On 10.12.24 16:27, Bjorn Helgaas wrote:
On Sun, Dec 08, 2024 at 07:49:50PM +0100, Gerhard Engleder wrote:
Link down and up triggers update of MTA table. This update executes many
PCIe writes and a final flush. Thus, PCIe will be blocked until all writes
are flushed. As a result, DMA transfers of other targets suffer from delay
in the range of 50us. This results in timing violations on real-time
systems during link down and up of e1000e.
These look like PCIe memory writes (not config or I/O writes), which
are posted and do not require Completions. Generally devices should
not delay acceptance of posted requests for more than 10us (PCIe r6.0,
sec 2.3.1).
Since you mention DMA to/from other targets, maybe there's some kind
of fairness issue in the interconnect, which would suggest a
platform-specific issue that could happen with devices other than
e1000e.
I think it would be useful to get to the root cause of this, or at
least mention the interconnect design where you saw the problem in
case somebody trips over this issue with other devices.
Getting the root cause would be interesting, but this problem happens on
a rather ancient platform: an Intel i3-2310E Ivy Bridge CPU launched in
2011 (which still does its job as robot controller). Intel support does
not answer questions for such old platforms. Even for other timing
issues on the interconnect the Intel support was limited. I will mention
the CPU more explicitly as the platform with this issue.
The PCIe spec does have an implementation note that says drivers might
need to restrict the programming model as you do here for designs that
can't process posted requests fast enough. If that's the case for
e1000e, I would ask Intel whether other related devices might also be
affected.
Even for newer CPUs the Intel support has already ended and this CPU is
not sold by Intel anymore. So I won't get an answer. But our experience
is that limiting the number of posted writes always make sense at least
for real-time. Even for our own FPGA based PCIe target, which is able to
consume posted writes at full speed, we limit the number of posted
writes to reduce negative effects on real-time. This experience was made
with multiple Intel platforms.
Gerhard
