Hardcoded buffer size in TcpClient/TcpDriver/Udp

[thomas-bc]

F´ Version
Affected Component

---

Problem Description

See

fprime/Drv/TcpClient/TcpClientComponentImpl.cpp

Lines 50 to 51 in da4b1da

	return allocate_out(0, 1024);
	}

We should do like UARTDriver (see here) and configure the size at configure() time so that it can be configured for each instance in the topology. Set a default size of 1024 so that it doesn't break any existing code.

[LeStarch]

Also in UDP and possibly UART. Should be a configuration parameter (on startup).

[menaman123]

So would a fix be having an allocation parameter in the getBuffer method?

[LeStarch]

@menaman123 the fix would be to add an "allocation size" to the configure method:

fprime/Drv/TcpClient/TcpClientComponentImpl.cpp

Lines 32 to 36 in da4b1da

	SocketIpStatus TcpClientComponentImpl::configure(const char* hostname,
	const U16 port,
	const U32 send_timeout_seconds,
	const U32 send_timeout_microseconds) {
	return m_socket.configure(hostname, port, send_timeout_seconds, send_timeout_microseconds);

[bdshenker]

This issue appears to have been fixed in commit 7FF002C by menamen123:

TcpClientComponentImpl.hpp
  PROTECTED: FwSizeType m_allocation_size;
  SocketIpStatus configure(const char* hostname,
                             const U16 port,
                             const U32 send_timeout_seconds = SOCKET_SEND_TIMEOUT_SECONDS,
                             const U32 send_timeout_microseconds = SOCKET_SEND_TIMEOUT_MICROSECONDS,
                             FwSizeType buffer_size = 1024);
TcpClientComponentImpl.cpp
  Fw::Buffer TcpClientComponentImpl::getBuffer() {
      return allocate_out(0, static_cast<U32>(m_allocation_size));
  }

If I am correct then this ticket can now be closed per GIT ticket # 2759.

[LeStarch]

@bdshenker the above ticket did resolve the problem in TcpClient, but not TcpServer nor Udp.

fprime/Drv/TcpServer/TcpServerComponentImpl.cpp

Line 50 in 0e4b4c2

return allocate_out(0, 1024);

[bdshenker]

Got it. Let me fix the other two. From: M Starch ***@***.***> Sent: Monday, October 28, 2024 5:55 PM To: nasa/fprime ***@***.***> Cc: Shenker, Boris D (US 348C) ***@***.***>; Mention ***@***.***> Subject: [EXTERNAL] Re: [nasa/fprime] Hardcoded buffer size in TcpClient/TcpDriver/Udp (Issue #2074) @bdshenker<https://urldefense.us/v3/__https:/github.com/bdshenker__;!!PvBDto6Hs4WbVuu7!ORjEyfqKG0YBosk6YTSLhzmWiqP-LZowUDCJnu7m8-T1TwapMkTix5jckU5SBvEp5dr-__50wu9rbLs8-wmRbftPmg5cDt01XQ$> the above ticket did resolve the problem in TcpClient, but not TcpServer nor Udp. https://github.com/nasa/fprime/blob/0e4b4c2aa219ed23ac7fe03b6b4196146ae7e8e3/Drv/TcpServer/TcpServerComponentImpl.cpp#L50<https://urldefense.us/v3/__https:/github.com/nasa/fprime/blob/0e4b4c2aa219ed23ac7fe03b6b4196146ae7e8e3/Drv/TcpServer/TcpServerComponentImpl.cpp*L50__;Iw!!PvBDto6Hs4WbVuu7!ORjEyfqKG0YBosk6YTSLhzmWiqP-LZowUDCJnu7m8-T1TwapMkTix5jckU5SBvEp5dr-__50wu9rbLs8-wmRbftPmg584uRD7A$> — Reply to this email directly, view it on GitHub<https://urldefense.us/v3/__https:/github.com/nasa/fprime/issues/2074*issuecomment-2442948746__;Iw!!PvBDto6Hs4WbVuu7!ORjEyfqKG0YBosk6YTSLhzmWiqP-LZowUDCJnu7m8-T1TwapMkTix5jckU5SBvEp5dr-__50wu9rbLs8-wmRbftPmg5Jx1y7tQ$>, or unsubscribe<https://urldefense.us/v3/__https:/github.com/notifications/unsubscribe-auth/AW762GPEGQJBUKL7SEZQ2L3Z53MG3AVCNFSM6AAAAAAY7TS2VSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDINBSHE2DQNZUGY__;!!PvBDto6Hs4WbVuu7!ORjEyfqKG0YBosk6YTSLhzmWiqP-LZowUDCJnu7m8-T1TwapMkTix5jckU5SBvEp5dr-__50wu9rbLs8-wmRbftPmg45wisqzA$>. You are receiving this because you were mentioned.Message ID: ***@***.******@***.***>>

[bdshenker]

@LeStarch I completed Git#2074 and reran regression suite locally.
However, when making pull request, one of the integration tests failed.
Please advise how I should proceed.

[LeStarch]

Closing as complete.