2019-05-29 16:57:47 -07:00
|
|
|
|
2010-12-28 14:25:21 -08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2012-05-26 06:20:19 -07:00
|
|
|
|
2010-12-28 14:25:21 -08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2014-06-06 14:37:31 -07:00
|
|
|
|
|
|
|
|
|
2010-12-28 14:25:21 -08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2012-05-26 06:20:19 -07:00
|
|
|
|
2023-06-27 13:25:41 -07:00
|
|
|
|
2010-12-28 14:25:21 -08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-08-11 15:14:39 -07:00
|
|
|
|
2010-12-28 14:25:21 -08:00
|
|
|
|
|
|
|
|
|
2012-05-26 06:20:28 -07:00
|
|
|
|
pstore: Replace crypto API compression with zlib_deflate library calls
Pstore supports compression using a variety of algorithms exposed by the
crypto API. This uses the deprecated comp (as opposed to scomp/acomp)
API, and so we should stop using that, and either move to the new API,
or switch to a different approach entirely.
Given that we only compress ASCII text in pstore, and considering that
this happens when the system is likely to be in an unstable state, the
flexibility that the complex crypto API provides does not outweigh its
impact on the risk that we might encounter additional problems when
trying to commit the kernel log contents to the pstore backend.
So let's switch [back] to the zlib deflate library API, and remove all
the complexity that really has no place in a low-level diagnostic
facility. Note that, while more modern compression algorithms have been
added to the kernel in recent years, the code size of zlib deflate is
substantially smaller than, e.g., zstd, while its performance in terms
of compression ratio is comparable for ASCII text, and speed is deemed
irrelevant in this context.
Note that this means that compressed pstore records may no longer be
accessible after a kernel upgrade, but this has never been part of the
contract. (The choice of compression algorithm is not stored in the
pstore records either)
Tested-by: "Guilherme G. Piccoli" <gpiccoli@igalia.com> # Steam Deck
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Eric Biggers <ebiggers@google.com>
Link: https://lore.kernel.org/r/20230712162332.2670437-3-ardb@kernel.org
Signed-off-by: Kees Cook <keescook@chromium.org>
2023-07-12 18:23:32 +02:00
|
|
|
|
2011-08-11 15:14:39 -07:00
|
|
|
|
pstore: Replace crypto API compression with zlib_deflate library calls
Pstore supports compression using a variety of algorithms exposed by the
crypto API. This uses the deprecated comp (as opposed to scomp/acomp)
API, and so we should stop using that, and either move to the new API,
or switch to a different approach entirely.
Given that we only compress ASCII text in pstore, and considering that
this happens when the system is likely to be in an unstable state, the
flexibility that the complex crypto API provides does not outweigh its
impact on the risk that we might encounter additional problems when
trying to commit the kernel log contents to the pstore backend.
So let's switch [back] to the zlib deflate library API, and remove all
the complexity that really has no place in a low-level diagnostic
facility. Note that, while more modern compression algorithms have been
added to the kernel in recent years, the code size of zlib deflate is
substantially smaller than, e.g., zstd, while its performance in terms
of compression ratio is comparable for ASCII text, and speed is deemed
irrelevant in this context.
Note that this means that compressed pstore records may no longer be
accessible after a kernel upgrade, but this has never been part of the
contract. (The choice of compression algorithm is not stored in the
pstore records either)
Tested-by: "Guilherme G. Piccoli" <gpiccoli@igalia.com> # Steam Deck
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Eric Biggers <ebiggers@google.com>
Link: https://lore.kernel.org/r/20230712162332.2670437-3-ardb@kernel.org
Signed-off-by: Kees Cook <keescook@chromium.org>
2023-07-12 18:23:32 +02:00
|
|
|
|
2010-12-28 14:25:21 -08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-08-11 15:14:39 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
pstore/platform: Disable automatic updates by default
Having automatic updates seems pointless for production system, and
even dangerous and thus counter-productive:
1. If we can mount pstore, or read files, we can as well read
/proc/kmsg. So, there's little point in duplicating the
functionality and present the same information but via another
userland ABI;
2. Expecting the kernel to behave sanely after oops/panic is naive.
It might work, but you'd rather not try it. Screwed up kernel
can do rather bad things, like recursive faults[1]; and pstore
rather provoking bad things to happen. It uses:
1. Timers (assumes sane interrupts state);
2. Workqueues and mutexes (assumes scheduler in a sane state);
3. kzalloc (a working slab allocator);
That's too much for a dead kernel, so the debugging facility
itself might just make debugging harder, which is not what
we want.
Maybe for non-oops message types it would make sense to re-enable
automatic updates, but so far I don't see any use case for this.
Even for tracing, it has its own run-time/normal ABI, so we're
only interested in pstore upon next boot, to retrieve what has
gone wrong with HW or SW.
So, let's disable the updates by default.
[1]
BUG: unable to handle kernel paging request at fffffffffffffff8
IP: [<ffffffff8104801b>] kthread_data+0xb/0x20
[...]
Process kworker/0:1 (pid: 14, threadinfo ffff8800072c0000, task ffff88000725b100)
[...
Call Trace:
[<ffffffff81043710>] wq_worker_sleeping+0x10/0xa0
[<ffffffff813687a8>] __schedule+0x568/0x7d0
[<ffffffff8106c24d>] ? trace_hardirqs_on+0xd/0x10
[<ffffffff81087e22>] ? call_rcu_sched+0x12/0x20
[<ffffffff8102b596>] ? release_task+0x156/0x2d0
[<ffffffff8102b45e>] ? release_task+0x1e/0x2d0
[<ffffffff8106c24d>] ? trace_hardirqs_on+0xd/0x10
[<ffffffff81368ac4>] schedule+0x24/0x70
[<ffffffff8102cba8>] do_exit+0x1f8/0x370
[<ffffffff810051e7>] oops_end+0x77/0xb0
[<ffffffff8135c301>] no_context+0x1a6/0x1b5
[<ffffffff8135c4de>] __bad_area_nosemaphore+0x1ce/0x1ed
[<ffffffff81053156>] ? ttwu_queue+0xc6/0xe0
[<ffffffff8135c50b>] bad_area_nosemaphore+0xe/0x10
[<ffffffff8101fa47>] do_page_fault+0x2c7/0x450
[<ffffffff8106e34b>] ? __lock_release+0x6b/0xe0
[<ffffffff8106bf21>] ? mark_held_locks+0x61/0x140
[<ffffffff810502fe>] ? __wake_up+0x4e/0x70
[<ffffffff81185f7d>] ? trace_hardirqs_off_thunk+0x3a/0x3c
[<ffffffff81158970>] ? pstore_register+0x120/0x120
[<ffffffff8136a37f>] page_fault+0x1f/0x30
[<ffffffff81158970>] ? pstore_register+0x120/0x120
[<ffffffff81185ab8>] ? memcpy+0x68/0x110
[<ffffffff8115875a>] ? pstore_get_records+0x3a/0x130
[<ffffffff811590f4>] ? persistent_ram_copy_old+0x64/0x90
[<ffffffff81158bf4>] ramoops_pstore_read+0x84/0x130
[<ffffffff81158799>] pstore_get_records+0x79/0x130
[<ffffffff81042536>] ? process_one_work+0x116/0x450
[<ffffffff81158970>] ? pstore_register+0x120/0x120
[<ffffffff8115897e>] pstore_dowork+0xe/0x10
[<ffffffff81042594>] process_one_work+0x174/0x450
[<ffffffff81042536>] ? process_one_work+0x116/0x450
[<ffffffff81042e13>] worker_thread+0x123/0x2d0
[<ffffffff81042cf0>] ? manage_workers.isra.28+0x120/0x120
[<ffffffff81047d8e>] kthread+0x8e/0xa0
[<ffffffff8136ba74>] kernel_thread_helper+0x4/0x10
[<ffffffff8136a199>] ? retint_restore_args+0xe/0xe
[<ffffffff81047d00>] ? __init_kthread_worker+0x70/0x70
[<ffffffff8136ba70>] ? gs_change+0xb/0xb
Code: be e2 00 00 00 48 c7 c7 d1 2a 4e 81 e8 bf fb fd ff 48 8b 5d f0 4c 8b 65 f8 c9 c3 0f 1f 44 00 00 48 8b 87 08 02 00 00 55 48 89 e5 <48> 8b 40 f8 5d c3 66 66 66 66 66 66 2e 0f 1f 84 00 00 00 00 00
RIP [<ffffffff8104801b>] kthread_data+0xb/0x20
RSP <ffff8800072c1888>
CR2: fffffffffffffff8
---[ end trace 996a332dc399111d ]---
Fixing recursive fault but reboot is needed!
Signed-off-by: Anton Vorontsov <anton.vorontsov@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-05-26 06:20:29 -07:00
|
|
|
|
2012-05-26 06:20:28 -07:00
|
|
|
|
|
|
|
|
|
pstore/platform: Disable automatic updates by default
Having automatic updates seems pointless for production system, and
even dangerous and thus counter-productive:
1. If we can mount pstore, or read files, we can as well read
/proc/kmsg. So, there's little point in duplicating the
functionality and present the same information but via another
userland ABI;
2. Expecting the kernel to behave sanely after oops/panic is naive.
It might work, but you'd rather not try it. Screwed up kernel
can do rather bad things, like recursive faults[1]; and pstore
rather provoking bad things to happen. It uses:
1. Timers (assumes sane interrupts state);
2. Workqueues and mutexes (assumes scheduler in a sane state);
3. kzalloc (a working slab allocator);
That's too much for a dead kernel, so the debugging facility
itself might just make debugging harder, which is not what
we want.
Maybe for non-oops message types it would make sense to re-enable
automatic updates, but so far I don't see any use case for this.
Even for tracing, it has its own run-time/normal ABI, so we're
only interested in pstore upon next boot, to retrieve what has
gone wrong with HW or SW.
So, let's disable the updates by default.
[1]
BUG: unable to handle kernel paging request at fffffffffffffff8
IP: [<ffffffff8104801b>] kthread_data+0xb/0x20
[...]
Process kworker/0:1 (pid: 14, threadinfo ffff8800072c0000, task ffff88000725b100)
[...
Call Trace:
[<ffffffff81043710>] wq_worker_sleeping+0x10/0xa0
[<ffffffff813687a8>] __schedule+0x568/0x7d0
[<ffffffff8106c24d>] ? trace_hardirqs_on+0xd/0x10
[<ffffffff81087e22>] ? call_rcu_sched+0x12/0x20
[<ffffffff8102b596>] ? release_task+0x156/0x2d0
[<ffffffff8102b45e>] ? release_task+0x1e/0x2d0
[<ffffffff8106c24d>] ? trace_hardirqs_on+0xd/0x10
[<ffffffff81368ac4>] schedule+0x24/0x70
[<ffffffff8102cba8>] do_exit+0x1f8/0x370
[<ffffffff810051e7>] oops_end+0x77/0xb0
[<ffffffff8135c301>] no_context+0x1a6/0x1b5
[<ffffffff8135c4de>] __bad_area_nosemaphore+0x1ce/0x1ed
[<ffffffff81053156>] ? ttwu_queue+0xc6/0xe0
[<ffffffff8135c50b>] bad_area_nosemaphore+0xe/0x10
[<ffffffff8101fa47>] do_page_fault+0x2c7/0x450
[<ffffffff8106e34b>] ? __lock_release+0x6b/0xe0
[<ffffffff8106bf21>] ? mark_held_locks+0x61/0x140
[<ffffffff810502fe>] ? __wake_up+0x4e/0x70
[<ffffffff81185f7d>] ? trace_hardirqs_off_thunk+0x3a/0x3c
[<ffffffff81158970>] ? pstore_register+0x120/0x120
[<ffffffff8136a37f>] page_fault+0x1f/0x30
[<ffffffff81158970>] ? pstore_register+0x120/0x120
[<ffffffff81185ab8>] ? memcpy+0x68/0x110
[<ffffffff8115875a>] ? pstore_get_records+0x3a/0x130
[<ffffffff811590f4>] ? persistent_ram_copy_old+0x64/0x90
[<ffffffff81158bf4>] ramoops_pstore_read+0x84/0x130
[<ffffffff81158799>] pstore_get_records+0x79/0x130
[<ffffffff81042536>] ? process_one_work+0x116/0x450
[<ffffffff81158970>] ? pstore_register+0x120/0x120
[<ffffffff8115897e>] pstore_dowork+0xe/0x10
[<ffffffff81042594>] process_one_work+0x174/0x450
[<ffffffff81042536>] ? process_one_work+0x116/0x450
[<ffffffff81042e13>] worker_thread+0x123/0x2d0
[<ffffffff81042cf0>] ? manage_workers.isra.28+0x120/0x120
[<ffffffff81047d8e>] kthread+0x8e/0xa0
[<ffffffff8136ba74>] kernel_thread_helper+0x4/0x10
[<ffffffff8136a199>] ? retint_restore_args+0xe/0xe
[<ffffffff81047d00>] ? __init_kthread_worker+0x70/0x70
[<ffffffff8136ba70>] ? gs_change+0xb/0xb
Code: be e2 00 00 00 48 c7 c7 d1 2a 4e 81 e8 bf fb fd ff 48 8b 5d f0 4c 8b 65 f8 c9 c3 0f 1f 44 00 00 48 8b 87 08 02 00 00 55 48 89 e5 <48> 8b 40 f8 5d c3 66 66 66 66 66 66 2e 0f 1f 84 00 00 00 00 00
RIP [<ffffffff8104801b>] kthread_data+0xb/0x20
RSP <ffff8800072c1888>
CR2: fffffffffffffff8
---[ end trace 996a332dc399111d ]---
Fixing recursive fault but reboot is needed!
Signed-off-by: Anton Vorontsov <anton.vorontsov@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-05-26 06:20:29 -07:00
|
|
|
|
2020-05-05 21:36:15 -07:00
|
|
|
|
pstore/platform: Disable automatic updates by default
Having automatic updates seems pointless for production system, and
even dangerous and thus counter-productive:
1. If we can mount pstore, or read files, we can as well read
/proc/kmsg. So, there's little point in duplicating the
functionality and present the same information but via another
userland ABI;
2. Expecting the kernel to behave sanely after oops/panic is naive.
It might work, but you'd rather not try it. Screwed up kernel
can do rather bad things, like recursive faults[1]; and pstore
rather provoking bad things to happen. It uses:
1. Timers (assumes sane interrupts state);
2. Workqueues and mutexes (assumes scheduler in a sane state);
3. kzalloc (a working slab allocator);
That's too much for a dead kernel, so the debugging facility
itself might just make debugging harder, which is not what
we want.
Maybe for non-oops message types it would make sense to re-enable
automatic updates, but so far I don't see any use case for this.
Even for tracing, it has its own run-time/normal ABI, so we're
only interested in pstore upon next boot, to retrieve what has
gone wrong with HW or SW.
So, let's disable the updates by default.
[1]
BUG: unable to handle kernel paging request at fffffffffffffff8
IP: [<ffffffff8104801b>] kthread_data+0xb/0x20
[...]
Process kworker/0:1 (pid: 14, threadinfo ffff8800072c0000, task ffff88000725b100)
[...
Call Trace:
[<ffffffff81043710>] wq_worker_sleeping+0x10/0xa0
[<ffffffff813687a8>] __schedule+0x568/0x7d0
[<ffffffff8106c24d>] ? trace_hardirqs_on+0xd/0x10
[<ffffffff81087e22>] ? call_rcu_sched+0x12/0x20
[<ffffffff8102b596>] ? release_task+0x156/0x2d0
[<ffffffff8102b45e>] ? release_task+0x1e/0x2d0
[<ffffffff8106c24d>] ? trace_hardirqs_on+0xd/0x10
[<ffffffff81368ac4>] schedule+0x24/0x70
[<ffffffff8102cba8>] do_exit+0x1f8/0x370
[<ffffffff810051e7>] oops_end+0x77/0xb0
[<ffffffff8135c301>] no_context+0x1a6/0x1b5
[<ffffffff8135c4de>] __bad_area_nosemaphore+0x1ce/0x1ed
[<ffffffff81053156>] ? ttwu_queue+0xc6/0xe0
[<ffffffff8135c50b>] bad_area_nosemaphore+0xe/0x10
[<ffffffff8101fa47>] do_page_fault+0x2c7/0x450
[<ffffffff8106e34b>] ? __lock_release+0x6b/0xe0
[<ffffffff8106bf21>] ? mark_held_locks+0x61/0x140
[<ffffffff810502fe>] ? __wake_up+0x4e/0x70
[<ffffffff81185f7d>] ? trace_hardirqs_off_thunk+0x3a/0x3c
[<ffffffff81158970>] ? pstore_register+0x120/0x120
[<ffffffff8136a37f>] page_fault+0x1f/0x30
[<ffffffff81158970>] ? pstore_register+0x120/0x120
[<ffffffff81185ab8>] ? memcpy+0x68/0x110
[<ffffffff8115875a>] ? pstore_get_records+0x3a/0x130
[<ffffffff811590f4>] ? persistent_ram_copy_old+0x64/0x90
[<ffffffff81158bf4>] ramoops_pstore_read+0x84/0x130
[<ffffffff81158799>] pstore_get_records+0x79/0x130
[<ffffffff81042536>] ? process_one_work+0x116/0x450
[<ffffffff81158970>] ? pstore_register+0x120/0x120
[<ffffffff8115897e>] pstore_dowork+0xe/0x10
[<ffffffff81042594>] process_one_work+0x174/0x450
[<ffffffff81042536>] ? process_one_work+0x116/0x450
[<ffffffff81042e13>] worker_thread+0x123/0x2d0
[<ffffffff81042cf0>] ? manage_workers.isra.28+0x120/0x120
[<ffffffff81047d8e>] kthread+0x8e/0xa0
[<ffffffff8136ba74>] kernel_thread_helper+0x4/0x10
[<ffffffff8136a199>] ? retint_restore_args+0xe/0xe
[<ffffffff81047d00>] ? __init_kthread_worker+0x70/0x70
[<ffffffff8136ba70>] ? gs_change+0xb/0xb
Code: be e2 00 00 00 48 c7 c7 d1 2a 4e 81 e8 bf fb fd ff 48 8b 5d f0 4c 8b 65 f8 c9 c3 0f 1f 44 00 00 48 8b 87 08 02 00 00 55 48 89 e5 <48> 8b 40 f8 5d c3 66 66 66 66 66 66 2e 0f 1f 84 00 00 00 00 00
RIP [<ffffffff8104801b>] kthread_data+0xb/0x20
RSP <ffff8800072c1888>
CR2: fffffffffffffff8
---[ end trace 996a332dc399111d ]---
Fixing recursive fault but reboot is needed!
Signed-off-by: Anton Vorontsov <anton.vorontsov@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2012-05-26 06:20:29 -07:00
|
|
|
|
2011-08-11 15:14:39 -07:00
|
|
|
|
2018-11-03 16:38:16 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-08-11 15:14:39 -07:00
|
|
|
|
|
|
|
|
|
2017-08-28 11:28:21 -07:00
|
|
|
|
2017-10-04 16:27:04 -07:00
|
|
|
|
2011-08-11 15:14:39 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2010-12-28 14:25:21 -08:00
|
|
|
|
2020-05-04 19:33:54 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2010-12-28 14:25:21 -08:00
|
|
|
|
2020-05-04 19:31:36 -07:00
|
|
|
|
2012-07-09 17:10:41 -07:00
|
|
|
|
2010-12-28 14:25:21 -08:00
|
|
|
|
2011-07-21 16:57:55 -04:00
|
|
|
|
2020-05-08 09:25:19 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
pstore: Replace crypto API compression with zlib_deflate library calls
Pstore supports compression using a variety of algorithms exposed by the
crypto API. This uses the deprecated comp (as opposed to scomp/acomp)
API, and so we should stop using that, and either move to the new API,
or switch to a different approach entirely.
Given that we only compress ASCII text in pstore, and considering that
this happens when the system is likely to be in an unstable state, the
flexibility that the complex crypto API provides does not outweigh its
impact on the risk that we might encounter additional problems when
trying to commit the kernel log contents to the pstore backend.
So let's switch [back] to the zlib deflate library API, and remove all
the complexity that really has no place in a low-level diagnostic
facility. Note that, while more modern compression algorithms have been
added to the kernel in recent years, the code size of zlib deflate is
substantially smaller than, e.g., zstd, while its performance in terms
of compression ratio is comparable for ASCII text, and speed is deemed
irrelevant in this context.
Note that this means that compressed pstore records may no longer be
accessible after a kernel upgrade, but this has never been part of the
contract. (The choice of compression algorithm is not stored in the
pstore records either)
Tested-by: "Guilherme G. Piccoli" <gpiccoli@igalia.com> # Steam Deck
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Eric Biggers <ebiggers@google.com>
Link: https://lore.kernel.org/r/20230712162332.2670437-3-ardb@kernel.org
Signed-off-by: Kees Cook <keescook@chromium.org>
2023-07-12 18:23:32 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2020-05-08 09:25:19 -07:00
|
|
|
|
|
|
|
|
|
2011-07-21 16:57:55 -04:00
|
|
|
|
2022-10-06 19:42:06 -03:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
pstore: Replace crypto API compression with zlib_deflate library calls
Pstore supports compression using a variety of algorithms exposed by the
crypto API. This uses the deprecated comp (as opposed to scomp/acomp)
API, and so we should stop using that, and either move to the new API,
or switch to a different approach entirely.
Given that we only compress ASCII text in pstore, and considering that
this happens when the system is likely to be in an unstable state, the
flexibility that the complex crypto API provides does not outweigh its
impact on the risk that we might encounter additional problems when
trying to commit the kernel log contents to the pstore backend.
So let's switch [back] to the zlib deflate library API, and remove all
the complexity that really has no place in a low-level diagnostic
facility. Note that, while more modern compression algorithms have been
added to the kernel in recent years, the code size of zlib deflate is
substantially smaller than, e.g., zstd, while its performance in terms
of compression ratio is comparable for ASCII text, and speed is deemed
irrelevant in this context.
Note that this means that compressed pstore records may no longer be
accessible after a kernel upgrade, but this has never been part of the
contract. (The choice of compression algorithm is not stored in the
pstore records either)
Tested-by: "Guilherme G. Piccoli" <gpiccoli@igalia.com> # Steam Deck
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Eric Biggers <ebiggers@google.com>
Link: https://lore.kernel.org/r/20230712162332.2670437-3-ardb@kernel.org
Signed-off-by: Kees Cook <keescook@chromium.org>
2023-07-12 18:23:32 +02:00
|
|
|
|
2016-02-18 22:04:22 +08:00
|
|
|
|
pstore: Base compression input buffer size on estimated compressed size
Commit 1756ddea6916 ("pstore: Remove worst-case compression size logic")
removed some clunky per-algorithm worst case size estimation routines on
the basis that we can always store pstore records uncompressed, and
these worst case estimations are about how much the size might
inadvertently *increase* due to encapsulation overhead when the input
cannot be compressed at all. So if compression results in a size
increase, we just store the original data instead.
However, it seems that the original code was misinterpreting these
calculations as an estimation of how much uncompressed data might fit
into a compressed buffer of a given size, and it was using the results
to consume the input data in larger chunks than the pstore record size,
relying on the compression to ensure that what ultimately gets stored
fits into the available space.
One result of this, as observed and reported by Linus, is that upgrading
to a newer kernel that includes the given commit may result in pstore
decompression errors reported in the kernel log. This is due to the fact
that the existing records may unexpectedly decompress to a size that is
larger than the pstore record size.
Another potential problem caused by this change is that we may
underutilize the fixed sized records on pstore backends such as ramoops.
And on pstore backends with variable sized records such as EFI, we will
end up creating many more entries than before to store the same amount
of compressed data.
So let's fix both issues, by bringing back the typical case estimation of
how much ASCII text captured from the dmesg log might fit into a pstore
record of a given size after compression. The original implementation
used the computation given below for zlib:
switch (size) {
/* buffer range for efivars */
case 1000 ... 2000:
cmpr = 56;
break;
case 2001 ... 3000:
cmpr = 54;
break;
case 3001 ... 3999:
cmpr = 52;
break;
/* buffer range for nvram, erst */
case 4000 ... 10000:
cmpr = 45;
break;
default:
cmpr = 60;
break;
}
return (size * 100) / cmpr;
We will use the previous worst-case of 60% for compression. For
decompression go extra large (3x) so we make sure there's enough space
for anything.
While at it, rate limit the error message so we don't flood the log
unnecessarily on systems that have accumulated a lot of pstore history.
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Eric Biggers <ebiggers@kernel.org>
Cc: Kees Cook <keescook@chromium.org>
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Link: https://lore.kernel.org/r/20230830212238.135900-1-ardb@kernel.org
Co-developed-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Kees Cook <keescook@chromium.org>
2023-08-30 23:22:38 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2013-08-16 13:53:10 -07:00
|
|
|
|
pstore: Base compression input buffer size on estimated compressed size
Commit 1756ddea6916 ("pstore: Remove worst-case compression size logic")
removed some clunky per-algorithm worst case size estimation routines on
the basis that we can always store pstore records uncompressed, and
these worst case estimations are about how much the size might
inadvertently *increase* due to encapsulation overhead when the input
cannot be compressed at all. So if compression results in a size
increase, we just store the original data instead.
However, it seems that the original code was misinterpreting these
calculations as an estimation of how much uncompressed data might fit
into a compressed buffer of a given size, and it was using the results
to consume the input data in larger chunks than the pstore record size,
relying on the compression to ensure that what ultimately gets stored
fits into the available space.
One result of this, as observed and reported by Linus, is that upgrading
to a newer kernel that includes the given commit may result in pstore
decompression errors reported in the kernel log. This is due to the fact
that the existing records may unexpectedly decompress to a size that is
larger than the pstore record size.
Another potential problem caused by this change is that we may
underutilize the fixed sized records on pstore backends such as ramoops.
And on pstore backends with variable sized records such as EFI, we will
end up creating many more entries than before to store the same amount
of compressed data.
So let's fix both issues, by bringing back the typical case estimation of
how much ASCII text captured from the dmesg log might fit into a pstore
record of a given size after compression. The original implementation
used the computation given below for zlib:
switch (size) {
/* buffer range for efivars */
case 1000 ... 2000:
cmpr = 56;
break;
case 2001 ... 3000:
cmpr = 54;
break;
case 3001 ... 3999:
cmpr = 52;
break;
/* buffer range for nvram, erst */
case 4000 ... 10000:
cmpr = 45;
break;
default:
cmpr = 60;
break;
}
return (size * 100) / cmpr;
We will use the previous worst-case of 60% for compression. For
decompression go extra large (3x) so we make sure there's enough space
for anything.
While at it, rate limit the error message so we don't flood the log
unnecessarily on systems that have accumulated a lot of pstore history.
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Eric Biggers <ebiggers@kernel.org>
Cc: Kees Cook <keescook@chromium.org>
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Link: https://lore.kernel.org/r/20230830212238.135900-1-ardb@kernel.org
Co-developed-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Kees Cook <keescook@chromium.org>
2023-08-30 23:22:38 +02:00
|
|
|
|
2013-08-16 13:53:10 -07:00
|
|
|
|
2011-03-18 15:33:43 -07:00
|
|
|
|
2010-12-28 14:25:21 -08:00
|
|
|
|
2011-03-18 15:33:43 -07:00
|
|
|
|
2010-12-28 14:25:21 -08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2018-11-03 16:38:16 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2020-05-05 21:36:15 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2022-03-14 19:59:53 +01:00
|
|
|
|
pstore: Avoid deadlock in panic and emergency-restart path
[Issue]
When pstore is in panic and emergency-restart paths, it may be blocked
in those paths because it simply takes spin_lock.
This is an example scenario which pstore may hang up in a panic path:
- cpuA grabs psinfo->buf_lock
- cpuB panics and calls smp_send_stop
- smp_send_stop sends IRQ to cpuA
- after 1 second, cpuB gives up on cpuA and sends an NMI instead
- cpuA is now in an NMI handler while still holding buf_lock
- cpuB is deadlocked
This case may happen if a firmware has a bug and
cpuA is stuck talking with it more than one second.
Also, this is a similar scenario in an emergency-restart path:
- cpuA grabs psinfo->buf_lock and stucks in a firmware
- cpuB kicks emergency-restart via either sysrq-b or hangcheck timer.
And then, cpuB is deadlocked by taking psinfo->buf_lock again.
[Solution]
This patch avoids the deadlocking issues in both panic and emergency_restart
paths by introducing a function, is_non_blocking_path(), to check if a cpu
can be blocked in current path.
With this patch, pstore is not blocked even if another cpu has
taken a spin_lock, in those paths by changing from spin_lock_irqsave
to spin_trylock_irqsave.
In addition, according to a comment of emergency_restart() in kernel/sys.c,
spin_lock shouldn't be taken in an emergency_restart path to avoid
deadlock. This patch fits the comment below.
<snip>
/**
* emergency_restart - reboot the system
*
* Without shutting down any hardware or taking any locks
* reboot the system. This is called when we know we are in
* trouble so this is our best effort to reboot. This is
* safe to call in interrupt context.
*/
void emergency_restart(void)
<snip>
Signed-off-by: Seiji Aguchi <seiji.aguchi@hds.com>
Acked-by: Don Zickus <dzickus@redhat.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
2013-01-11 18:09:41 +00:00
|
|
|
|
2022-03-14 19:59:53 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
pstore: Avoid deadlock in panic and emergency-restart path
[Issue]
When pstore is in panic and emergency-restart paths, it may be blocked
in those paths because it simply takes spin_lock.
This is an example scenario which pstore may hang up in a panic path:
- cpuA grabs psinfo->buf_lock
- cpuB panics and calls smp_send_stop
- smp_send_stop sends IRQ to cpuA
- after 1 second, cpuB gives up on cpuA and sends an NMI instead
- cpuA is now in an NMI handler while still holding buf_lock
- cpuB is deadlocked
This case may happen if a firmware has a bug and
cpuA is stuck talking with it more than one second.
Also, this is a similar scenario in an emergency-restart path:
- cpuA grabs psinfo->buf_lock and stucks in a firmware
- cpuB kicks emergency-restart via either sysrq-b or hangcheck timer.
And then, cpuB is deadlocked by taking psinfo->buf_lock again.
[Solution]
This patch avoids the deadlocking issues in both panic and emergency_restart
paths by introducing a function, is_non_blocking_path(), to check if a cpu
can be blocked in current path.
With this patch, pstore is not blocked even if another cpu has
taken a spin_lock, in those paths by changing from spin_lock_irqsave
to spin_trylock_irqsave.
In addition, according to a comment of emergency_restart() in kernel/sys.c,
spin_lock shouldn't be taken in an emergency_restart path to avoid
deadlock. This patch fits the comment below.
<snip>
/**
* emergency_restart - reboot the system
*
* Without shutting down any hardware or taking any locks
* reboot the system. This is called when we know we are in
* trouble so this is our best effort to reboot. This is
* safe to call in interrupt context.
*/
void emergency_restart(void)
<snip>
Signed-off-by: Seiji Aguchi <seiji.aguchi@hds.com>
Acked-by: Don Zickus <dzickus@redhat.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
2013-01-11 18:09:41 +00:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2022-03-14 19:59:53 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
pstore: Avoid deadlock in panic and emergency-restart path
[Issue]
When pstore is in panic and emergency-restart paths, it may be blocked
in those paths because it simply takes spin_lock.
This is an example scenario which pstore may hang up in a panic path:
- cpuA grabs psinfo->buf_lock
- cpuB panics and calls smp_send_stop
- smp_send_stop sends IRQ to cpuA
- after 1 second, cpuB gives up on cpuA and sends an NMI instead
- cpuA is now in an NMI handler while still holding buf_lock
- cpuB is deadlocked
This case may happen if a firmware has a bug and
cpuA is stuck talking with it more than one second.
Also, this is a similar scenario in an emergency-restart path:
- cpuA grabs psinfo->buf_lock and stucks in a firmware
- cpuB kicks emergency-restart via either sysrq-b or hangcheck timer.
And then, cpuB is deadlocked by taking psinfo->buf_lock again.
[Solution]
This patch avoids the deadlocking issues in both panic and emergency_restart
paths by introducing a function, is_non_blocking_path(), to check if a cpu
can be blocked in current path.
With this patch, pstore is not blocked even if another cpu has
taken a spin_lock, in those paths by changing from spin_lock_irqsave
to spin_trylock_irqsave.
In addition, according to a comment of emergency_restart() in kernel/sys.c,
spin_lock shouldn't be taken in an emergency_restart path to avoid
deadlock. This patch fits the comment below.
<snip>
/**
* emergency_restart - reboot the system
*
* Without shutting down any hardware or taking any locks
* reboot the system. This is called when we know we are in
* trouble so this is our best effort to reboot. This is
* safe to call in interrupt context.
*/
void emergency_restart(void)
<snip>
Signed-off-by: Seiji Aguchi <seiji.aguchi@hds.com>
Acked-by: Don Zickus <dzickus@redhat.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
2013-01-11 18:09:41 +00:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2016-02-18 22:04:22 +08:00
|
|
|
|
pstore: Use crypto compress API
In the pstore compression part, we use zlib/lzo/lz4/lz4hc/842
compression algorithm API to implement pstore compression backends. But
there are many repeat codes in these implementations. This patch uses
crypto compress API to simplify these codes.
1) rewrite allocate_buf_for_compression, free_buf_for_compression,
pstore_compress, pstore_decompress functions using crypto compress API.
2) drop compress, decompress, allocate, free functions in pstore_zbackend,
and add zbufsize function to get each different compress buffer size.
3) use late_initcall to call ramoops_init later, to make sure the crypto
subsystem has already initialized.
4) use 'unsigned int' type instead of 'size_t' in pstore_compress,
pstore_decompress functions' length arguments.
5) rename 'zlib' to 'deflate' to follow the crypto API's name convention.
Signed-off-by: Geliang Tang <geliangtang@gmail.com>
[kees: tweaked error messages on allocation failures and Kconfig help]
Signed-off-by: Kees Cook <keescook@chromium.org>
2018-03-09 18:51:07 +08:00
|
|
|
|
2016-02-18 22:04:22 +08:00
|
|
|
|
pstore: Replace crypto API compression with zlib_deflate library calls
Pstore supports compression using a variety of algorithms exposed by the
crypto API. This uses the deprecated comp (as opposed to scomp/acomp)
API, and so we should stop using that, and either move to the new API,
or switch to a different approach entirely.
Given that we only compress ASCII text in pstore, and considering that
this happens when the system is likely to be in an unstable state, the
flexibility that the complex crypto API provides does not outweigh its
impact on the risk that we might encounter additional problems when
trying to commit the kernel log contents to the pstore backend.
So let's switch [back] to the zlib deflate library API, and remove all
the complexity that really has no place in a low-level diagnostic
facility. Note that, while more modern compression algorithms have been
added to the kernel in recent years, the code size of zlib deflate is
substantially smaller than, e.g., zstd, while its performance in terms
of compression ratio is comparable for ASCII text, and speed is deemed
irrelevant in this context.
Note that this means that compressed pstore records may no longer be
accessible after a kernel upgrade, but this has never been part of the
contract. (The choice of compression algorithm is not stored in the
pstore records either)
Tested-by: "Guilherme G. Piccoli" <gpiccoli@igalia.com> # Steam Deck
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Eric Biggers <ebiggers@google.com>
Link: https://lore.kernel.org/r/20230712162332.2670437-3-ardb@kernel.org
Signed-off-by: Kees Cook <keescook@chromium.org>
2023-07-12 18:23:32 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
pstore: Use crypto compress API
In the pstore compression part, we use zlib/lzo/lz4/lz4hc/842
compression algorithm API to implement pstore compression backends. But
there are many repeat codes in these implementations. This patch uses
crypto compress API to simplify these codes.
1) rewrite allocate_buf_for_compression, free_buf_for_compression,
pstore_compress, pstore_decompress functions using crypto compress API.
2) drop compress, decompress, allocate, free functions in pstore_zbackend,
and add zbufsize function to get each different compress buffer size.
3) use late_initcall to call ramoops_init later, to make sure the crypto
subsystem has already initialized.
4) use 'unsigned int' type instead of 'size_t' in pstore_compress,
pstore_decompress functions' length arguments.
5) rename 'zlib' to 'deflate' to follow the crypto API's name convention.
Signed-off-by: Geliang Tang <geliangtang@gmail.com>
[kees: tweaked error messages on allocation failures and Kconfig help]
Signed-off-by: Kees Cook <keescook@chromium.org>
2018-03-09 18:51:07 +08:00
|
|
|
|
|
|
|
|
|
2021-02-18 12:15:47 +01:00
|
|
|
|
2020-07-06 19:37:36 -07:00
|
|
|
|
|
|
|
|
|
pstore: Replace crypto API compression with zlib_deflate library calls
Pstore supports compression using a variety of algorithms exposed by the
crypto API. This uses the deprecated comp (as opposed to scomp/acomp)
API, and so we should stop using that, and either move to the new API,
or switch to a different approach entirely.
Given that we only compress ASCII text in pstore, and considering that
this happens when the system is likely to be in an unstable state, the
flexibility that the complex crypto API provides does not outweigh its
impact on the risk that we might encounter additional problems when
trying to commit the kernel log contents to the pstore backend.
So let's switch [back] to the zlib deflate library API, and remove all
the complexity that really has no place in a low-level diagnostic
facility. Note that, while more modern compression algorithms have been
added to the kernel in recent years, the code size of zlib deflate is
substantially smaller than, e.g., zstd, while its performance in terms
of compression ratio is comparable for ASCII text, and speed is deemed
irrelevant in this context.
Note that this means that compressed pstore records may no longer be
accessible after a kernel upgrade, but this has never been part of the
contract. (The choice of compression algorithm is not stored in the
pstore records either)
Tested-by: "Guilherme G. Piccoli" <gpiccoli@igalia.com> # Steam Deck
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Eric Biggers <ebiggers@google.com>
Link: https://lore.kernel.org/r/20230712162332.2670437-3-ardb@kernel.org
Signed-off-by: Kees Cook <keescook@chromium.org>
2023-07-12 18:23:32 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
pstore: Use crypto compress API
In the pstore compression part, we use zlib/lzo/lz4/lz4hc/842
compression algorithm API to implement pstore compression backends. But
there are many repeat codes in these implementations. This patch uses
crypto compress API to simplify these codes.
1) rewrite allocate_buf_for_compression, free_buf_for_compression,
pstore_compress, pstore_decompress functions using crypto compress API.
2) drop compress, decompress, allocate, free functions in pstore_zbackend,
and add zbufsize function to get each different compress buffer size.
3) use late_initcall to call ramoops_init later, to make sure the crypto
subsystem has already initialized.
4) use 'unsigned int' type instead of 'size_t' in pstore_compress,
pstore_decompress functions' length arguments.
5) rename 'zlib' to 'deflate' to follow the crypto API's name convention.
Signed-off-by: Geliang Tang <geliangtang@gmail.com>
[kees: tweaked error messages on allocation failures and Kconfig help]
Signed-off-by: Kees Cook <keescook@chromium.org>
2018-03-09 18:51:07 +08:00
|
|
|
|
pstore: Replace crypto API compression with zlib_deflate library calls
Pstore supports compression using a variety of algorithms exposed by the
crypto API. This uses the deprecated comp (as opposed to scomp/acomp)
API, and so we should stop using that, and either move to the new API,
or switch to a different approach entirely.
Given that we only compress ASCII text in pstore, and considering that
this happens when the system is likely to be in an unstable state, the
flexibility that the complex crypto API provides does not outweigh its
impact on the risk that we might encounter additional problems when
trying to commit the kernel log contents to the pstore backend.
So let's switch [back] to the zlib deflate library API, and remove all
the complexity that really has no place in a low-level diagnostic
facility. Note that, while more modern compression algorithms have been
added to the kernel in recent years, the code size of zlib deflate is
substantially smaller than, e.g., zstd, while its performance in terms
of compression ratio is comparable for ASCII text, and speed is deemed
irrelevant in this context.
Note that this means that compressed pstore records may no longer be
accessible after a kernel upgrade, but this has never been part of the
contract. (The choice of compression algorithm is not stored in the
pstore records either)
Tested-by: "Guilherme G. Piccoli" <gpiccoli@igalia.com> # Steam Deck
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Eric Biggers <ebiggers@google.com>
Link: https://lore.kernel.org/r/20230712162332.2670437-3-ardb@kernel.org
Signed-off-by: Kees Cook <keescook@chromium.org>
2023-07-12 18:23:32 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2016-02-18 22:04:22 +08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
pstore: Base compression input buffer size on estimated compressed size
Commit 1756ddea6916 ("pstore: Remove worst-case compression size logic")
removed some clunky per-algorithm worst case size estimation routines on
the basis that we can always store pstore records uncompressed, and
these worst case estimations are about how much the size might
inadvertently *increase* due to encapsulation overhead when the input
cannot be compressed at all. So if compression results in a size
increase, we just store the original data instead.
However, it seems that the original code was misinterpreting these
calculations as an estimation of how much uncompressed data might fit
into a compressed buffer of a given size, and it was using the results
to consume the input data in larger chunks than the pstore record size,
relying on the compression to ensure that what ultimately gets stored
fits into the available space.
One result of this, as observed and reported by Linus, is that upgrading
to a newer kernel that includes the given commit may result in pstore
decompression errors reported in the kernel log. This is due to the fact
that the existing records may unexpectedly decompress to a size that is
larger than the pstore record size.
Another potential problem caused by this change is that we may
underutilize the fixed sized records on pstore backends such as ramoops.
And on pstore backends with variable sized records such as EFI, we will
end up creating many more entries than before to store the same amount
of compressed data.
So let's fix both issues, by bringing back the typical case estimation of
how much ASCII text captured from the dmesg log might fit into a pstore
record of a given size after compression. The original implementation
used the computation given below for zlib:
switch (size) {
/* buffer range for efivars */
case 1000 ... 2000:
cmpr = 56;
break;
case 2001 ... 3000:
cmpr = 54;
break;
case 3001 ... 3999:
cmpr = 52;
break;
/* buffer range for nvram, erst */
case 4000 ... 10000:
cmpr = 45;
break;
default:
cmpr = 60;
break;
}
return (size * 100) / cmpr;
We will use the previous worst-case of 60% for compression. For
decompression go extra large (3x) so we make sure there's enough space
for anything.
While at it, rate limit the error message so we don't flood the log
unnecessarily on systems that have accumulated a lot of pstore history.
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Eric Biggers <ebiggers@kernel.org>
Cc: Kees Cook <keescook@chromium.org>
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Link: https://lore.kernel.org/r/20230830212238.135900-1-ardb@kernel.org
Co-developed-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Kees Cook <keescook@chromium.org>
2023-08-30 23:22:38 +02:00
|
|
|
|
2018-10-17 14:00:12 -07:00
|
|
|
|
|
|
|
|
|
pstore: Replace crypto API compression with zlib_deflate library calls
Pstore supports compression using a variety of algorithms exposed by the
crypto API. This uses the deprecated comp (as opposed to scomp/acomp)
API, and so we should stop using that, and either move to the new API,
or switch to a different approach entirely.
Given that we only compress ASCII text in pstore, and considering that
this happens when the system is likely to be in an unstable state, the
flexibility that the complex crypto API provides does not outweigh its
impact on the risk that we might encounter additional problems when
trying to commit the kernel log contents to the pstore backend.
So let's switch [back] to the zlib deflate library API, and remove all
the complexity that really has no place in a low-level diagnostic
facility. Note that, while more modern compression algorithms have been
added to the kernel in recent years, the code size of zlib deflate is
substantially smaller than, e.g., zstd, while its performance in terms
of compression ratio is comparable for ASCII text, and speed is deemed
irrelevant in this context.
Note that this means that compressed pstore records may no longer be
accessible after a kernel upgrade, but this has never been part of the
contract. (The choice of compression algorithm is not stored in the
pstore records either)
Tested-by: "Guilherme G. Piccoli" <gpiccoli@igalia.com> # Steam Deck
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Eric Biggers <ebiggers@google.com>
Link: https://lore.kernel.org/r/20230712162332.2670437-3-ardb@kernel.org
Signed-off-by: Kees Cook <keescook@chromium.org>
2023-07-12 18:23:32 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
pstore: Use crypto compress API
In the pstore compression part, we use zlib/lzo/lz4/lz4hc/842
compression algorithm API to implement pstore compression backends. But
there are many repeat codes in these implementations. This patch uses
crypto compress API to simplify these codes.
1) rewrite allocate_buf_for_compression, free_buf_for_compression,
pstore_compress, pstore_decompress functions using crypto compress API.
2) drop compress, decompress, allocate, free functions in pstore_zbackend,
and add zbufsize function to get each different compress buffer size.
3) use late_initcall to call ramoops_init later, to make sure the crypto
subsystem has already initialized.
4) use 'unsigned int' type instead of 'size_t' in pstore_compress,
pstore_decompress functions' length arguments.
5) rename 'zlib' to 'deflate' to follow the crypto API's name convention.
Signed-off-by: Geliang Tang <geliangtang@gmail.com>
[kees: tweaked error messages on allocation failures and Kconfig help]
Signed-off-by: Kees Cook <keescook@chromium.org>
2018-03-09 18:51:07 +08:00
|
|
|
|
pstore: Replace crypto API compression with zlib_deflate library calls
Pstore supports compression using a variety of algorithms exposed by the
crypto API. This uses the deprecated comp (as opposed to scomp/acomp)
API, and so we should stop using that, and either move to the new API,
or switch to a different approach entirely.
Given that we only compress ASCII text in pstore, and considering that
this happens when the system is likely to be in an unstable state, the
flexibility that the complex crypto API provides does not outweigh its
impact on the risk that we might encounter additional problems when
trying to commit the kernel log contents to the pstore backend.
So let's switch [back] to the zlib deflate library API, and remove all
the complexity that really has no place in a low-level diagnostic
facility. Note that, while more modern compression algorithms have been
added to the kernel in recent years, the code size of zlib deflate is
substantially smaller than, e.g., zstd, while its performance in terms
of compression ratio is comparable for ASCII text, and speed is deemed
irrelevant in this context.
Note that this means that compressed pstore records may no longer be
accessible after a kernel upgrade, but this has never been part of the
contract. (The choice of compression algorithm is not stored in the
pstore records either)
Tested-by: "Guilherme G. Piccoli" <gpiccoli@igalia.com> # Steam Deck
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Eric Biggers <ebiggers@google.com>
Link: https://lore.kernel.org/r/20230712162332.2670437-3-ardb@kernel.org
Signed-off-by: Kees Cook <keescook@chromium.org>
2023-07-12 18:23:32 +02:00
|
|
|
|
pstore: Use crypto compress API
In the pstore compression part, we use zlib/lzo/lz4/lz4hc/842
compression algorithm API to implement pstore compression backends. But
there are many repeat codes in these implementations. This patch uses
crypto compress API to simplify these codes.
1) rewrite allocate_buf_for_compression, free_buf_for_compression,
pstore_compress, pstore_decompress functions using crypto compress API.
2) drop compress, decompress, allocate, free functions in pstore_zbackend,
and add zbufsize function to get each different compress buffer size.
3) use late_initcall to call ramoops_init later, to make sure the crypto
subsystem has already initialized.
4) use 'unsigned int' type instead of 'size_t' in pstore_compress,
pstore_decompress functions' length arguments.
5) rename 'zlib' to 'deflate' to follow the crypto API's name convention.
Signed-off-by: Geliang Tang <geliangtang@gmail.com>
[kees: tweaked error messages on allocation failures and Kconfig help]
Signed-off-by: Kees Cook <keescook@chromium.org>
2018-03-09 18:51:07 +08:00
|
|
|
|
pstore: Replace crypto API compression with zlib_deflate library calls
Pstore supports compression using a variety of algorithms exposed by the
crypto API. This uses the deprecated comp (as opposed to scomp/acomp)
API, and so we should stop using that, and either move to the new API,
or switch to a different approach entirely.
Given that we only compress ASCII text in pstore, and considering that
this happens when the system is likely to be in an unstable state, the
flexibility that the complex crypto API provides does not outweigh its
impact on the risk that we might encounter additional problems when
trying to commit the kernel log contents to the pstore backend.
So let's switch [back] to the zlib deflate library API, and remove all
the complexity that really has no place in a low-level diagnostic
facility. Note that, while more modern compression algorithms have been
added to the kernel in recent years, the code size of zlib deflate is
substantially smaller than, e.g., zstd, while its performance in terms
of compression ratio is comparable for ASCII text, and speed is deemed
irrelevant in this context.
Note that this means that compressed pstore records may no longer be
accessible after a kernel upgrade, but this has never been part of the
contract. (The choice of compression algorithm is not stored in the
pstore records either)
Tested-by: "Guilherme G. Piccoli" <gpiccoli@igalia.com> # Steam Deck
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Eric Biggers <ebiggers@google.com>
Link: https://lore.kernel.org/r/20230712162332.2670437-3-ardb@kernel.org
Signed-off-by: Kees Cook <keescook@chromium.org>
2023-07-12 18:23:32 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
pstore: Use crypto compress API
In the pstore compression part, we use zlib/lzo/lz4/lz4hc/842
compression algorithm API to implement pstore compression backends. But
there are many repeat codes in these implementations. This patch uses
crypto compress API to simplify these codes.
1) rewrite allocate_buf_for_compression, free_buf_for_compression,
pstore_compress, pstore_decompress functions using crypto compress API.
2) drop compress, decompress, allocate, free functions in pstore_zbackend,
and add zbufsize function to get each different compress buffer size.
3) use late_initcall to call ramoops_init later, to make sure the crypto
subsystem has already initialized.
4) use 'unsigned int' type instead of 'size_t' in pstore_compress,
pstore_decompress functions' length arguments.
5) rename 'zlib' to 'deflate' to follow the crypto API's name convention.
Signed-off-by: Geliang Tang <geliangtang@gmail.com>
[kees: tweaked error messages on allocation failures and Kconfig help]
Signed-off-by: Kees Cook <keescook@chromium.org>
2018-03-09 18:51:07 +08:00
|
|
|
|
|
|
|
|
|
2023-07-12 18:23:31 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
pstore: Base compression input buffer size on estimated compressed size
Commit 1756ddea6916 ("pstore: Remove worst-case compression size logic")
removed some clunky per-algorithm worst case size estimation routines on
the basis that we can always store pstore records uncompressed, and
these worst case estimations are about how much the size might
inadvertently *increase* due to encapsulation overhead when the input
cannot be compressed at all. So if compression results in a size
increase, we just store the original data instead.
However, it seems that the original code was misinterpreting these
calculations as an estimation of how much uncompressed data might fit
into a compressed buffer of a given size, and it was using the results
to consume the input data in larger chunks than the pstore record size,
relying on the compression to ensure that what ultimately gets stored
fits into the available space.
One result of this, as observed and reported by Linus, is that upgrading
to a newer kernel that includes the given commit may result in pstore
decompression errors reported in the kernel log. This is due to the fact
that the existing records may unexpectedly decompress to a size that is
larger than the pstore record size.
Another potential problem caused by this change is that we may
underutilize the fixed sized records on pstore backends such as ramoops.
And on pstore backends with variable sized records such as EFI, we will
end up creating many more entries than before to store the same amount
of compressed data.
So let's fix both issues, by bringing back the typical case estimation of
how much ASCII text captured from the dmesg log might fit into a pstore
record of a given size after compression. The original implementation
used the computation given below for zlib:
switch (size) {
/* buffer range for efivars */
case 1000 ... 2000:
cmpr = 56;
break;
case 2001 ... 3000:
cmpr = 54;
break;
case 3001 ... 3999:
cmpr = 52;
break;
/* buffer range for nvram, erst */
case 4000 ... 10000:
cmpr = 45;
break;
default:
cmpr = 60;
break;
}
return (size * 100) / cmpr;
We will use the previous worst-case of 60% for compression. For
decompression go extra large (3x) so we make sure there's enough space
for anything.
While at it, rate limit the error message so we don't flood the log
unnecessarily on systems that have accumulated a lot of pstore history.
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Eric Biggers <ebiggers@kernel.org>
Cc: Kees Cook <keescook@chromium.org>
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Link: https://lore.kernel.org/r/20230830212238.135900-1-ardb@kernel.org
Co-developed-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Kees Cook <keescook@chromium.org>
2023-08-30 23:22:38 +02:00
|
|
|
|
|
|
|
|
|
2018-10-17 14:00:12 -07:00
|
|
|
|
2023-07-12 18:23:31 +02:00
|
|
|
|
|
|
|
|
|
pstore: Use crypto compress API
In the pstore compression part, we use zlib/lzo/lz4/lz4hc/842
compression algorithm API to implement pstore compression backends. But
there are many repeat codes in these implementations. This patch uses
crypto compress API to simplify these codes.
1) rewrite allocate_buf_for_compression, free_buf_for_compression,
pstore_compress, pstore_decompress functions using crypto compress API.
2) drop compress, decompress, allocate, free functions in pstore_zbackend,
and add zbufsize function to get each different compress buffer size.
3) use late_initcall to call ramoops_init later, to make sure the crypto
subsystem has already initialized.
4) use 'unsigned int' type instead of 'size_t' in pstore_compress,
pstore_decompress functions' length arguments.
5) rename 'zlib' to 'deflate' to follow the crypto API's name convention.
Signed-off-by: Geliang Tang <geliangtang@gmail.com>
[kees: tweaked error messages on allocation failures and Kconfig help]
Signed-off-by: Kees Cook <keescook@chromium.org>
2018-03-09 18:51:07 +08:00
|
|
|
|
2016-02-18 22:04:22 +08:00
|
|
|
|
2018-10-17 14:00:12 -07:00
|
|
|
|
pstore: Replace crypto API compression with zlib_deflate library calls
Pstore supports compression using a variety of algorithms exposed by the
crypto API. This uses the deprecated comp (as opposed to scomp/acomp)
API, and so we should stop using that, and either move to the new API,
or switch to a different approach entirely.
Given that we only compress ASCII text in pstore, and considering that
this happens when the system is likely to be in an unstable state, the
flexibility that the complex crypto API provides does not outweigh its
impact on the risk that we might encounter additional problems when
trying to commit the kernel log contents to the pstore backend.
So let's switch [back] to the zlib deflate library API, and remove all
the complexity that really has no place in a low-level diagnostic
facility. Note that, while more modern compression algorithms have been
added to the kernel in recent years, the code size of zlib deflate is
substantially smaller than, e.g., zstd, while its performance in terms
of compression ratio is comparable for ASCII text, and speed is deemed
irrelevant in this context.
Note that this means that compressed pstore records may no longer be
accessible after a kernel upgrade, but this has never been part of the
contract. (The choice of compression algorithm is not stored in the
pstore records either)
Tested-by: "Guilherme G. Piccoli" <gpiccoli@igalia.com> # Steam Deck
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Eric Biggers <ebiggers@google.com>
Link: https://lore.kernel.org/r/20230712162332.2670437-3-ardb@kernel.org
Signed-off-by: Kees Cook <keescook@chromium.org>
2023-07-12 18:23:32 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-06-27 13:25:41 -07:00
|
|
|
|
2018-10-17 14:00:12 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
pstore: Base compression input buffer size on estimated compressed size
Commit 1756ddea6916 ("pstore: Remove worst-case compression size logic")
removed some clunky per-algorithm worst case size estimation routines on
the basis that we can always store pstore records uncompressed, and
these worst case estimations are about how much the size might
inadvertently *increase* due to encapsulation overhead when the input
cannot be compressed at all. So if compression results in a size
increase, we just store the original data instead.
However, it seems that the original code was misinterpreting these
calculations as an estimation of how much uncompressed data might fit
into a compressed buffer of a given size, and it was using the results
to consume the input data in larger chunks than the pstore record size,
relying on the compression to ensure that what ultimately gets stored
fits into the available space.
One result of this, as observed and reported by Linus, is that upgrading
to a newer kernel that includes the given commit may result in pstore
decompression errors reported in the kernel log. This is due to the fact
that the existing records may unexpectedly decompress to a size that is
larger than the pstore record size.
Another potential problem caused by this change is that we may
underutilize the fixed sized records on pstore backends such as ramoops.
And on pstore backends with variable sized records such as EFI, we will
end up creating many more entries than before to store the same amount
of compressed data.
So let's fix both issues, by bringing back the typical case estimation of
how much ASCII text captured from the dmesg log might fit into a pstore
record of a given size after compression. The original implementation
used the computation given below for zlib:
switch (size) {
/* buffer range for efivars */
case 1000 ... 2000:
cmpr = 56;
break;
case 2001 ... 3000:
cmpr = 54;
break;
case 3001 ... 3999:
cmpr = 52;
break;
/* buffer range for nvram, erst */
case 4000 ... 10000:
cmpr = 45;
break;
default:
cmpr = 60;
break;
}
return (size * 100) / cmpr;
We will use the previous worst-case of 60% for compression. For
decompression go extra large (3x) so we make sure there's enough space
for anything.
While at it, rate limit the error message so we don't flood the log
unnecessarily on systems that have accumulated a lot of pstore history.
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Eric Biggers <ebiggers@kernel.org>
Cc: Kees Cook <keescook@chromium.org>
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Link: https://lore.kernel.org/r/20230830212238.135900-1-ardb@kernel.org
Co-developed-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Kees Cook <keescook@chromium.org>
2023-08-30 23:22:38 +02:00
|
|
|
|
2018-10-17 14:00:12 -07:00
|
|
|
|
2023-07-12 18:23:31 +02:00
|
|
|
|
2016-02-18 22:04:22 +08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
pstore: Replace crypto API compression with zlib_deflate library calls
Pstore supports compression using a variety of algorithms exposed by the
crypto API. This uses the deprecated comp (as opposed to scomp/acomp)
API, and so we should stop using that, and either move to the new API,
or switch to a different approach entirely.
Given that we only compress ASCII text in pstore, and considering that
this happens when the system is likely to be in an unstable state, the
flexibility that the complex crypto API provides does not outweigh its
impact on the risk that we might encounter additional problems when
trying to commit the kernel log contents to the pstore backend.
So let's switch [back] to the zlib deflate library API, and remove all
the complexity that really has no place in a low-level diagnostic
facility. Note that, while more modern compression algorithms have been
added to the kernel in recent years, the code size of zlib deflate is
substantially smaller than, e.g., zstd, while its performance in terms
of compression ratio is comparable for ASCII text, and speed is deemed
irrelevant in this context.
Note that this means that compressed pstore records may no longer be
accessible after a kernel upgrade, but this has never been part of the
contract. (The choice of compression algorithm is not stored in the
pstore records either)
Tested-by: "Guilherme G. Piccoli" <gpiccoli@igalia.com> # Steam Deck
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Eric Biggers <ebiggers@google.com>
Link: https://lore.kernel.org/r/20230712162332.2670437-3-ardb@kernel.org
Signed-off-by: Kees Cook <keescook@chromium.org>
2023-07-12 18:23:32 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-05-20 14:51:19 +08:00
|
|
|
|
pstore: Replace crypto API compression with zlib_deflate library calls
Pstore supports compression using a variety of algorithms exposed by the
crypto API. This uses the deprecated comp (as opposed to scomp/acomp)
API, and so we should stop using that, and either move to the new API,
or switch to a different approach entirely.
Given that we only compress ASCII text in pstore, and considering that
this happens when the system is likely to be in an unstable state, the
flexibility that the complex crypto API provides does not outweigh its
impact on the risk that we might encounter additional problems when
trying to commit the kernel log contents to the pstore backend.
So let's switch [back] to the zlib deflate library API, and remove all
the complexity that really has no place in a low-level diagnostic
facility. Note that, while more modern compression algorithms have been
added to the kernel in recent years, the code size of zlib deflate is
substantially smaller than, e.g., zstd, while its performance in terms
of compression ratio is comparable for ASCII text, and speed is deemed
irrelevant in this context.
Note that this means that compressed pstore records may no longer be
accessible after a kernel upgrade, but this has never been part of the
contract. (The choice of compression algorithm is not stored in the
pstore records either)
Tested-by: "Guilherme G. Piccoli" <gpiccoli@igalia.com> # Steam Deck
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Eric Biggers <ebiggers@google.com>
Link: https://lore.kernel.org/r/20230712162332.2670437-3-ardb@kernel.org
Signed-off-by: Kees Cook <keescook@chromium.org>
2023-07-12 18:23:32 +02:00
|
|
|
|
2023-06-27 13:25:41 -07:00
|
|
|
|
pstore: Use crypto compress API
In the pstore compression part, we use zlib/lzo/lz4/lz4hc/842
compression algorithm API to implement pstore compression backends. But
there are many repeat codes in these implementations. This patch uses
crypto compress API to simplify these codes.
1) rewrite allocate_buf_for_compression, free_buf_for_compression,
pstore_compress, pstore_decompress functions using crypto compress API.
2) drop compress, decompress, allocate, free functions in pstore_zbackend,
and add zbufsize function to get each different compress buffer size.
3) use late_initcall to call ramoops_init later, to make sure the crypto
subsystem has already initialized.
4) use 'unsigned int' type instead of 'size_t' in pstore_compress,
pstore_decompress functions' length arguments.
5) rename 'zlib' to 'deflate' to follow the crypto API's name convention.
Signed-off-by: Geliang Tang <geliangtang@gmail.com>
[kees: tweaked error messages on allocation failures and Kconfig help]
Signed-off-by: Kees Cook <keescook@chromium.org>
2018-03-09 18:51:07 +08:00
|
|
|
|
pstore: Base compression input buffer size on estimated compressed size
Commit 1756ddea6916 ("pstore: Remove worst-case compression size logic")
removed some clunky per-algorithm worst case size estimation routines on
the basis that we can always store pstore records uncompressed, and
these worst case estimations are about how much the size might
inadvertently *increase* due to encapsulation overhead when the input
cannot be compressed at all. So if compression results in a size
increase, we just store the original data instead.
However, it seems that the original code was misinterpreting these
calculations as an estimation of how much uncompressed data might fit
into a compressed buffer of a given size, and it was using the results
to consume the input data in larger chunks than the pstore record size,
relying on the compression to ensure that what ultimately gets stored
fits into the available space.
One result of this, as observed and reported by Linus, is that upgrading
to a newer kernel that includes the given commit may result in pstore
decompression errors reported in the kernel log. This is due to the fact
that the existing records may unexpectedly decompress to a size that is
larger than the pstore record size.
Another potential problem caused by this change is that we may
underutilize the fixed sized records on pstore backends such as ramoops.
And on pstore backends with variable sized records such as EFI, we will
end up creating many more entries than before to store the same amount
of compressed data.
So let's fix both issues, by bringing back the typical case estimation of
how much ASCII text captured from the dmesg log might fit into a pstore
record of a given size after compression. The original implementation
used the computation given below for zlib:
switch (size) {
/* buffer range for efivars */
case 1000 ... 2000:
cmpr = 56;
break;
case 2001 ... 3000:
cmpr = 54;
break;
case 3001 ... 3999:
cmpr = 52;
break;
/* buffer range for nvram, erst */
case 4000 ... 10000:
cmpr = 45;
break;
default:
cmpr = 60;
break;
}
return (size * 100) / cmpr;
We will use the previous worst-case of 60% for compression. For
decompression go extra large (3x) so we make sure there's enough space
for anything.
While at it, rate limit the error message so we don't flood the log
unnecessarily on systems that have accumulated a lot of pstore history.
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Eric Biggers <ebiggers@kernel.org>
Cc: Kees Cook <keescook@chromium.org>
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Link: https://lore.kernel.org/r/20230830212238.135900-1-ardb@kernel.org
Co-developed-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Kees Cook <keescook@chromium.org>
2023-08-30 23:22:38 +02:00
|
|
|
|
2013-08-16 13:53:10 -07:00
|
|
|
|
|
|
|
|
|
2017-05-19 15:10:31 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2017-05-19 15:29:10 -07:00
|
|
|
|
|
|
|
|
|
2018-05-14 15:50:52 -07:00
|
|
|
|
2017-05-19 15:10:31 -07:00
|
|
|
|
|
|
|
|
|
2010-12-28 14:25:21 -08:00
|
|
|
|
2018-11-01 14:03:07 -07:00
|
|
|
|
|
|
|
|
|
2010-12-28 14:25:21 -08:00
|
|
|
|
|
|
|
|
|
2024-07-02 14:26:04 +02:00
|
|
|
|
2010-12-28 14:25:21 -08:00
|
|
|
|
2021-03-03 11:15:25 +01:00
|
|
|
|
2012-06-15 14:07:51 +02:00
|
|
|
|
2012-03-16 15:36:59 -07:00
|
|
|
|
2011-07-21 16:57:54 -04:00
|
|
|
|
2022-03-14 19:59:53 +01:00
|
|
|
|
2022-10-13 18:06:46 -03:00
|
|
|
|
2012-06-15 14:07:51 +02:00
|
|
|
|
2010-12-28 14:25:21 -08:00
|
|
|
|
2024-07-02 14:26:04 +02:00
|
|
|
|
2011-03-22 16:01:49 -07:00
|
|
|
|
2024-07-02 14:26:04 +02:00
|
|
|
|
2024-08-19 22:59:45 +08:00
|
|
|
|
2022-03-14 19:59:53 +01:00
|
|
|
|
|
|
|
|
|
2016-11-05 10:15:59 +08:00
|
|
|
|
pstore: Avoid deadlock in panic and emergency-restart path
[Issue]
When pstore is in panic and emergency-restart paths, it may be blocked
in those paths because it simply takes spin_lock.
This is an example scenario which pstore may hang up in a panic path:
- cpuA grabs psinfo->buf_lock
- cpuB panics and calls smp_send_stop
- smp_send_stop sends IRQ to cpuA
- after 1 second, cpuB gives up on cpuA and sends an NMI instead
- cpuA is now in an NMI handler while still holding buf_lock
- cpuB is deadlocked
This case may happen if a firmware has a bug and
cpuA is stuck talking with it more than one second.
Also, this is a similar scenario in an emergency-restart path:
- cpuA grabs psinfo->buf_lock and stucks in a firmware
- cpuB kicks emergency-restart via either sysrq-b or hangcheck timer.
And then, cpuB is deadlocked by taking psinfo->buf_lock again.
[Solution]
This patch avoids the deadlocking issues in both panic and emergency_restart
paths by introducing a function, is_non_blocking_path(), to check if a cpu
can be blocked in current path.
With this patch, pstore is not blocked even if another cpu has
taken a spin_lock, in those paths by changing from spin_lock_irqsave
to spin_trylock_irqsave.
In addition, according to a comment of emergency_restart() in kernel/sys.c,
spin_lock shouldn't be taken in an emergency_restart path to avoid
deadlock. This patch fits the comment below.
<snip>
/**
* emergency_restart - reboot the system
*
* Without shutting down any hardware or taking any locks
* reboot the system. This is called when we know we are in
* trouble so this is our best effort to reboot. This is
* safe to call in interrupt context.
*/
void emergency_restart(void)
<snip>
Signed-off-by: Seiji Aguchi <seiji.aguchi@hds.com>
Acked-by: Don Zickus <dzickus@redhat.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
2013-01-11 18:09:41 +00:00
|
|
|
|
2022-03-14 19:59:53 +01:00
|
|
|
|
2024-08-19 22:59:45 +08:00
|
|
|
|
2016-05-18 21:00:05 +09:00
|
|
|
|
2018-11-30 14:36:58 -08:00
|
|
|
|
2021-03-03 11:15:25 +01:00
|
|
|
|
|
|
|
|
|
2010-12-28 14:25:21 -08:00
|
|
|
|
|
|
|
|
|
2012-06-15 14:07:51 +02:00
|
|
|
|
2017-03-03 23:28:53 -08:00
|
|
|
|
|
|
|
|
|
2013-08-16 13:53:10 -07:00
|
|
|
|
2017-03-03 23:28:53 -08:00
|
|
|
|
2017-05-19 15:10:31 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-07-02 14:26:04 +02:00
|
|
|
|
2017-05-19 15:10:31 -07:00
|
|
|
|
|
|
|
|
|
2012-06-15 14:07:51 +02:00
|
|
|
|
2023-07-12 18:23:31 +02:00
|
|
|
|
pstore: Base compression input buffer size on estimated compressed size
Commit 1756ddea6916 ("pstore: Remove worst-case compression size logic")
removed some clunky per-algorithm worst case size estimation routines on
the basis that we can always store pstore records uncompressed, and
these worst case estimations are about how much the size might
inadvertently *increase* due to encapsulation overhead when the input
cannot be compressed at all. So if compression results in a size
increase, we just store the original data instead.
However, it seems that the original code was misinterpreting these
calculations as an estimation of how much uncompressed data might fit
into a compressed buffer of a given size, and it was using the results
to consume the input data in larger chunks than the pstore record size,
relying on the compression to ensure that what ultimately gets stored
fits into the available space.
One result of this, as observed and reported by Linus, is that upgrading
to a newer kernel that includes the given commit may result in pstore
decompression errors reported in the kernel log. This is due to the fact
that the existing records may unexpectedly decompress to a size that is
larger than the pstore record size.
Another potential problem caused by this change is that we may
underutilize the fixed sized records on pstore backends such as ramoops.
And on pstore backends with variable sized records such as EFI, we will
end up creating many more entries than before to store the same amount
of compressed data.
So let's fix both issues, by bringing back the typical case estimation of
how much ASCII text captured from the dmesg log might fit into a pstore
record of a given size after compression. The original implementation
used the computation given below for zlib:
switch (size) {
/* buffer range for efivars */
case 1000 ... 2000:
cmpr = 56;
break;
case 2001 ... 3000:
cmpr = 54;
break;
case 3001 ... 3999:
cmpr = 52;
break;
/* buffer range for nvram, erst */
case 4000 ... 10000:
cmpr = 45;
break;
default:
cmpr = 60;
break;
}
return (size * 100) / cmpr;
We will use the previous worst-case of 60% for compression. For
decompression go extra large (3x) so we make sure there's enough space
for anything.
While at it, rate limit the error message so we don't flood the log
unnecessarily on systems that have accumulated a lot of pstore history.
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Eric Biggers <ebiggers@kernel.org>
Cc: Kees Cook <keescook@chromium.org>
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Link: https://lore.kernel.org/r/20230830212238.135900-1-ardb@kernel.org
Co-developed-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Kees Cook <keescook@chromium.org>
2023-08-30 23:22:38 +02:00
|
|
|
|
2010-12-28 14:25:21 -08:00
|
|
|
|
2017-03-03 23:28:53 -08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2013-08-16 13:53:10 -07:00
|
|
|
|
2017-03-03 23:28:53 -08:00
|
|
|
|
2021-03-03 11:15:25 +01:00
|
|
|
|
2017-03-03 23:28:53 -08:00
|
|
|
|
2016-05-18 21:00:06 +09:00
|
|
|
|
|
|
|
|
|
2018-11-30 14:36:58 -08:00
|
|
|
|
2013-08-16 13:53:10 -07:00
|
|
|
|
2017-03-03 23:28:53 -08:00
|
|
|
|
|
|
|
|
|
2013-08-16 13:53:10 -07:00
|
|
|
|
|
|
|
|
|
2017-03-03 23:28:53 -08:00
|
|
|
|
|
|
|
|
|
2013-08-16 13:53:10 -07:00
|
|
|
|
pstore: Base compression input buffer size on estimated compressed size
Commit 1756ddea6916 ("pstore: Remove worst-case compression size logic")
removed some clunky per-algorithm worst case size estimation routines on
the basis that we can always store pstore records uncompressed, and
these worst case estimations are about how much the size might
inadvertently *increase* due to encapsulation overhead when the input
cannot be compressed at all. So if compression results in a size
increase, we just store the original data instead.
However, it seems that the original code was misinterpreting these
calculations as an estimation of how much uncompressed data might fit
into a compressed buffer of a given size, and it was using the results
to consume the input data in larger chunks than the pstore record size,
relying on the compression to ensure that what ultimately gets stored
fits into the available space.
One result of this, as observed and reported by Linus, is that upgrading
to a newer kernel that includes the given commit may result in pstore
decompression errors reported in the kernel log. This is due to the fact
that the existing records may unexpectedly decompress to a size that is
larger than the pstore record size.
Another potential problem caused by this change is that we may
underutilize the fixed sized records on pstore backends such as ramoops.
And on pstore backends with variable sized records such as EFI, we will
end up creating many more entries than before to store the same amount
of compressed data.
So let's fix both issues, by bringing back the typical case estimation of
how much ASCII text captured from the dmesg log might fit into a pstore
record of a given size after compression. The original implementation
used the computation given below for zlib:
switch (size) {
/* buffer range for efivars */
case 1000 ... 2000:
cmpr = 56;
break;
case 2001 ... 3000:
cmpr = 54;
break;
case 3001 ... 3999:
cmpr = 52;
break;
/* buffer range for nvram, erst */
case 4000 ... 10000:
cmpr = 45;
break;
default:
cmpr = 60;
break;
}
return (size * 100) / cmpr;
We will use the previous worst-case of 60% for compression. For
decompression go extra large (3x) so we make sure there's enough space
for anything.
While at it, rate limit the error message so we don't flood the log
unnecessarily on systems that have accumulated a lot of pstore history.
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Eric Biggers <ebiggers@kernel.org>
Cc: Kees Cook <keescook@chromium.org>
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Link: https://lore.kernel.org/r/20230830212238.135900-1-ardb@kernel.org
Co-developed-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Kees Cook <keescook@chromium.org>
2023-08-30 23:22:38 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2013-08-16 13:53:10 -07:00
|
|
|
|
|
|
|
|
|
2017-03-03 23:28:53 -08:00
|
|
|
|
2013-08-16 13:53:10 -07:00
|
|
|
|
2010-12-28 14:25:21 -08:00
|
|
|
|
2017-03-03 23:28:53 -08:00
|
|
|
|
2024-07-02 14:26:04 +02:00
|
|
|
|
2011-08-11 15:14:39 -07:00
|
|
|
|
2020-05-05 21:36:15 -07:00
|
|
|
|
2022-10-13 18:06:46 -03:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2020-05-05 21:36:15 -07:00
|
|
|
|
2012-06-15 14:07:51 +02:00
|
|
|
|
2017-03-03 23:28:53 -08:00
|
|
|
|
2011-07-21 16:57:53 -04:00
|
|
|
|
2010-12-28 14:25:21 -08:00
|
|
|
|
2024-08-19 22:59:45 +08:00
|
|
|
|
2022-10-13 18:06:46 -03:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2010-12-28 14:25:21 -08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2015-10-31 23:23:15 +08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2015-10-20 00:39:02 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2015-10-20 00:39:03 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2012-05-26 06:20:19 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2018-11-01 14:08:07 -07:00
|
|
|
|
2012-05-26 06:20:19 -07:00
|
|
|
|
2019-01-31 18:12:46 +08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2018-11-01 14:08:07 -07:00
|
|
|
|
|
|
|
|
|
2012-05-26 06:20:19 -07:00
|
|
|
|
2018-11-01 14:08:07 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2012-05-26 06:20:19 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2020-05-08 09:26:28 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2020-05-06 16:34:42 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2012-05-26 06:20:19 -07:00
|
|
|
|
|
|
|
|
|
2015-10-20 00:39:03 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2012-05-26 06:20:19 -07:00
|
|
|
|
|
|
|
|
|
2015-10-20 00:39:03 -07:00
|
|
|
|
2012-05-26 06:20:19 -07:00
|
|
|
|
|
|
|
|
|
2017-03-05 22:41:10 -08:00
|
|
|
|
|
|
|
|
|
2016-09-01 08:13:46 -07:00
|
|
|
|
2017-03-07 13:57:11 -08:00
|
|
|
|
2016-09-01 08:13:46 -07:00
|
|
|
|
2017-03-07 13:57:11 -08:00
|
|
|
|
|
|
|
|
|
2016-09-01 08:13:46 -07:00
|
|
|
|
2023-06-27 13:25:41 -07:00
|
|
|
|
2017-09-26 03:21:27 +09:00
|
|
|
|
2017-04-29 09:45:16 +08:00
|
|
|
|
2017-03-07 13:57:11 -08:00
|
|
|
|
2016-09-01 08:13:46 -07:00
|
|
|
|
2017-03-07 13:57:11 -08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-06-27 13:25:41 -07:00
|
|
|
|
2017-04-29 09:45:16 +08:00
|
|
|
|
2017-03-07 13:57:11 -08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2016-09-01 08:13:46 -07:00
|
|
|
|
|
|
|
|
|
2010-12-28 14:25:21 -08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-06-23 10:27:06 +08:00
|
|
|
|
|
|
|
|
|
2017-03-03 12:11:40 -08:00
|
|
|
|
2022-10-06 19:42:05 -03:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2017-03-03 12:11:40 -08:00
|
|
|
|
2013-06-28 17:11:33 -04:00
|
|
|
|
2017-03-05 22:41:10 -08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-06-23 10:27:06 +08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2020-05-04 19:31:36 -07:00
|
|
|
|
2010-12-28 14:25:21 -08:00
|
|
|
|
2017-03-03 12:11:40 -08:00
|
|
|
|
|
|
|
|
|
2020-05-04 19:31:36 -07:00
|
|
|
|
2023-06-23 10:27:06 +08:00
|
|
|
|
2010-12-28 14:25:21 -08:00
|
|
|
|
|
|
|
|
|
2011-07-21 16:57:55 -04:00
|
|
|
|
2017-03-05 22:41:10 -08:00
|
|
|
|
|
|
|
|
|
2010-12-28 14:25:21 -08:00
|
|
|
|
2011-11-17 12:58:07 -08:00
|
|
|
|
2024-08-19 22:59:45 +08:00
|
|
|
|
2010-12-28 14:25:21 -08:00
|
|
|
|
2019-05-30 23:37:29 -07:00
|
|
|
|
|
|
|
|
|
2013-08-16 13:53:10 -07:00
|
|
|
|
2020-05-04 19:43:41 -07:00
|
|
|
|
2010-12-28 14:25:21 -08:00
|
|
|
|
2020-05-05 11:45:07 -04:00
|
|
|
|
|
|
|
|
|
2016-07-28 00:08:25 +09:00
|
|
|
|
2020-05-05 11:45:07 -04:00
|
|
|
|
2016-07-28 00:08:25 +09:00
|
|
|
|
2013-12-18 15:17:10 -08:00
|
|
|
|
2016-07-28 00:08:25 +09:00
|
|
|
|
2013-12-18 15:17:10 -08:00
|
|
|
|
2016-07-28 00:08:25 +09:00
|
|
|
|
2015-01-16 16:01:10 -08:00
|
|
|
|
2010-12-28 14:25:21 -08:00
|
|
|
|
2017-03-06 12:42:12 -08:00
|
|
|
|
2020-05-05 21:36:15 -07:00
|
|
|
|
2011-08-11 15:14:39 -07:00
|
|
|
|
2015-05-21 09:34:22 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-06-23 10:27:06 +08:00
|
|
|
|
2015-05-21 09:34:22 -07:00
|
|
|
|
2014-06-06 14:37:31 -07:00
|
|
|
|
2013-06-28 17:11:33 -04:00
|
|
|
|
2020-05-04 19:33:54 -07:00
|
|
|
|
2010-12-28 14:25:21 -08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2015-10-20 00:39:03 -07:00
|
|
|
|
|
|
|
|
|
2020-05-04 19:33:54 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2020-05-05 21:36:15 -07:00
|
|
|
|
2016-07-28 00:08:25 +09:00
|
|
|
|
2016-05-19 10:59:03 -04:00
|
|
|
|
2016-07-28 00:08:25 +09:00
|
|
|
|
2016-05-19 10:59:03 -04:00
|
|
|
|
2016-07-28 00:08:25 +09:00
|
|
|
|
2016-05-19 10:59:03 -04:00
|
|
|
|
2016-07-28 00:08:25 +09:00
|
|
|
|
|
|
|
|
|
2015-10-20 00:39:03 -07:00
|
|
|
|
2020-05-05 21:36:15 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2020-05-04 19:46:53 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2015-10-20 00:39:03 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2020-05-08 09:16:02 -07:00
|
|
|
|
2015-10-20 00:39:03 -07:00
|
|
|
|
2022-10-06 19:42:07 -03:00
|
|
|
|
|
|
|
|
|
2020-05-04 19:33:54 -07:00
|
|
|
|
2015-10-20 00:39:03 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
pstore: Replace crypto API compression with zlib_deflate library calls
Pstore supports compression using a variety of algorithms exposed by the
crypto API. This uses the deprecated comp (as opposed to scomp/acomp)
API, and so we should stop using that, and either move to the new API,
or switch to a different approach entirely.
Given that we only compress ASCII text in pstore, and considering that
this happens when the system is likely to be in an unstable state, the
flexibility that the complex crypto API provides does not outweigh its
impact on the risk that we might encounter additional problems when
trying to commit the kernel log contents to the pstore backend.
So let's switch [back] to the zlib deflate library API, and remove all
the complexity that really has no place in a low-level diagnostic
facility. Note that, while more modern compression algorithms have been
added to the kernel in recent years, the code size of zlib deflate is
substantially smaller than, e.g., zstd, while its performance in terms
of compression ratio is comparable for ASCII text, and speed is deemed
irrelevant in this context.
Note that this means that compressed pstore records may no longer be
accessible after a kernel upgrade, but this has never been part of the
contract. (The choice of compression algorithm is not stored in the
pstore records either)
Tested-by: "Guilherme G. Piccoli" <gpiccoli@igalia.com> # Steam Deck
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Eric Biggers <ebiggers@google.com>
Link: https://lore.kernel.org/r/20230712162332.2670437-3-ardb@kernel.org
Signed-off-by: Kees Cook <keescook@chromium.org>
2023-07-12 18:23:32 +02:00
|
|
|
|
|
|
|
|
|
2017-03-03 17:35:25 -08:00
|
|
|
|
2018-10-26 01:17:07 -07:00
|
|
|
|
2017-03-03 17:35:25 -08:00
|
|
|
|
2018-10-26 01:17:07 -07:00
|
|
|
|
pstore: Base compression input buffer size on estimated compressed size
Commit 1756ddea6916 ("pstore: Remove worst-case compression size logic")
removed some clunky per-algorithm worst case size estimation routines on
the basis that we can always store pstore records uncompressed, and
these worst case estimations are about how much the size might
inadvertently *increase* due to encapsulation overhead when the input
cannot be compressed at all. So if compression results in a size
increase, we just store the original data instead.
However, it seems that the original code was misinterpreting these
calculations as an estimation of how much uncompressed data might fit
into a compressed buffer of a given size, and it was using the results
to consume the input data in larger chunks than the pstore record size,
relying on the compression to ensure that what ultimately gets stored
fits into the available space.
One result of this, as observed and reported by Linus, is that upgrading
to a newer kernel that includes the given commit may result in pstore
decompression errors reported in the kernel log. This is due to the fact
that the existing records may unexpectedly decompress to a size that is
larger than the pstore record size.
Another potential problem caused by this change is that we may
underutilize the fixed sized records on pstore backends such as ramoops.
And on pstore backends with variable sized records such as EFI, we will
end up creating many more entries than before to store the same amount
of compressed data.
So let's fix both issues, by bringing back the typical case estimation of
how much ASCII text captured from the dmesg log might fit into a pstore
record of a given size after compression. The original implementation
used the computation given below for zlib:
switch (size) {
/* buffer range for efivars */
case 1000 ... 2000:
cmpr = 56;
break;
case 2001 ... 3000:
cmpr = 54;
break;
case 3001 ... 3999:
cmpr = 52;
break;
/* buffer range for nvram, erst */
case 4000 ... 10000:
cmpr = 45;
break;
default:
cmpr = 60;
break;
}
return (size * 100) / cmpr;
We will use the previous worst-case of 60% for compression. For
decompression go extra large (3x) so we make sure there's enough space
for anything.
While at it, rate limit the error message so we don't flood the log
unnecessarily on systems that have accumulated a lot of pstore history.
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Eric Biggers <ebiggers@kernel.org>
Cc: Kees Cook <keescook@chromium.org>
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Link: https://lore.kernel.org/r/20230830212238.135900-1-ardb@kernel.org
Co-developed-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Kees Cook <keescook@chromium.org>
2023-08-30 23:22:38 +02:00
|
|
|
|
2017-03-03 17:35:25 -08:00
|
|
|
|
2021-02-18 12:15:47 +01:00
|
|
|
|
2017-05-23 11:16:52 +05:30
|
|
|
|
|
|
|
|
|
2017-03-03 17:35:25 -08:00
|
|
|
|
2017-05-23 11:16:52 +05:30
|
|
|
|
2017-03-03 17:35:25 -08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2018-10-26 01:17:07 -07:00
|
|
|
|
pstore: Replace crypto API compression with zlib_deflate library calls
Pstore supports compression using a variety of algorithms exposed by the
crypto API. This uses the deprecated comp (as opposed to scomp/acomp)
API, and so we should stop using that, and either move to the new API,
or switch to a different approach entirely.
Given that we only compress ASCII text in pstore, and considering that
this happens when the system is likely to be in an unstable state, the
flexibility that the complex crypto API provides does not outweigh its
impact on the risk that we might encounter additional problems when
trying to commit the kernel log contents to the pstore backend.
So let's switch [back] to the zlib deflate library API, and remove all
the complexity that really has no place in a low-level diagnostic
facility. Note that, while more modern compression algorithms have been
added to the kernel in recent years, the code size of zlib deflate is
substantially smaller than, e.g., zstd, while its performance in terms
of compression ratio is comparable for ASCII text, and speed is deemed
irrelevant in this context.
Note that this means that compressed pstore records may no longer be
accessible after a kernel upgrade, but this has never been part of the
contract. (The choice of compression algorithm is not stored in the
pstore records either)
Tested-by: "Guilherme G. Piccoli" <gpiccoli@igalia.com> # Steam Deck
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Eric Biggers <ebiggers@google.com>
Link: https://lore.kernel.org/r/20230712162332.2670437-3-ardb@kernel.org
Signed-off-by: Kees Cook <keescook@chromium.org>
2023-07-12 18:23:32 +02:00
|
|
|
|
2018-10-26 01:17:07 -07:00
|
|
|
|
2017-03-03 17:35:25 -08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
pstore: Replace crypto API compression with zlib_deflate library calls
Pstore supports compression using a variety of algorithms exposed by the
crypto API. This uses the deprecated comp (as opposed to scomp/acomp)
API, and so we should stop using that, and either move to the new API,
or switch to a different approach entirely.
Given that we only compress ASCII text in pstore, and considering that
this happens when the system is likely to be in an unstable state, the
flexibility that the complex crypto API provides does not outweigh its
impact on the risk that we might encounter additional problems when
trying to commit the kernel log contents to the pstore backend.
So let's switch [back] to the zlib deflate library API, and remove all
the complexity that really has no place in a low-level diagnostic
facility. Note that, while more modern compression algorithms have been
added to the kernel in recent years, the code size of zlib deflate is
substantially smaller than, e.g., zstd, while its performance in terms
of compression ratio is comparable for ASCII text, and speed is deemed
irrelevant in this context.
Note that this means that compressed pstore records may no longer be
accessible after a kernel upgrade, but this has never been part of the
contract. (The choice of compression algorithm is not stored in the
pstore records either)
Tested-by: "Guilherme G. Piccoli" <gpiccoli@igalia.com> # Steam Deck
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Eric Biggers <ebiggers@google.com>
Link: https://lore.kernel.org/r/20230712162332.2670437-3-ardb@kernel.org
Signed-off-by: Kees Cook <keescook@chromium.org>
2023-07-12 18:23:32 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2018-10-26 01:17:07 -07:00
|
|
|
|
pstore: Base compression input buffer size on estimated compressed size
Commit 1756ddea6916 ("pstore: Remove worst-case compression size logic")
removed some clunky per-algorithm worst case size estimation routines on
the basis that we can always store pstore records uncompressed, and
these worst case estimations are about how much the size might
inadvertently *increase* due to encapsulation overhead when the input
cannot be compressed at all. So if compression results in a size
increase, we just store the original data instead.
However, it seems that the original code was misinterpreting these
calculations as an estimation of how much uncompressed data might fit
into a compressed buffer of a given size, and it was using the results
to consume the input data in larger chunks than the pstore record size,
relying on the compression to ensure that what ultimately gets stored
fits into the available space.
One result of this, as observed and reported by Linus, is that upgrading
to a newer kernel that includes the given commit may result in pstore
decompression errors reported in the kernel log. This is due to the fact
that the existing records may unexpectedly decompress to a size that is
larger than the pstore record size.
Another potential problem caused by this change is that we may
underutilize the fixed sized records on pstore backends such as ramoops.
And on pstore backends with variable sized records such as EFI, we will
end up creating many more entries than before to store the same amount
of compressed data.
So let's fix both issues, by bringing back the typical case estimation of
how much ASCII text captured from the dmesg log might fit into a pstore
record of a given size after compression. The original implementation
used the computation given below for zlib:
switch (size) {
/* buffer range for efivars */
case 1000 ... 2000:
cmpr = 56;
break;
case 2001 ... 3000:
cmpr = 54;
break;
case 3001 ... 3999:
cmpr = 52;
break;
/* buffer range for nvram, erst */
case 4000 ... 10000:
cmpr = 45;
break;
default:
cmpr = 60;
break;
}
return (size * 100) / cmpr;
We will use the previous worst-case of 60% for compression. For
decompression go extra large (3x) so we make sure there's enough space
for anything.
While at it, rate limit the error message so we don't flood the log
unnecessarily on systems that have accumulated a lot of pstore history.
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Eric Biggers <ebiggers@kernel.org>
Cc: Kees Cook <keescook@chromium.org>
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Link: https://lore.kernel.org/r/20230830212238.135900-1-ardb@kernel.org
Co-developed-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Kees Cook <keescook@chromium.org>
2023-08-30 23:22:38 +02:00
|
|
|
|
|
|
|
|
|
2023-06-27 13:25:41 -07:00
|
|
|
|
2018-10-26 01:17:07 -07:00
|
|
|
|
2017-03-04 22:28:46 -08:00
|
|
|
|
|
|
|
|
|
pstore: Replace crypto API compression with zlib_deflate library calls
Pstore supports compression using a variety of algorithms exposed by the
crypto API. This uses the deprecated comp (as opposed to scomp/acomp)
API, and so we should stop using that, and either move to the new API,
or switch to a different approach entirely.
Given that we only compress ASCII text in pstore, and considering that
this happens when the system is likely to be in an unstable state, the
flexibility that the complex crypto API provides does not outweigh its
impact on the risk that we might encounter additional problems when
trying to commit the kernel log contents to the pstore backend.
So let's switch [back] to the zlib deflate library API, and remove all
the complexity that really has no place in a low-level diagnostic
facility. Note that, while more modern compression algorithms have been
added to the kernel in recent years, the code size of zlib deflate is
substantially smaller than, e.g., zstd, while its performance in terms
of compression ratio is comparable for ASCII text, and speed is deemed
irrelevant in this context.
Note that this means that compressed pstore records may no longer be
accessible after a kernel upgrade, but this has never been part of the
contract. (The choice of compression algorithm is not stored in the
pstore records either)
Tested-by: "Guilherme G. Piccoli" <gpiccoli@igalia.com> # Steam Deck
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Eric Biggers <ebiggers@google.com>
Link: https://lore.kernel.org/r/20230712162332.2670437-3-ardb@kernel.org
Signed-off-by: Kees Cook <keescook@chromium.org>
2023-07-12 18:23:32 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
pstore: Base compression input buffer size on estimated compressed size
Commit 1756ddea6916 ("pstore: Remove worst-case compression size logic")
removed some clunky per-algorithm worst case size estimation routines on
the basis that we can always store pstore records uncompressed, and
these worst case estimations are about how much the size might
inadvertently *increase* due to encapsulation overhead when the input
cannot be compressed at all. So if compression results in a size
increase, we just store the original data instead.
However, it seems that the original code was misinterpreting these
calculations as an estimation of how much uncompressed data might fit
into a compressed buffer of a given size, and it was using the results
to consume the input data in larger chunks than the pstore record size,
relying on the compression to ensure that what ultimately gets stored
fits into the available space.
One result of this, as observed and reported by Linus, is that upgrading
to a newer kernel that includes the given commit may result in pstore
decompression errors reported in the kernel log. This is due to the fact
that the existing records may unexpectedly decompress to a size that is
larger than the pstore record size.
Another potential problem caused by this change is that we may
underutilize the fixed sized records on pstore backends such as ramoops.
And on pstore backends with variable sized records such as EFI, we will
end up creating many more entries than before to store the same amount
of compressed data.
So let's fix both issues, by bringing back the typical case estimation of
how much ASCII text captured from the dmesg log might fit into a pstore
record of a given size after compression. The original implementation
used the computation given below for zlib:
switch (size) {
/* buffer range for efivars */
case 1000 ... 2000:
cmpr = 56;
break;
case 2001 ... 3000:
cmpr = 54;
break;
case 3001 ... 3999:
cmpr = 52;
break;
/* buffer range for nvram, erst */
case 4000 ... 10000:
cmpr = 45;
break;
default:
cmpr = 60;
break;
}
return (size * 100) / cmpr;
We will use the previous worst-case of 60% for compression. For
decompression go extra large (3x) so we make sure there's enough space
for anything.
While at it, rate limit the error message so we don't flood the log
unnecessarily on systems that have accumulated a lot of pstore history.
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Eric Biggers <ebiggers@kernel.org>
Cc: Kees Cook <keescook@chromium.org>
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Link: https://lore.kernel.org/r/20230830212238.135900-1-ardb@kernel.org
Co-developed-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Kees Cook <keescook@chromium.org>
2023-08-30 23:22:38 +02:00
|
|
|
|
pstore: Replace crypto API compression with zlib_deflate library calls
Pstore supports compression using a variety of algorithms exposed by the
crypto API. This uses the deprecated comp (as opposed to scomp/acomp)
API, and so we should stop using that, and either move to the new API,
or switch to a different approach entirely.
Given that we only compress ASCII text in pstore, and considering that
this happens when the system is likely to be in an unstable state, the
flexibility that the complex crypto API provides does not outweigh its
impact on the risk that we might encounter additional problems when
trying to commit the kernel log contents to the pstore backend.
So let's switch [back] to the zlib deflate library API, and remove all
the complexity that really has no place in a low-level diagnostic
facility. Note that, while more modern compression algorithms have been
added to the kernel in recent years, the code size of zlib deflate is
substantially smaller than, e.g., zstd, while its performance in terms
of compression ratio is comparable for ASCII text, and speed is deemed
irrelevant in this context.
Note that this means that compressed pstore records may no longer be
accessible after a kernel upgrade, but this has never been part of the
contract. (The choice of compression algorithm is not stored in the
pstore records either)
Tested-by: "Guilherme G. Piccoli" <gpiccoli@igalia.com> # Steam Deck
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Eric Biggers <ebiggers@google.com>
Link: https://lore.kernel.org/r/20230712162332.2670437-3-ardb@kernel.org
Signed-off-by: Kees Cook <keescook@chromium.org>
2023-07-12 18:23:32 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
pstore: Base compression input buffer size on estimated compressed size
Commit 1756ddea6916 ("pstore: Remove worst-case compression size logic")
removed some clunky per-algorithm worst case size estimation routines on
the basis that we can always store pstore records uncompressed, and
these worst case estimations are about how much the size might
inadvertently *increase* due to encapsulation overhead when the input
cannot be compressed at all. So if compression results in a size
increase, we just store the original data instead.
However, it seems that the original code was misinterpreting these
calculations as an estimation of how much uncompressed data might fit
into a compressed buffer of a given size, and it was using the results
to consume the input data in larger chunks than the pstore record size,
relying on the compression to ensure that what ultimately gets stored
fits into the available space.
One result of this, as observed and reported by Linus, is that upgrading
to a newer kernel that includes the given commit may result in pstore
decompression errors reported in the kernel log. This is due to the fact
that the existing records may unexpectedly decompress to a size that is
larger than the pstore record size.
Another potential problem caused by this change is that we may
underutilize the fixed sized records on pstore backends such as ramoops.
And on pstore backends with variable sized records such as EFI, we will
end up creating many more entries than before to store the same amount
of compressed data.
So let's fix both issues, by bringing back the typical case estimation of
how much ASCII text captured from the dmesg log might fit into a pstore
record of a given size after compression. The original implementation
used the computation given below for zlib:
switch (size) {
/* buffer range for efivars */
case 1000 ... 2000:
cmpr = 56;
break;
case 2001 ... 3000:
cmpr = 54;
break;
case 3001 ... 3999:
cmpr = 52;
break;
/* buffer range for nvram, erst */
case 4000 ... 10000:
cmpr = 45;
break;
default:
cmpr = 60;
break;
}
return (size * 100) / cmpr;
We will use the previous worst-case of 60% for compression. For
decompression go extra large (3x) so we make sure there's enough space
for anything.
While at it, rate limit the error message so we don't flood the log
unnecessarily on systems that have accumulated a lot of pstore history.
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Eric Biggers <ebiggers@kernel.org>
Cc: Kees Cook <keescook@chromium.org>
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Link: https://lore.kernel.org/r/20230830212238.135900-1-ardb@kernel.org
Co-developed-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Kees Cook <keescook@chromium.org>
2023-08-30 23:22:38 +02:00
|
|
|
|
2023-06-27 13:25:41 -07:00
|
|
|
|
2017-03-04 22:28:46 -08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
pstore: Replace crypto API compression with zlib_deflate library calls
Pstore supports compression using a variety of algorithms exposed by the
crypto API. This uses the deprecated comp (as opposed to scomp/acomp)
API, and so we should stop using that, and either move to the new API,
or switch to a different approach entirely.
Given that we only compress ASCII text in pstore, and considering that
this happens when the system is likely to be in an unstable state, the
flexibility that the complex crypto API provides does not outweigh its
impact on the risk that we might encounter additional problems when
trying to commit the kernel log contents to the pstore backend.
So let's switch [back] to the zlib deflate library API, and remove all
the complexity that really has no place in a low-level diagnostic
facility. Note that, while more modern compression algorithms have been
added to the kernel in recent years, the code size of zlib deflate is
substantially smaller than, e.g., zstd, while its performance in terms
of compression ratio is comparable for ASCII text, and speed is deemed
irrelevant in this context.
Note that this means that compressed pstore records may no longer be
accessible after a kernel upgrade, but this has never been part of the
contract. (The choice of compression algorithm is not stored in the
pstore records either)
Tested-by: "Guilherme G. Piccoli" <gpiccoli@igalia.com> # Steam Deck
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Eric Biggers <ebiggers@google.com>
Link: https://lore.kernel.org/r/20230712162332.2670437-3-ardb@kernel.org
Signed-off-by: Kees Cook <keescook@chromium.org>
2023-07-12 18:23:32 +02:00
|
|
|
|
|
|
|
|
|
2017-03-04 22:28:46 -08:00
|
|
|
|
2018-10-26 01:17:07 -07:00
|
|
|
|
2017-03-04 22:28:46 -08:00
|
|
|
|
|
|
|
|
|
2018-10-26 01:17:07 -07:00
|
|
|
|
2023-06-27 13:25:41 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2018-10-26 01:17:07 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-06-27 13:25:41 -07:00
|
|
|
|
2018-10-26 01:17:07 -07:00
|
|
|
|
2017-03-04 22:28:46 -08:00
|
|
|
|
|
|
|
|
|
2017-03-03 17:35:25 -08:00
|
|
|
|
|
|
|
|
|
2010-12-28 14:25:21 -08:00
|
|
|
|
2017-04-27 15:53:21 -07:00
|
|
|
|
2011-08-11 15:14:39 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2010-12-28 14:25:21 -08:00
|
|
|
|
2017-04-27 15:53:21 -07:00
|
|
|
|
|
|
|
|
|
2010-12-28 14:25:21 -08:00
|
|
|
|
2017-03-04 22:57:26 -08:00
|
|
|
|
2017-05-16 12:03:31 -07:00
|
|
|
|
pstore: Replace crypto API compression with zlib_deflate library calls
Pstore supports compression using a variety of algorithms exposed by the
crypto API. This uses the deprecated comp (as opposed to scomp/acomp)
API, and so we should stop using that, and either move to the new API,
or switch to a different approach entirely.
Given that we only compress ASCII text in pstore, and considering that
this happens when the system is likely to be in an unstable state, the
flexibility that the complex crypto API provides does not outweigh its
impact on the risk that we might encounter additional problems when
trying to commit the kernel log contents to the pstore backend.
So let's switch [back] to the zlib deflate library API, and remove all
the complexity that really has no place in a low-level diagnostic
facility. Note that, while more modern compression algorithms have been
added to the kernel in recent years, the code size of zlib deflate is
substantially smaller than, e.g., zstd, while its performance in terms
of compression ratio is comparable for ASCII text, and speed is deemed
irrelevant in this context.
Note that this means that compressed pstore records may no longer be
accessible after a kernel upgrade, but this has never been part of the
contract. (The choice of compression algorithm is not stored in the
pstore records either)
Tested-by: "Guilherme G. Piccoli" <gpiccoli@igalia.com> # Steam Deck
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Eric Biggers <ebiggers@google.com>
Link: https://lore.kernel.org/r/20230712162332.2670437-3-ardb@kernel.org
Signed-off-by: Kees Cook <keescook@chromium.org>
2023-07-12 18:23:32 +02:00
|
|
|
|
2010-12-28 14:25:21 -08:00
|
|
|
|
2017-04-27 15:53:21 -07:00
|
|
|
|
2010-12-28 14:25:21 -08:00
|
|
|
|
|
|
|
|
|
pstore: Replace crypto API compression with zlib_deflate library calls
Pstore supports compression using a variety of algorithms exposed by the
crypto API. This uses the deprecated comp (as opposed to scomp/acomp)
API, and so we should stop using that, and either move to the new API,
or switch to a different approach entirely.
Given that we only compress ASCII text in pstore, and considering that
this happens when the system is likely to be in an unstable state, the
flexibility that the complex crypto API provides does not outweigh its
impact on the risk that we might encounter additional problems when
trying to commit the kernel log contents to the pstore backend.
So let's switch [back] to the zlib deflate library API, and remove all
the complexity that really has no place in a low-level diagnostic
facility. Note that, while more modern compression algorithms have been
added to the kernel in recent years, the code size of zlib deflate is
substantially smaller than, e.g., zstd, while its performance in terms
of compression ratio is comparable for ASCII text, and speed is deemed
irrelevant in this context.
Note that this means that compressed pstore records may no longer be
accessible after a kernel upgrade, but this has never been part of the
contract. (The choice of compression algorithm is not stored in the
pstore records either)
Tested-by: "Guilherme G. Piccoli" <gpiccoli@igalia.com> # Steam Deck
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Eric Biggers <ebiggers@google.com>
Link: https://lore.kernel.org/r/20230712162332.2670437-3-ardb@kernel.org
Signed-off-by: Kees Cook <keescook@chromium.org>
2023-07-12 18:23:32 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2011-11-17 12:58:07 -08:00
|
|
|
|
2011-11-18 13:49:00 -08:00
|
|
|
|
2011-05-16 11:00:27 -07:00
|
|
|
|
|
|
|
|
|
2017-03-04 22:46:41 -08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2017-05-16 12:03:31 -07:00
|
|
|
|
2017-03-04 22:57:26 -08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2017-05-19 15:10:31 -07:00
|
|
|
|
2017-03-04 22:57:26 -08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2017-05-30 15:50:38 -07:00
|
|
|
|
|
|
|
|
|
2017-03-04 22:57:26 -08:00
|
|
|
|
2017-05-30 15:50:38 -07:00
|
|
|
|
2017-03-04 22:57:26 -08:00
|
|
|
|
pstore: Replace crypto API compression with zlib_deflate library calls
Pstore supports compression using a variety of algorithms exposed by the
crypto API. This uses the deprecated comp (as opposed to scomp/acomp)
API, and so we should stop using that, and either move to the new API,
or switch to a different approach entirely.
Given that we only compress ASCII text in pstore, and considering that
this happens when the system is likely to be in an unstable state, the
flexibility that the complex crypto API provides does not outweigh its
impact on the risk that we might encounter additional problems when
trying to commit the kernel log contents to the pstore backend.
So let's switch [back] to the zlib deflate library API, and remove all
the complexity that really has no place in a low-level diagnostic
facility. Note that, while more modern compression algorithms have been
added to the kernel in recent years, the code size of zlib deflate is
substantially smaller than, e.g., zstd, while its performance in terms
of compression ratio is comparable for ASCII text, and speed is deemed
irrelevant in this context.
Note that this means that compressed pstore records may no longer be
accessible after a kernel upgrade, but this has never been part of the
contract. (The choice of compression algorithm is not stored in the
pstore records either)
Tested-by: "Guilherme G. Piccoli" <gpiccoli@igalia.com> # Steam Deck
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Eric Biggers <ebiggers@google.com>
Link: https://lore.kernel.org/r/20230712162332.2670437-3-ardb@kernel.org
Signed-off-by: Kees Cook <keescook@chromium.org>
2023-07-12 18:23:32 +02:00
|
|
|
|
2017-04-27 15:53:21 -07:00
|
|
|
|
2017-03-04 22:46:41 -08:00
|
|
|
|
2017-03-04 23:12:24 -08:00
|
|
|
|
2023-06-27 13:25:41 -07:00
|
|
|
|
2022-06-22 09:31:31 +02:00
|
|
|
|
2017-03-04 23:12:24 -08:00
|
|
|
|
2017-03-04 22:46:41 -08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2010-12-28 14:25:21 -08:00
|
|
|
|
2011-11-18 13:49:00 -08:00
|
|
|
|
|
|
|
|
|
2011-05-16 11:00:27 -07:00
|
|
|
|
2011-11-17 12:58:07 -08:00
|
|
|
|
2010-12-28 14:25:21 -08:00
|
|
|
|
pstore: Replace crypto API compression with zlib_deflate library calls
Pstore supports compression using a variety of algorithms exposed by the
crypto API. This uses the deprecated comp (as opposed to scomp/acomp)
API, and so we should stop using that, and either move to the new API,
or switch to a different approach entirely.
Given that we only compress ASCII text in pstore, and considering that
this happens when the system is likely to be in an unstable state, the
flexibility that the complex crypto API provides does not outweigh its
impact on the risk that we might encounter additional problems when
trying to commit the kernel log contents to the pstore backend.
So let's switch [back] to the zlib deflate library API, and remove all
the complexity that really has no place in a low-level diagnostic
facility. Note that, while more modern compression algorithms have been
added to the kernel in recent years, the code size of zlib deflate is
substantially smaller than, e.g., zstd, while its performance in terms
of compression ratio is comparable for ASCII text, and speed is deemed
irrelevant in this context.
Note that this means that compressed pstore records may no longer be
accessible after a kernel upgrade, but this has never been part of the
contract. (The choice of compression algorithm is not stored in the
pstore records either)
Tested-by: "Guilherme G. Piccoli" <gpiccoli@igalia.com> # Steam Deck
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Eric Biggers <ebiggers@google.com>
Link: https://lore.kernel.org/r/20230712162332.2670437-3-ardb@kernel.org
Signed-off-by: Kees Cook <keescook@chromium.org>
2023-07-12 18:23:32 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2010-12-28 14:25:21 -08:00
|
|
|
|
2017-05-16 12:03:31 -07:00
|
|
|
|
2014-06-06 14:37:31 -07:00
|
|
|
|
2017-05-16 12:03:31 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2010-12-28 14:25:21 -08:00
|
|
|
|
|
|
|
|
|
2011-08-11 15:14:39 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2017-08-28 11:28:21 -07:00
|
|
|
|
2011-08-11 15:14:39 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2020-05-05 21:36:15 -07:00
|
|
|
|
2011-08-11 15:14:39 -07:00
|
|
|
|
|
|
|
|
|
2018-10-18 11:17:42 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2020-02-07 17:46:39 +08:00
|
|
|
|
2018-10-18 11:17:42 -07:00
|
|
|
|
2020-02-07 17:46:39 +08:00
|
|
|
|
2018-10-18 11:17:42 -07:00
|
|
|
|
2018-10-17 03:13:55 -07:00
|
|
|
|
2018-10-18 11:17:42 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-06-13 21:39:42 -07:00
|
|
|
|
2018-10-18 11:17:42 -07:00
|
|
|
|