability-space of the pd), they are inappropriate to use as
Native_capability. Therefore, Native_capability is implemented as a reference
to Cap_index objects. This design seems to be a good pre-condition to implement
smart-pointers for entries in the capability-space, and thereby closing existing
leaks (please refer to issue #32).
Cap_index, Cap_map, and the allocator for Cap_index objects are designed in a way,
that it should be relatively easy to apply the same concept to NOVA also. By now,
these classes are located in the `base-foc` repository, but they intentionally
contain no Fiasco.OC specific elements.
The previously explained changes had extensive impact on the whole Fiasco.OC
platform implementation, due to various dependencies. The following things had to
be changed:
* The Thread object's startup and destruction routine is re-arranged, to
enable another thread (that calls the Thread destructor) gaining the
capability id of the thread's gate to remove it from the Cap_map, the
thread's UTCB had to be made available to the caller, because there
is the current location of that id. After having the UTCB available
in the Thread object for that reason, the whole thread bootstrapping
could be simplified.
* In the course of changing the Native_capability's semantic, a new Cap_mapping
class was introduced in core, that facilitates the establishment and
destruction of capability mappings between core and it's client's, especially
mappings related to Platform_thread and Platform_task, that are relevant to
task and thread creation and destruction. Thereby, the destruction of
threads had to be reworked, which effectively removed a bug (issue #149)
where some threads weren't destroyed properly.
* In the quick fix for issue #112, something similar to the Cap_map was
introduced available in all processes. Moreover, some kind of a capability
map already existed in core, to handle cap-session request properly. The
introduction of the Cap_map unified both structures, so that the
cap-session component code in core had to be reworked too.
* The platform initialization code had to be changed sligthly due to the
changes in Native_capability
* The vcpu initialization in the L4Linux support library had to be adapted
according to the already mentioned changes in the Thread object's bootstrap
code.
2012-03-15 11:41:24 +00:00
|
|
|
*/
|
Fiasco.OC: several capability ref-counter fixes.
This commit fixes several issues that were triggered e.g. by the
'noux_tool_chain' run-script (fix #208 in part). The following problems
are tackled:
* Don't reference count capability selectors within a task that are actually
controlled by core (all beneath 0x200000), because it's undecideable which
"version" of a capability selector we currently use, e.g. a thread gets
destroyed and a new one gets created immediately some other thread might
have a Native_capability pointing to the already destroyed thread's gate
capability-slot, that is now a new valid one (the one of the new thread)
* In core we cannot invalidate and remove a capability from the so called
Cap_map before each reference to it is destroyed, so don't do this in
Cap_session_component::free, but only reference-decrement within there,
the actual removal can only be done in Cap_map::remove. Because core also
has to invalidate a capability to be removed in all protection-domains
we have to implement a core specific Cap_map::remove method
* When a capability gets inserted into the Cap_map, and we detect an old
invalid entry with the dame id in the tree, don't just overmap that
invalid entry (as there exist remaining references to it), but just remove
it from the tree and allocate an new entry.
* Use the Cap_session_component interface to free a Pager_object when it
gets dissolved, as its also used for allocation
2012-08-29 13:52:18 +00:00
|
|
|
virtual Cap_index* alloc(addr_t kcap) = 0;
|
Fiasco.OC: introduce Cap_index (fixes #149, #112)
This commit introduces a Cap_index class for Fiasco.OC's capabilities.
A Cap_index is a combination of the global capability id, that is used by Genode
to correctly identify a kernel-object, and a corresponding entry in a
protection-domain's (kernel-)capability-space. The cap-indices are non-copyable,
unique objects, that are held in a Cap_map. The Cap_map is used to re-find
capabilities already present in the protection-domain, when a capability is
received via IPC. The retrieval of capabilities effectively fixes issue #112,
meaning the waste of capability-space entries.
Because Cap_index objects are non-copyable (their address indicates the position
in the capability-space of the pd), they are inappropriate to use as
Native_capability. Therefore, Native_capability is implemented as a reference
to Cap_index objects. This design seems to be a good pre-condition to implement
smart-pointers for entries in the capability-space, and thereby closing existing
leaks (please refer to issue #32).
Cap_index, Cap_map, and the allocator for Cap_index objects are designed in a way,
that it should be relatively easy to apply the same concept to NOVA also. By now,
these classes are located in the `base-foc` repository, but they intentionally
contain no Fiasco.OC specific elements.
The previously explained changes had extensive impact on the whole Fiasco.OC
platform implementation, due to various dependencies. The following things had to
be changed:
* The Thread object's startup and destruction routine is re-arranged, to
enable another thread (that calls the Thread destructor) gaining the
capability id of the thread's gate to remove it from the Cap_map, the
thread's UTCB had to be made available to the caller, because there
is the current location of that id. After having the UTCB available
in the Thread object for that reason, the whole thread bootstrapping
could be simplified.
* In the course of changing the Native_capability's semantic, a new Cap_mapping
class was introduced in core, that facilitates the establishment and
destruction of capability mappings between core and it's client's, especially
mappings related to Platform_thread and Platform_task, that are relevant to
task and thread creation and destruction. Thereby, the destruction of
threads had to be reworked, which effectively removed a bug (issue #149)
where some threads weren't destroyed properly.
* In the quick fix for issue #112, something similar to the Cap_map was
introduced available in all processes. Moreover, some kind of a capability
map already existed in core, to handle cap-session request properly. The
introduction of the Cap_map unified both structures, so that the
cap-session component code in core had to be reworked too.
* The platform initialization code had to be changed sligthly due to the
changes in Native_capability
* The vcpu initialization in the L4Linux support library had to be adapted
according to the already mentioned changes in the Thread object's bootstrap
code.
2012-03-15 11:41:24 +00:00
|
|
|
|
|
|
|
/**
|
|
|
|
* Free a range of Cap_index objects
|
|
|
|
*
|
|
|
|
* \param idx pointer to first object in range
|
|
|
|
* \param cnt number of objects to free
|
|
|
|
* \throw Index_out_of_bounds if address is out of scope
|
|
|
|
*/
|
|
|
|
virtual void free(Cap_index *idx, size_t cnt) = 0;
|
|
|
|
|
|
|
|
/**
|
|
|
|
* Get the Cap_index object's address in capability space
|
|
|
|
*
|
|
|
|
* \param idx pointer to the Cap_index object in question
|
|
|
|
*/
|
|
|
|
virtual addr_t idx_to_kcap(Cap_index *idx) = 0;
|
|
|
|
|
|
|
|
/**
|
|
|
|
* Get the Cap_index object of a specific location
|
|
|
|
* in the capability space
|
|
|
|
*
|
|
|
|
* \param kcap the address in the capability space
|
|
|
|
*/
|
|
|
|
virtual Cap_index* kcap_to_idx(addr_t kcap) = 0;
|
Fiasco.OC: several capability ref-counter fixes.
This commit fixes several issues that were triggered e.g. by the
'noux_tool_chain' run-script (fix #208 in part). The following problems
are tackled:
* Don't reference count capability selectors within a task that are actually
controlled by core (all beneath 0x200000), because it's undecideable which
"version" of a capability selector we currently use, e.g. a thread gets
destroyed and a new one gets created immediately some other thread might
have a Native_capability pointing to the already destroyed thread's gate
capability-slot, that is now a new valid one (the one of the new thread)
* In core we cannot invalidate and remove a capability from the so called
Cap_map before each reference to it is destroyed, so don't do this in
Cap_session_component::free, but only reference-decrement within there,
the actual removal can only be done in Cap_map::remove. Because core also
has to invalidate a capability to be removed in all protection-domains
we have to implement a core specific Cap_map::remove method
* When a capability gets inserted into the Cap_map, and we detect an old
invalid entry with the dame id in the tree, don't just overmap that
invalid entry (as there exist remaining references to it), but just remove
it from the tree and allocate an new entry.
* Use the Cap_session_component interface to free a Pager_object when it
gets dissolved, as its also used for allocation
2012-08-29 13:52:18 +00:00
|
|
|
|
|
|
|
/**
|
|
|
|
* Returns whether a Cap_index object is from the region
|
|
|
|
* controlled by core, or not.
|
|
|
|
*
|
|
|
|
* \param idx pointer to the Cap_index object in question
|
|
|
|
*/
|
|
|
|
virtual bool static_idx(Cap_index *idx) = 0;
|
Fiasco.OC: introduce Cap_index (fixes #149, #112)
This commit introduces a Cap_index class for Fiasco.OC's capabilities.
A Cap_index is a combination of the global capability id, that is used by Genode
to correctly identify a kernel-object, and a corresponding entry in a
protection-domain's (kernel-)capability-space. The cap-indices are non-copyable,
unique objects, that are held in a Cap_map. The Cap_map is used to re-find
capabilities already present in the protection-domain, when a capability is
received via IPC. The retrieval of capabilities effectively fixes issue #112,
meaning the waste of capability-space entries.
Because Cap_index objects are non-copyable (their address indicates the position
in the capability-space of the pd), they are inappropriate to use as
Native_capability. Therefore, Native_capability is implemented as a reference
to Cap_index objects. This design seems to be a good pre-condition to implement
smart-pointers for entries in the capability-space, and thereby closing existing
leaks (please refer to issue #32).
Cap_index, Cap_map, and the allocator for Cap_index objects are designed in a way,
that it should be relatively easy to apply the same concept to NOVA also. By now,
these classes are located in the `base-foc` repository, but they intentionally
contain no Fiasco.OC specific elements.
The previously explained changes had extensive impact on the whole Fiasco.OC
platform implementation, due to various dependencies. The following things had to
be changed:
* The Thread object's startup and destruction routine is re-arranged, to
enable another thread (that calls the Thread destructor) gaining the
capability id of the thread's gate to remove it from the Cap_map, the
thread's UTCB had to be made available to the caller, because there
is the current location of that id. After having the UTCB available
in the Thread object for that reason, the whole thread bootstrapping
could be simplified.
* In the course of changing the Native_capability's semantic, a new Cap_mapping
class was introduced in core, that facilitates the establishment and
destruction of capability mappings between core and it's client's, especially
mappings related to Platform_thread and Platform_task, that are relevant to
task and thread creation and destruction. Thereby, the destruction of
threads had to be reworked, which effectively removed a bug (issue #149)
where some threads weren't destroyed properly.
* In the quick fix for issue #112, something similar to the Cap_map was
introduced available in all processes. Moreover, some kind of a capability
map already existed in core, to handle cap-session request properly. The
introduction of the Cap_map unified both structures, so that the
cap-session component code in core had to be reworked too.
* The platform initialization code had to be changed sligthly due to the
changes in Native_capability
* The vcpu initialization in the L4Linux support library had to be adapted
according to the already mentioned changes in the Thread object's bootstrap
code.
2012-03-15 11:41:24 +00:00
|
|
|
};
|
|
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
* Get the global Cap_index_allocator of the process.
|
|
|
|
*/
|
|
|
|
Cap_index_allocator *cap_idx_alloc();
|
|
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
* Low-level spin-lock to protect Cap_index_allocator and the Cap_map
|
|
|
|
*
|
|
|
|
* We cannot use a normal Genode lock because this lock is used by code
|
|
|
|
* executed prior the initialization of Genode
|
|
|
|
*/
|
|
|
|
class Spin_lock
|
|
|
|
{
|
|
|
|
private:
|
|
|
|
|
|
|
|
volatile int _spinlock;
|
|
|
|
|
|
|
|
public:
|
|
|
|
|
|
|
|
/**
|
|
|
|
* Constructor
|
|
|
|
*/
|
|
|
|
Spin_lock();
|
|
|
|
|
|
|
|
void lock();
|
|
|
|
void unlock();
|
|
|
|
|
|
|
|
/**
|
|
|
|
* Lock guard
|
|
|
|
*/
|
|
|
|
typedef Genode::Lock_guard<Spin_lock> Guard;
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
|
|
class Native_capability;
|
|
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
* The Capability_map is an AVL-tree of Cap_index objects that can be
|
|
|
|
* found via the global capability id
|
|
|
|
*
|
|
|
|
* It is used to re-find capabilities whenever a capability gets
|
|
|
|
* transfered to a process, so that we can re-use an existing one
|
|
|
|
* to save entries in the capability space, and prevent leaks of
|
|
|
|
* them.
|
|
|
|
*/
|
|
|
|
class Capability_map : Noncopyable
|
|
|
|
{
|
|
|
|
private:
|
|
|
|
|
|
|
|
Avl_tree<Cap_index> _tree;
|
|
|
|
Spin_lock _lock;
|
|
|
|
|
|
|
|
public:
|
|
|
|
|
|
|
|
/**
|
|
|
|
* Find an existing Cap_index via a capability id
|
|
|
|
*
|
|
|
|
* \param id the global capability id
|
|
|
|
* \return pointer of Cap_index when found, otherwise zero
|
|
|
|
*/
|
|
|
|
Cap_index* find(int id);
|
|
|
|
|
|
|
|
/**
|
|
|
|
* Create and insert a new Cap_index with a specific capability id
|
|
|
|
*
|
|
|
|
* Allocation of the Cap_index is done via the global
|
|
|
|
* Cap_index_allocator, which might throw exceptions that aren't
|
|
|
|
* caught by this method
|
|
|
|
*
|
|
|
|
* \param id the global capability id
|
|
|
|
* \return pointer to the new Cap_index object, or zero
|
|
|
|
* when allocation failed
|
|
|
|
*/
|
|
|
|
Cap_index* insert(int id);
|
|
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
* Create and insert a new Cap_index with a specific capability id,
|
|
|
|
* and location in capability space
|
|
|
|
*
|
Fiasco.OC: several capability ref-counter fixes.
This commit fixes several issues that were triggered e.g. by the
'noux_tool_chain' run-script (fix #208 in part). The following problems
are tackled:
* Don't reference count capability selectors within a task that are actually
controlled by core (all beneath 0x200000), because it's undecideable which
"version" of a capability selector we currently use, e.g. a thread gets
destroyed and a new one gets created immediately some other thread might
have a Native_capability pointing to the already destroyed thread's gate
capability-slot, that is now a new valid one (the one of the new thread)
* In core we cannot invalidate and remove a capability from the so called
Cap_map before each reference to it is destroyed, so don't do this in
Cap_session_component::free, but only reference-decrement within there,
the actual removal can only be done in Cap_map::remove. Because core also
has to invalidate a capability to be removed in all protection-domains
we have to implement a core specific Cap_map::remove method
* When a capability gets inserted into the Cap_map, and we detect an old
invalid entry with the dame id in the tree, don't just overmap that
invalid entry (as there exist remaining references to it), but just remove
it from the tree and allocate an new entry.
* Use the Cap_session_component interface to free a Pager_object when it
gets dissolved, as its also used for allocation
2012-08-29 13:52:18 +00:00
|
|
|
* A previously existent entry with the same id gets removed!
|
Fiasco.OC: introduce Cap_index (fixes #149, #112)
This commit introduces a Cap_index class for Fiasco.OC's capabilities.
A Cap_index is a combination of the global capability id, that is used by Genode
to correctly identify a kernel-object, and a corresponding entry in a
protection-domain's (kernel-)capability-space. The cap-indices are non-copyable,
unique objects, that are held in a Cap_map. The Cap_map is used to re-find
capabilities already present in the protection-domain, when a capability is
received via IPC. The retrieval of capabilities effectively fixes issue #112,
meaning the waste of capability-space entries.
Because Cap_index objects are non-copyable (their address indicates the position
in the capability-space of the pd), they are inappropriate to use as
Native_capability. Therefore, Native_capability is implemented as a reference
to Cap_index objects. This design seems to be a good pre-condition to implement
smart-pointers for entries in the capability-space, and thereby closing existing
leaks (please refer to issue #32).
Cap_index, Cap_map, and the allocator for Cap_index objects are designed in a way,
that it should be relatively easy to apply the same concept to NOVA also. By now,
these classes are located in the `base-foc` repository, but they intentionally
contain no Fiasco.OC specific elements.
The previously explained changes had extensive impact on the whole Fiasco.OC
platform implementation, due to various dependencies. The following things had to
be changed:
* The Thread object's startup and destruction routine is re-arranged, to
enable another thread (that calls the Thread destructor) gaining the
capability id of the thread's gate to remove it from the Cap_map, the
thread's UTCB had to be made available to the caller, because there
is the current location of that id. After having the UTCB available
in the Thread object for that reason, the whole thread bootstrapping
could be simplified.
* In the course of changing the Native_capability's semantic, a new Cap_mapping
class was introduced in core, that facilitates the establishment and
destruction of capability mappings between core and it's client's, especially
mappings related to Platform_thread and Platform_task, that are relevant to
task and thread creation and destruction. Thereby, the destruction of
threads had to be reworked, which effectively removed a bug (issue #149)
where some threads weren't destroyed properly.
* In the quick fix for issue #112, something similar to the Cap_map was
introduced available in all processes. Moreover, some kind of a capability
map already existed in core, to handle cap-session request properly. The
introduction of the Cap_map unified both structures, so that the
cap-session component code in core had to be reworked too.
* The platform initialization code had to be changed sligthly due to the
changes in Native_capability
* The vcpu initialization in the L4Linux support library had to be adapted
according to the already mentioned changes in the Thread object's bootstrap
code.
2012-03-15 11:41:24 +00:00
|
|
|
*
|
|
|
|
* \param id the global capability id
|
|
|
|
* \param kcap address in capability space
|
|
|
|
* \return pointer to the new Cap_index object, or zero
|
|
|
|
* when allocation failed
|
|
|
|
*/
|
|
|
|
Cap_index* insert(int id, addr_t kcap);
|
|
|
|
|
2012-05-29 08:00:28 +00:00
|
|
|
/**
|
|
|
|
* Create and insert a new Cap_index with a specific capability id
|
Fiasco.OC: several capability ref-counter fixes.
This commit fixes several issues that were triggered e.g. by the
'noux_tool_chain' run-script (fix #208 in part). The following problems
are tackled:
* Don't reference count capability selectors within a task that are actually
controlled by core (all beneath 0x200000), because it's undecideable which
"version" of a capability selector we currently use, e.g. a thread gets
destroyed and a new one gets created immediately some other thread might
have a Native_capability pointing to the already destroyed thread's gate
capability-slot, that is now a new valid one (the one of the new thread)
* In core we cannot invalidate and remove a capability from the so called
Cap_map before each reference to it is destroyed, so don't do this in
Cap_session_component::free, but only reference-decrement within there,
the actual removal can only be done in Cap_map::remove. Because core also
has to invalidate a capability to be removed in all protection-domains
we have to implement a core specific Cap_map::remove method
* When a capability gets inserted into the Cap_map, and we detect an old
invalid entry with the dame id in the tree, don't just overmap that
invalid entry (as there exist remaining references to it), but just remove
it from the tree and allocate an new entry.
* Use the Cap_session_component interface to free a Pager_object when it
gets dissolved, as its also used for allocation
2012-08-29 13:52:18 +00:00
|
|
|
* and map from given kcap to newly allocated one,
|
|
|
|
* if the an entry with the same id exists already,
|
|
|
|
* it is returned if it points to the same kernel-object,
|
|
|
|
* or gets overridden if it's already invalid.
|
2012-05-29 08:00:28 +00:00
|
|
|
*
|
|
|
|
* Allocation of the Cap_index is done via the global
|
|
|
|
* Cap_index_allocator, which might throw exceptions that aren't
|
|
|
|
* caught by this method
|
|
|
|
*
|
|
|
|
* \param id the global capability id
|
|
|
|
* \return pointer to the new Cap_index object, or zero
|
Fiasco.OC: several capability ref-counter fixes.
This commit fixes several issues that were triggered e.g. by the
'noux_tool_chain' run-script (fix #208 in part). The following problems
are tackled:
* Don't reference count capability selectors within a task that are actually
controlled by core (all beneath 0x200000), because it's undecideable which
"version" of a capability selector we currently use, e.g. a thread gets
destroyed and a new one gets created immediately some other thread might
have a Native_capability pointing to the already destroyed thread's gate
capability-slot, that is now a new valid one (the one of the new thread)
* In core we cannot invalidate and remove a capability from the so called
Cap_map before each reference to it is destroyed, so don't do this in
Cap_session_component::free, but only reference-decrement within there,
the actual removal can only be done in Cap_map::remove. Because core also
has to invalidate a capability to be removed in all protection-domains
we have to implement a core specific Cap_map::remove method
* When a capability gets inserted into the Cap_map, and we detect an old
invalid entry with the dame id in the tree, don't just overmap that
invalid entry (as there exist remaining references to it), but just remove
it from the tree and allocate an new entry.
* Use the Cap_session_component interface to free a Pager_object when it
gets dissolved, as its also used for allocation
2012-08-29 13:52:18 +00:00
|
|
|
* when allocation failed, or when a valid entry
|
|
|
|
* with the same id exists and it's kernel-object
|
|
|
|
* differs to the one given by kcap
|
2012-05-29 08:00:28 +00:00
|
|
|
*/
|
|
|
|
Cap_index* insert_map(int id, addr_t kcap);
|
|
|
|
|
Fiasco.OC: introduce Cap_index (fixes #149, #112)
This commit introduces a Cap_index class for Fiasco.OC's capabilities.
A Cap_index is a combination of the global capability id, that is used by Genode
to correctly identify a kernel-object, and a corresponding entry in a
protection-domain's (kernel-)capability-space. The cap-indices are non-copyable,
unique objects, that are held in a Cap_map. The Cap_map is used to re-find
capabilities already present in the protection-domain, when a capability is
received via IPC. The retrieval of capabilities effectively fixes issue #112,
meaning the waste of capability-space entries.
Because Cap_index objects are non-copyable (their address indicates the position
in the capability-space of the pd), they are inappropriate to use as
Native_capability. Therefore, Native_capability is implemented as a reference
to Cap_index objects. This design seems to be a good pre-condition to implement
smart-pointers for entries in the capability-space, and thereby closing existing
leaks (please refer to issue #32).
Cap_index, Cap_map, and the allocator for Cap_index objects are designed in a way,
that it should be relatively easy to apply the same concept to NOVA also. By now,
these classes are located in the `base-foc` repository, but they intentionally
contain no Fiasco.OC specific elements.
The previously explained changes had extensive impact on the whole Fiasco.OC
platform implementation, due to various dependencies. The following things had to
be changed:
* The Thread object's startup and destruction routine is re-arranged, to
enable another thread (that calls the Thread destructor) gaining the
capability id of the thread's gate to remove it from the Cap_map, the
thread's UTCB had to be made available to the caller, because there
is the current location of that id. After having the UTCB available
in the Thread object for that reason, the whole thread bootstrapping
could be simplified.
* In the course of changing the Native_capability's semantic, a new Cap_mapping
class was introduced in core, that facilitates the establishment and
destruction of capability mappings between core and it's client's, especially
mappings related to Platform_thread and Platform_task, that are relevant to
task and thread creation and destruction. Thereby, the destruction of
threads had to be reworked, which effectively removed a bug (issue #149)
where some threads weren't destroyed properly.
* In the quick fix for issue #112, something similar to the Cap_map was
introduced available in all processes. Moreover, some kind of a capability
map already existed in core, to handle cap-session request properly. The
introduction of the Cap_map unified both structures, so that the
cap-session component code in core had to be reworked too.
* The platform initialization code had to be changed sligthly due to the
changes in Native_capability
* The vcpu initialization in the L4Linux support library had to be adapted
according to the already mentioned changes in the Thread object's bootstrap
code.
2012-03-15 11:41:24 +00:00
|
|
|
/**
|
Fiasco.OC: several capability ref-counter fixes.
This commit fixes several issues that were triggered e.g. by the
'noux_tool_chain' run-script (fix #208 in part). The following problems
are tackled:
* Don't reference count capability selectors within a task that are actually
controlled by core (all beneath 0x200000), because it's undecideable which
"version" of a capability selector we currently use, e.g. a thread gets
destroyed and a new one gets created immediately some other thread might
have a Native_capability pointing to the already destroyed thread's gate
capability-slot, that is now a new valid one (the one of the new thread)
* In core we cannot invalidate and remove a capability from the so called
Cap_map before each reference to it is destroyed, so don't do this in
Cap_session_component::free, but only reference-decrement within there,
the actual removal can only be done in Cap_map::remove. Because core also
has to invalidate a capability to be removed in all protection-domains
we have to implement a core specific Cap_map::remove method
* When a capability gets inserted into the Cap_map, and we detect an old
invalid entry with the dame id in the tree, don't just overmap that
invalid entry (as there exist remaining references to it), but just remove
it from the tree and allocate an new entry.
* Use the Cap_session_component interface to free a Pager_object when it
gets dissolved, as its also used for allocation
2012-08-29 13:52:18 +00:00
|
|
|
* Remove (resp. invalidate) a Cap_index object
|
Fiasco.OC: introduce Cap_index (fixes #149, #112)
This commit introduces a Cap_index class for Fiasco.OC's capabilities.
A Cap_index is a combination of the global capability id, that is used by Genode
to correctly identify a kernel-object, and a corresponding entry in a
protection-domain's (kernel-)capability-space. The cap-indices are non-copyable,
unique objects, that are held in a Cap_map. The Cap_map is used to re-find
capabilities already present in the protection-domain, when a capability is
received via IPC. The retrieval of capabilities effectively fixes issue #112,
meaning the waste of capability-space entries.
Because Cap_index objects are non-copyable (their address indicates the position
in the capability-space of the pd), they are inappropriate to use as
Native_capability. Therefore, Native_capability is implemented as a reference
to Cap_index objects. This design seems to be a good pre-condition to implement
smart-pointers for entries in the capability-space, and thereby closing existing
leaks (please refer to issue #32).
Cap_index, Cap_map, and the allocator for Cap_index objects are designed in a way,
that it should be relatively easy to apply the same concept to NOVA also. By now,
these classes are located in the `base-foc` repository, but they intentionally
contain no Fiasco.OC specific elements.
The previously explained changes had extensive impact on the whole Fiasco.OC
platform implementation, due to various dependencies. The following things had to
be changed:
* The Thread object's startup and destruction routine is re-arranged, to
enable another thread (that calls the Thread destructor) gaining the
capability id of the thread's gate to remove it from the Cap_map, the
thread's UTCB had to be made available to the caller, because there
is the current location of that id. After having the UTCB available
in the Thread object for that reason, the whole thread bootstrapping
could be simplified.
* In the course of changing the Native_capability's semantic, a new Cap_mapping
class was introduced in core, that facilitates the establishment and
destruction of capability mappings between core and it's client's, especially
mappings related to Platform_thread and Platform_task, that are relevant to
task and thread creation and destruction. Thereby, the destruction of
threads had to be reworked, which effectively removed a bug (issue #149)
where some threads weren't destroyed properly.
* In the quick fix for issue #112, something similar to the Cap_map was
introduced available in all processes. Moreover, some kind of a capability
map already existed in core, to handle cap-session request properly. The
introduction of the Cap_map unified both structures, so that the
cap-session component code in core had to be reworked too.
* The platform initialization code had to be changed sligthly due to the
changes in Native_capability
* The vcpu initialization in the L4Linux support library had to be adapted
according to the already mentioned changes in the Thread object's bootstrap
code.
2012-03-15 11:41:24 +00:00
|
|
|
*
|
|
|
|
* \param i pointer to Cap_index object to remove
|
|
|
|
*/
|
|
|
|
void remove(Cap_index* i);
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
* Get the global Capability_map of the process.
|
|
|
|
*/
|
|
|
|
Capability_map *cap_map();
|
|
|
|
}
|
|
|
|
|
|
|
|
#endif /* _INCLUDE__BASE__CAP_MAP_H_ */
|