Ignored.
Up until now I haven't met with any specific issue with the current
implementation.
Nevertheless, I believe that EDK2 implementation shall adhere as much as it can
to the UEFI spec as vendors and developers rely on it.
That's to say, I think that because SnpDxe currently supports S/W UNDI it
shouldn't ignore any possible flow that is defined in the spec, including
FRAGMENTED/BLOCK/DON'T_BLOCK etc. You might consider it a bug/call for
implementation.
Ari
From: Li, Ruth [mailto:ruth...@intel.com]
Sent: Wednesday, February 27, 2013 12:52 PM
To: edk2-devel@lists.sourceforge.net
Subject: Re: [edk2] UNDI's Fragmented Transmit & Transmit Block
Hi Ari
I may misunderstand, you can ignore my statement of 'edk2 not support S/W
UNDI'. Sorry.
I am wondering, do you meet any specific issue in using current code?
Thanks,
Ruth
From: Ari Zigler [mailto:a...@mellanox.com]
Sent: Wednesday, February 27, 2013 4:57 PM
To: edk2-devel@lists.sourceforge.net<mailto:edk2-devel@lists.sourceforge.net>
Subject: Re: [edk2] UNDI's Fragmented Transmit & Transmit Block
Hi Ruth, thanks for the prompt reply.
First, I must say that I'm a little surprised by your claim that EDK2 does not
support S/W UNDI as there is that E100b UNDI driver which happens to work just
fine under EDK2 implementation and it doesn't seem to be an H/W one.
(Particularly, it explicitly exposes PXE_SW_UNDI as its !PXE struct. See
edk2/OptionRomPkg/UndiRuntimeDxe/Init.c:20 and the corresponding implementation
in SnpDxe/Snp.c:422). Moreover, our device also managed to function under EDK2
as a S/W UNDI.
Second, assuming that EDK2 will eventually work with S/W UNDI - which module in
the network stack is responsible for the fragmentation?
Thanks,
Ari
From: Li, Ruth [mailto:ruth...@intel.com]
Sent: Wednesday, February 27, 2013 4:34 AM
To: edk2-devel@lists.sourceforge.net<mailto:edk2-devel@lists.sourceforge.net>
Subject: Re: [edk2] UNDI's Fragmented Transmit & Transmit Block
Hi,
>From UEFI 2.3.1 Appendix E, the
>PXE_OPFLAGS_TRANSMIT_FRAGEMENTED/BLOCK/DON'T_BLOCK are only for S/W UNDI. I
>think current EDK2 implementation only handle the H/W UNDI now.
Thanks,
Ruth
From: Ari Zigler [mailto:a...@mellanox.com]
Sent: Wednesday, February 27, 2013 12:45 AM
To: edk2-devel@lists.sourceforge.net<mailto:edk2-devel@lists.sourceforge.net>
Subject: [edk2] UNDI's Fragmented Transmit & Transmit Block
Hi,
As no driver in the EDK2 network stack uses Fragmented Transmit
(PXE_OPFLAGS_TRANSMIT_FRAGMENTED), not even the SnpDxe one, and nevertheless,
fragmented Ip packets manage to go on the wire (I checked it), what is the
purpose of this opflag? Who might partition the packets into fragments? Does
it need special care?
Another issue is the PXE_OPFLAGS_TRANSMIT_BLOCK/DON'T_BLOCK:
The SnpDxe driver ignores these flags and do not pass them to the UNDI driver.
The edk2 platform performs UNDI GetStatus periodically to handle the completed
TX buffers, so I presume the system works in a non-blocking mode, doesn't it?
Or it might be a bug in the SnpDxe implementation...
Thnaks,
Ari
------------------------------------------------------------------------------
Everyone hates slow websites. So do we.
Make your web apps faster with AppDynamics
Download AppDynamics Lite for free today:
http://p.sf.net/sfu/appdyn_d2d_feb
_______________________________________________
edk2-devel mailing list
edk2-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/edk2-devel
