zeroize memory copy #78
Merged
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
bai-charisu
reviewed
Sep 18, 2020
| * @return 返回错误码 | ||
| */ | ||
| virtual int AioRead(int fd, CurveAioContext* aioctx); | ||
| virtual int AioRead(int fd, CurveAioContext* aioctx, UserDataType dataType); |
There was a problem hiding this comment.
这个只是个枚举类型，没有必要加。
There was a problem hiding this comment.
这个只是个枚举类型，没有必要加。
| * @return 返回错误码 | ||
| */ | ||
| virtual int AioWrite(int fd, CurveAioContext* aioctx); | ||
| virtual int AioWrite(int fd, CurveAioContext* aioctx, | ||
| UserDataType dataType); |
| namespace curve { | ||
| namespace client { | ||
|
|
||
| const char* OpTypeToString(OpType optype) { |
There was a problem hiding this comment.
这个文件就一个函数，而且函数逻辑不复杂，就改成inline放到头文件了。
src/client/io_tracker.cpp
Outdated
| reqlist_.reserve(length_ / Splitor::IOSplitSize() + 1); | ||
|
|
||
| int ret = Splitor::IO2ChunkRequests( | ||
| this, mc_, &reqlist_, nullptr, offset_, length_, mdsclient, fi); |
There was a problem hiding this comment.
修改之前的splitor，需要传入用户buffer，拆分的sub-io的buffer会指向用户buffer不同的offset。
- 写请求会把sub-io的buffer中的数据复制到rpc controller中。
- 读请求读到的数据，会从rpc controller复制到sub-io的buffer中。这里当所有的sub-io完成后，用户buffer里面就存储了需要读取的数据。
修改之后，小io的buffer，改成了IOBuf。
- 写请求，需要把用户IOBuf中的内容cut到sub-io的iobuf中，然后传给rpc controller。
- 读请求，只需要把rpc controller中的iobuf复制给sub-io的iobuf中就可以了。当所有sub-io完成后，会把所有sub-io的iobuf按顺序拼接到一起，返回给用户。
| IOTracker* temp = new IOTracker(this, &mc_, scheduler_); | ||
| temp->SetUserDataType(UserDataType::RawBuffer); |
There was a problem hiding this comment.
这里是直接读快照chunk，接口传入的是char* buf。
src/client/iomanager4file.cpp
Outdated
| return -LIBCURVE_ERROR::FAILED; | ||
| } | ||
|
|
||
| memcpy(buf, data.to_string().c_str(), length); |
There was a problem hiding this comment.
需要零拷贝的只有nebd使用的接口。上层nbd/qemu都是使用的异步读写接口，所以同步这里就没修改了。
wu-hanqing
force-pushed
the
zeroize-memory-copy
branch
3 times, most recently
from
5303c14
to
c4c20f2
Compare
bai-charisu
approved these changes
Sep 21, 2020
xu-chaojie
approved these changes
Sep 22, 2020
xu-chaojie
requested changes
Sep 22, 2020
wu-hanqing
force-pushed
the
zeroize-memory-copy
branch
from
c4c20f2
to
455cbec
Compare
xu-chaojie
approved these changes
Sep 23, 2020
src/client/io_tracker.cpp
Outdated
| type_ = OpType::WRITE; | ||
| type_ = OpType::WRITE; | ||
|
|
||
| DVLOG(9) << "aiowrite op, offset = " << offset << ", length = " << length; |
Change-Id: I2d389b24446623e8f639642a2b354120f20dbe56
wu-hanqing
force-pushed
the
zeroize-memory-copy
branch
from
455cbec
to
d98a59d
Compare
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.
Change-Id: I2d389b24446623e8f639642a2b354120f20dbe56
What problem does this PR solve?
Issue Number: None
Problem Summary:
What is changed and how it works?
What's Changed: zeroize memory copy between nebd and curve-client
How it Works:
Side effects(Breaking backward compatibility? Performance regression?):
Check List