Show an error if CIA contend is encrypted #5130
Conversation
|
The braces should be placed on the same line as the if statements. |
B3n30
force-pushed
the
warn_if_cia_contend_is_encrypted
branch
from
2b70160
to
706a329
Compare
B3n30
force-pushed
the
warn_if_cia_contend_is_encrypted
branch
from
706a329
to
391580c
Compare
src/core/file_sys/ncch_container.cpp
Outdated
| @@ -133,8 +133,37 @@ Loader::ResultStatus NCCHContainer::OpenFile(const std::string& filepath, u32 nc | |||
| return Loader::ResultStatus::Success; | |||
| } | |||
|
|
|||
| Loader::ResultStatus NCCHContainer::LoadHeader() { | |||
| if (has_header) | |||
There was a problem hiding this comment.
I think we are following the convention that all if statements should be braced. Dito below.
B3n30
force-pushed
the
warn_if_cia_contend_is_encrypted
branch
from
1a3cbf7
to
a4457d8
Compare
|
Since some users requested it, I added a log that shows which key is missing when a Ticket gets decrypted during cia installation |
src/core/hle/service/am/am.cpp
Outdated
| const auto file_ext = FileUtil::GetExtensionFromFilename(path); | ||
| return std::find(extensions.begin(), extensions.end(), file_ext) != extensions.end(); | ||
| } | ||
| } // namespace |
There was a problem hiding this comment.
I personally don't see a reason for this hardcoded list of extensions, why can't we just try to load everything in the folder?
There was a problem hiding this comment.
This was just to don't do any unnecessary stuff (on big game lists this could really be costly)
There was a problem hiding this comment.
This is only executed once when a CIA is installed so I don't think it's going to influence the game list performance
| @@ -697,7 +728,7 @@ Loader::ResultStatus NCCHContainer::ReadOverrideRomFS(std::shared_ptr<RomFSReade | |||
| } | |||
|
|
|||
| Loader::ResultStatus NCCHContainer::ReadProgramId(u64_le& program_id) { | |||
| Loader::ResultStatus result = Load(); | |||
| Loader::ResultStatus result = LoadHeader(); | |||
There was a problem hiding this comment.
Why is this change necessary?
There was a problem hiding this comment.
This ensure that we get a ProgramID even if the NCCH is encrypted and the keys aren't available. It probably still requires to change the ui to show the programId instead of the filename in the game list though
|
|
||
| has_header = true; | ||
| return Loader::ResultStatus::Success; | ||
| } |
There was a problem hiding this comment.
Should probably convert the first part of the Load() to LoadHeader() then
There was a problem hiding this comment.
Thats not that easy because it seems the remaining part of Load() needs the file to be at the position after the header
src/core/hle/service/am/am.cpp
Outdated
| return FileUtil::ForeachDirectoryEntry(nullptr, physical_name, callback); | ||
| } | ||
| }; | ||
| if (!FileUtil::ForeachDirectoryEntry(nullptr, path, callback)) { |
There was a problem hiding this comment.
Is this correct? won't path point to the CIA file (instead of the installed directory)?
There was a problem hiding this comment.
Oh yeah, thx for spotting this
src/core/hle/service/am/am.cpp
Outdated
| const auto file_ext = FileUtil::GetExtensionFromFilename(path); | ||
| return std::find(extensions.begin(), extensions.end(), file_ext) != extensions.end(); | ||
| } | ||
| } // namespace |
There was a problem hiding this comment.
This is only executed once when a CIA is installed so I don't think it's going to influence the game list performance
We already show an error if the cia is encrypted but not if the cia itself isn't encrypted but it's content is.
This lead to a lot of confusion where users thought they successfully installed the cia, but then nothing appeared in the game list. Now the encryption error is also shown if the cia contained any executable that Citra isn't able to decrypt (e.g. due to missing aes keys).
In addition any executable that is encrypted and Citra can't decrypt is still shown in the game list (if executables without icons are visible). Such executables will still be shown in an ugly way (no proper title, icon, region etc), since we can't access the SMDH.
As a last change in that PR I removed a debugging line we added that showed the path of each NCCH file, since it was only added to debug an issue where loading encrypted NCCHs caused a crash.
This change is