4.1.4 RAM tests
Principles
There are three RAM area concepts:
• Block: a contiguous area of 16 bytes (RAM_BLOCK_SIZE), hard coded by the STL (no link with the memory
physical sectors).
• Section: a contiguous area of 128 bytes (RAM_SECTION_SIZE), hard coded by the STL.
• Subset: a contiguous area, with the size being a multiple of two blocks and with a 32-bit aligned start
address. Then a subset size is not necessarily a multiple of the section size, because the last part of a
subset can be less than one section.
• Subset size = N * RAM_SECTION_SIZE+ 2 * M RAM_BLOCK_SIZE, where:
– N is an integer ≥ 0
– M is an integer < (RAM_SECTION_SIZE / (2 * RAM_BLOCK_SIZE))
The subset is defined by the User application. It is possible to define several subsets.
The STL implements a RAM memory test with the following principles:
• RAM tests are performed on RAM blocks defined by the User application
• RAM tests are performed either in a row (one shot), or partially in a single atomic step for a number of
sections defined by the User application
• The test implementation is based on March C- algorithm
• RAM tests are performed on RAM content (not on DCache content).
The mandatory steps (for the User application) to perform RAM tests are:
• initialization of RAM test
• configuration of the RAM subset(s)
• execution of the RAM test
Once all subsets are tested, the application must reset the RAM test module to perform the test again.
In the case of an STL_ERROR / STL_FAILED test result, the test module is stuck at the failed memory subset. It
is therefore adviseable to deinitialize then initialize the RAM test module (or to reset it) before performing the test
again.
RAM test and interrupts
By default, the STM32 interrupts and Cortex exceptions with configurable priority are temporarily masked during
execution of the smallest data granularity (block), except when the User application activates the dedicated
compilation switch STL_ENABLE_IT (see Section 5.5.2 Steps to build an application from scratch). If the flag is
activated, the end user is responsible for managing interferences between the STL and its application software
that could possibly lead to false STL-error reports or application software malfunctions, due to RAM content
destruction by the test (see March C- test principle ).
March C- test principle
The RAM test is based on a March C- method where memory is written with a specific pattern and then read
back. To restore the initial content inside the memory, an optional backup buffer may be used. This optional
backup buffer is enabled by default and can be disabled by activating the compilation switch,
STL_DISABLE_RAM_BCKUP_BUF (see Section 5.5.2 Steps to build an application from scratch). The address
of the RAM backup buffer is configurable by the end user at compilation time, it must be outside RAM subset
configuration. There is only one RAM backup buffer for the whole test. The RAM backup buffer, when enabled, is
also tested by March C- test.
ST Restricted - Subject to Non-Disclosure Agreement – Do not copy
UM2601
STL functional description
UM2601 - Rev 3
page 12/71
剩余70页未读，继续阅读
嵌入式基地
- 粉丝: 4w+
- 资源: 347
我的内容管理
收起
- 我的资源
快来上传第一个资源
我的收益 登录查看自己的收益
我的积分
登录查看自己的积分
我的C币
登录后查看C币余额
我的收藏 
我的下载 
下载帮助
会员权益专享
最新资源
- BSC绩效考核指标汇总 (2).docx
- BSC资料.pdf
- BSC绩效考核指标汇总 (3).pdf
- C5000W常见问题解决方案.docx
- BSC概念 (2).pdf
- ESP8266智能家居.docx
- ESP8266智能家居.pdf
- BSC概念 HR猫猫.docx
- C5000W常见问题解决方案.pdf
- BSC模板：关键绩效指标示例(财务、客户、内部运营、学习成长四个方面).docx
- BSC概念.docx
- BSC模板：关键绩效指标示例(财务、客户、内部运营、学习成长四个方面).pdf
- BSC概念.pdf
- 各种智能算法的总结汇总.docx
- BSC概念 HR猫猫.pdf
- bsc概念hr猫猫.pdf
资源上传下载、课程学习等过程中有任何疑问或建议，欢迎提出宝贵意见哦~我们会及时处理！
点击此处反馈
