starred/terminal
1
0
Fork 0
mirror of https://github.com/microsoft/terminal.git synced 2026-02-04 05:35:20 +00:00
Code Issues 19.0k Packages Projects Releases 20 Wiki Activity

Document ConDrv headers (condrv.h and conmsgl*.h) #14273

New Issue
Open
opened 2026-01-31 04:05:53 +00:00 by claunia · 1 comment
claunia commented 2026-01-31 04:05:53 +00:00
Owner
Copy Link

Originally created by @lhecker on GitHub (Jun 19, 2021).

As the ConDrv headers and \Device\ConDrv\Server are undocumented, it:

  • Prevents the community from building their own, alternative conhost
  • is a hindrance for MS employees to quickly understand how ConDrv works

In the name of https://github.com/microsoft/terminal/issues/10462#issuecomment-864446348 I'd like to thus ask for ConDrv to be documented.

Originally created by @lhecker on GitHub (Jun 19, 2021). As the ConDrv headers and `\Device\ConDrv\Server` are undocumented, it: * Prevents the community from building their own, alternative conhost * is a hindrance for MS employees to quickly understand how ConDrv works In the name of https://github.com/microsoft/terminal/issues/10462#issuecomment-864446348 I'd like to thus ask for ConDrv to be documented.
claunia added the Product-ConhostIssue-TaskPriority-3Area-ServerIssue-Docs labels 2026-01-31 04:05:54 +00:00
claunia commented 2026-01-31 04:05:55 +00:00
Author
Owner
Copy Link

@malxau commented on GitHub (Jul 4, 2021):

Can I ask what the expected gain is here?

~Six months ago I went through the exercise of figuring out how this works. The conhost side is straightforward because all of the code pulling from condrv is in the tree, and I'm sure with this any competent dev could write/rewrite/replace the conhost side. The client side was a bit more interesting, because that code would normally be in kernel32.dll but with a usermode debugger and the headers here it wasn't hard to figure out.

What I was trying to do with this exercise was investigate the feasibility of a Win32++ that exposes newer capabilities via a familiar API, and the natural way to do that is with condrv messages. As luck would have it, condrv doesn't do much sanity checking of messages so adding new ones is trivial. I didn't speak to anyone here about it because I thought this isn't something you'd be interested in (with all those deprecation notices on MSDN and all), and eventually lost interest in pursuing it. There is a nice bonus of having a functional/RPC interface, which is the client can detect when the server can't implement something and gracefully fall back.

Ultimately though it ended in a pile of "learnings":

  • The condrv API is intrinsically synchronous and the context switches here are killer. Trying to display RGB color squares with my extended Win32 API was painful because it takes two process context switches per cell (set color, write text), and of course has to switch back again. Given a 30usec context switch cost, a 64x64 2D grid of colored cells with 2x2 switches per cell means ~500ms of switches - rendering is extra. Note that the basic console model predates condrv/conhost, so the cost of this interface became more expensive later in its life.
  • Following on from the above, there's an "opportunity" to generate and buffer VT sequences on the client so that simple requests can be batched.
  • There's very limited opportunities for "mix and match" because conhost is hardcoded to a condrv version, and individual clients will talk whatever version kernel32 implements. Even though it wouldn't be hard to build a functional condrv for Windows XP and run the conhost from this tree on it, no command line programs on XP would talk to it without going full Detours on kernel32.
  • Following on from the above, if it becomes a goal to run WT or the conhost from this tree across a broad range of versions of Windows, more condrv version negotiation would be helpful. Because historically all of the components were shipped in a single release, this looks to be quite absent.
  • Although open sourcing the existing condrv, or even writing a new one implementing its interface, wouldn't be overly hard, it's not (that) helpful for open source developers given kernel driver signing requirements.

So even as someone who generally finishes anything he starts, this is one thing I walked away from.

@malxau commented on GitHub (Jul 4, 2021): Can I ask what the expected gain is here? ~Six months ago I went through the exercise of figuring out how this works. The conhost side is straightforward because all of the code pulling from condrv is in the tree, and I'm sure with this any competent dev could write/rewrite/replace the conhost side. The client side was a bit more interesting, because that code would normally be in kernel32.dll but with a usermode debugger and the headers here it wasn't hard to figure out. What I was trying to do with this exercise was investigate the feasibility of a Win32++ that exposes newer capabilities via a familiar API, and the natural way to do that is with condrv messages. As luck would have it, condrv doesn't do much sanity checking of messages so adding new ones is trivial. I didn't speak to anyone here about it because I thought this isn't something you'd be interested in (with all those deprecation notices on MSDN and all), and eventually lost interest in pursuing it. There is a nice bonus of having a functional/RPC interface, which is the client can detect when the server can't implement something and gracefully fall back. Ultimately though it ended in a pile of "learnings": - The condrv API is intrinsically synchronous and the context switches here are killer. Trying to display RGB color squares with my extended Win32 API was painful because it takes two process context switches per cell (set color, write text), and of course has to switch back again. Given a 30usec context switch cost, a 64x64 2D grid of colored cells with 2x2 switches per cell means ~500ms of switches - rendering is extra. Note that the basic console model predates condrv/conhost, so the cost of this interface became more expensive later in its life. - Following on from the above, there's an "opportunity" to generate and buffer VT sequences on the client so that simple requests can be batched. - There's very limited opportunities for "mix and match" because conhost is hardcoded to a condrv version, and individual clients will talk whatever version kernel32 implements. Even though it wouldn't be hard to build a functional condrv for Windows XP and run the conhost from this tree on it, no command line programs on XP would talk to it without going full Detours on kernel32. - Following on from the above, if it becomes a goal to run WT or the conhost from this tree across a broad range of versions of Windows, more condrv version negotiation would be helpful. Because historically all of the components were shipped in a single release, this looks to be quite absent. - Although open sourcing the existing condrv, or even writing a new one implementing its interface, wouldn't be overly hard, it's not (that) helpful for open source developers given kernel driver signing requirements. So even as someone who generally finishes anything he starts, this is one thing I walked away from.
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
⛺ Reserved
A11yCO
A11yMAS
A11ySev1
A11ySev2
A11ySev3
A11yTTValidated
A11yUsable
A11yVoiceAccess
A11yWCAG
Area-Accessibility
Area-AtlasEngine
Area-AzureShell
Area-Build
Area-Build
Area-Chat
Area-CmdPal
Area-CodeHealth
Area-Commandline
Area-CookedRead
Area-DefApp
Area-Extensibility
Area-Fonts
Area-GroupPolicy
Area-i18n
Area-Input
Area-Interaction
Area-Interop
Area-Localization
Area-Output
Area-Performance
Area-Portable
Area-Quality
Area-Remoting
Area-Rendering
Area-Schema
Area-Server
Area-Settings
Area-SettingsUI
Area-ShellExtension
Area-ShellExtension
Area-ShellExtension
Area-Suggestions
Area-Suggestions
Area-TerminalConnection
Area-TerminalControl
Area-Theming
Area-UserInterface
Area-VT
Area-Windowing
Area-WPFControl
AutoMerge
Blocking-Ingestion
Culprit-Centennial
Culprit-WinUI
Disability-All
Disability-Blind
Disability-LowVision
Disability-Mobility
External-Blocked-WinUI3
Fixed
Gathering-Data
good first issue
HCL-E+D
HCL-WindowsTerminal
Help Wanted
Impact-Compatibility
Impact-Compliance
Impact-Correctness
Impact-Visual
In-PR
InclusionBacklog
InclusionBacklog-Windows TerminalWin32
InclusionCommitted-202206
Issue-Bug
Issue-Docs
Issue-Feature
Issue-Feature
Issue-Question
Issue-Samples
Issue-Scenario
Issue-Task
Needs-Attention
Needs-Author-Feedback
Needs-Bisect
Needs-Discussion
Needs-Repro
Needs-Tag-Fix
Needs-Tag-Fix
Needs-Triage
No-Recent-Activity
Priority-0
Priority-1
Priority-2
Priority-3
Product-Cmd.exe
Product-Colortool
Product-Colortool
Product-Colortool
Product-Conhost
Product-Conpty
Product-Meta
Product-Powershell
Product-Terminal
Product-WSL
pull-request
Mirrored from GitHub Pull Request
 Resolution-Answered
Resolution-By-Design
Resolution-Duplicate
Resolution-External
Resolution-Fix-Available
Resolution-Fix-Committed
Resolution-No-Repro
Resolution-Won't-Fix
Severity-Blocking
Severity-Crash
Severity-DataLoss
spam
this-will-be-a-breaking-change
Tracking-External
WindowsTerminal_Win32
Work-Item
zAskModeBug
zInbox-Bug
No Label Area-Server Issue-Docs Issue-Task Priority-3 Product-Conhost
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
claunia
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: starred/terminal#14273
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?