TDBStore: Must use work buffer with size of programming unit. #13731
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
ciarmcom
added
the
release-type: patch
|
@SeppoTakalo, thank you for your changes. |
0xc0170
previously approved these changes
Oct 14, 2020
evedon
previously approved these changes
Oct 14, 2020
|
CI started |
Jenkins CI Test : ❌ FAILEDBuild Number: 1 | 🔒 Jenkins CI Job | 🌐 Logs & ArtifactsCLICK for Detailed Summary
|
|
Unittests failed ( |
Please update the unittests or if not, someone else shall pick this up ? |
In TDBStore::copy_record() we must prepare our content in size of full programming unit. If we don't have big enough buffer we end up writing garbage from our stack. In most cases, this is prevent by usage of output buffer, but when writing the record header, due the padding, we must write one full programming unit.
SeppoTakalo
force-pushed
the
IOTCLT-4438
branch
from
67f3b34 to
c8426e8
Compare
mergify
bot
dismissed stale reviews from 0xc0170 and evedon
Pull request has been modified.
|
PR updated. Found a bug in the PR, fixed it. |
0xc0170
approved these changes
Nov 12, 2020
|
CI started |
Jenkins CI Test : ❌ FAILEDBuild Number: 2 | 🔒 Jenkins CI Job | 🌐 Logs & ArtifactsCLICK for Detailed Summary
|
|
I restarted tests. @ARMmbed/mbed-os-core would you be able to take over? |
Jenkins CI Test : ❌ FAILEDBuild Number: 3 | 🔒 Jenkins CI Job | 🌐 Logs & ArtifactsCLICK for Detailed Summary
|
@LDong-Arm is taking over |
LDong-Arm
reviewed
Dec 1, 2020
| @@ -824,17 +823,19 @@ int TDBStore::copy_record(uint8_t from_area, uint32_t from_offset, uint32_t to_o | |||
| uint32_t &to_next_offset) | |||
| { | |||
| int ret; | |||
| record_header_t header; | |||
| record_header_t *header = (record_header_t *) _work_buf; | |||
There was a problem hiding this comment.
There's no guarantee _work_buf is large enough
| @@ -1041,7 +1042,7 @@ int TDBStore::init() | |||
| } | |||
|
|
|||
| _prog_size = _bd->get_program_size(); | |||
| _work_buf = new uint8_t[work_buf_size]; | |||
| _work_buf = new uint8_t[_prog_size]; | |||
There was a problem hiding this comment.
As above, we read record_header_t (a 32-bit aligned struct of size 24 bytes) into _work_buff, but
- here it's only 8-bit aligned
- the program size can be less than that - some flashes can be programmed to 1-byte, in which case the buffer is insufficient
This could be why the Greentea test failed and hard faulted.
|
OK, I'll close this one now. |
mergify
bot
removed
needs: work
release-type: patch
LDong-Arm
reviewed
Dec 1, 2020
| @@ -847,7 +848,7 @@ int TDBStore::copy_record(uint8_t from_area, uint32_t from_offset, uint32_t to_o | |||
| } | |||
|
|
|||
| chunk_size = align_up(sizeof(record_header_t), _prog_size); | |||
| ret = write_area(1 - from_area, to_offset, chunk_size, &header); | |||
There was a problem hiding this comment.
Okay, I get that this is what the PR intended to fix
LDong-Arm
added a commit
to LDong-Arm/mbed-os
that referenced
this pull request
Dec 2, 2020
Co-authored-by: Seppo Takalo <seppo.takalo@arm.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Summary of changes
In TDBStore::copy_record() we must prepare our content in
size of full programming unit. If we don't have big enough buffer
we end up writing garbage from our stack.
In most cases, this is prevent by usage of output buffer, but when
writing the record header, due the padding, we must write one full
programming unit.
Impact of changes
Migration actions required
Documentation
Pull request type
Test results
Reviewers