On Mon, Nov 16, 2026 at 6:14 AM Andrew Lunn [off-list ref] wrote:
quoted
quoted
Think about it.
If you don't have a MAC, you don't have FCS.
There are existing mainline PHY drivers exposing this functionality.
The QCA807x/QCA808x drivers use qcom_phy_counter_config() to enable
CRC checking, and qcom_phy_get_stats() reports the RX CRC-error count
through ethtool_phy_stats.rx_errors [0]. The IP101G driver also exposes
a phy_crc_errors counter separately from phy_symbol_errors [1].
This demonstrates frame-aware diagnostic functionality within these
PHYs, rather than something every PHY necessarily implements.
The generic byte counter need not require that functionality. My
question was whether rx_bytes should count received bytes without
conditioning the count on FCS validity. ACx00 could then report its
hardware byte total through that field, including bytes from frames
which its separate CRC checker identifies as bad.
Read 802.3. What does an 802.3 defined PHY actually allow. That is
ideally what the kernel should support with these counters. Lowest
common denominator.
FCS validation is part of the MAC functionality defined by 802.3.
Counting received bytes at the PHY need not depend on that validation.
The PHY CRC counters I mentioned are additional hardware diagnostics,
not functionality that every PHY must implement.
The original statistics interface also distinguishes the IEEE-defined
eth-phy statistics from the additional phydev counters [0]. For
rx_bytes, the common meaning can be received-byte activity without
requiring the PHY to determine whether the MAC will accept the frame.
So i would say RX bytes does not differentiate between the frames the
MAC will later throw away because the FCS is wrong. I would say it is
also O.K. to include the preamble, and the FCS, any padding up to 64
byte frames, etc.
That matches the ACx00 byte counter: it includes the Ethernet header,
padding and FCS, including bytes from frames with an invalid FCS. It
excludes the preamble and SFD.
The separate counter was prompted by the existing kernel-doc wording,
"Total successfully received bytes", rather than an IEEE requirement
or a need for successful-only byte accounting.
Since this counter is not actually defined by 802.3, i would also
allow some flexibility in its meaning. Different PHYs might have
slightly different meanings.
For this use case, documenting those differences is sufficient. I
don't need identical byte boundaries across vendors, and whether a
counter includes the preamble does not change its diagnostic value.
Look at it from the other direction. What is your use case? Does the
exact definition actually matter? Does it matter if a different
vendors PHY counts it differently? Does your use case require cross
device compatibility?
The use case is the board bring-up and networking diagnosis you
describe: exposing the counters already available in the PHY to help
locate where traffic is being lost. It is the same PHY/MAC comparison
described in the diagnostics work accompanying this interface [1].
For example, in our tests the PHY byte and CRC-error counters advanced
for traffic that the MAC discarded before DMA. That distinguishes
traffic reaching the PHY but being rejected downstream from traffic
not reaching the PHY at all.
A common interface is useful for inspecting that activity across
devices, but exact cross-device byte totals are not required. The
important point here is that received traffic remains visible even
when the MAC subsequently rejects it.
[0] https://lore.kernel.org/r/20240829174342.3255168-3-kuba@kernel.org/ (local)
[1] https://lore.kernel.org/r/20241221081530.3003900-5-o.rempel@pengutronix.de/ (local)
You have not actually said what your use case is. For new board
bringup, debugging broken networking, the exact definition generally
does not matter. The PHY RX counter is going up, the MAC is not,
something between the MAC and the PHY. Neither the MAC nor the PHY
counters are going up, something between the RJ45 and the PHY. The
absolute values don't matter, its that they change which matters. But
maybe your use case is different?
Andrew