| Message ID | 20231211223346.2497157-7-tobias@waldekranz.com (mailing list archive) |
|---|---|
| State | Superseded |
| Delegated to: | Netdev Maintainers |
| Headers | show |
| Series | net: dsa: mv88e6xxx: Add "eth-mac" and "rmon" counter group support | expand |
On 12/11/23 14:33, Tobias Waldekranz wrote: > Chips in this family only has one set of histogram counters, which can > be used to count ingressing and/or egressing traffic. mv88e6xxx has, > up until this point, kept the hardware default of counting both > directions. s/has/have/ > > In the mean time, standard counter group support has been added to > ethtool. Via that interface, drivers may report ingress-only and > egress-only histograms separately - but not combined. > > In order for mv88e6xxx to maximalize amount of diagnostic information > that can be exported via standard interfaces, we opt to limit the > histogram counters to ingress traffic only. Which will allow us to > export them via the standard "rmon" group in an upcoming commit. s/maximalize/maximize/ > > The reason for choosing ingress-only over egress-only, is to be > compatible with RFC2819 (RMON MIB). > > Signed-off-by: Tobias Waldekranz <tobias@waldekranz.com> Out of curiosity: does this commit and the next one need to be swapped in order to surprises if someone happened to be bisecting across this patch series? Unless there is something else that needs to be addressed, please address the two typos above, regardless: Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com>
On mon, dec 11, 2023 at 15:03, Florian Fainelli <f.fainelli@gmail.com> wrote: > On 12/11/23 14:33, Tobias Waldekranz wrote: >> Chips in this family only has one set of histogram counters, which can >> be used to count ingressing and/or egressing traffic. mv88e6xxx has, >> up until this point, kept the hardware default of counting both >> directions. > > s/has/have/ > >> >> In the mean time, standard counter group support has been added to >> ethtool. Via that interface, drivers may report ingress-only and >> egress-only histograms separately - but not combined. >> >> In order for mv88e6xxx to maximalize amount of diagnostic information >> that can be exported via standard interfaces, we opt to limit the >> histogram counters to ingress traffic only. Which will allow us to >> export them via the standard "rmon" group in an upcoming commit. > > s/maximalize/maximize/ > >> >> The reason for choosing ingress-only over egress-only, is to be >> compatible with RFC2819 (RMON MIB). >> >> Signed-off-by: Tobias Waldekranz <tobias@waldekranz.com> > > Out of curiosity: does this commit and the next one need to be swapped > in order to surprises if someone happened to be bisecting across this > patch series? I'm not sure I follow. This commit only changes the behavior of the existing counters (ethtool -S). If it was swapped with the next one, then there would be one commit in the history in which the "rmon" histogram counters would report the wrong values (the bug pointed out by Vladimir on v2) > Unless there is something else that needs to be addressed, please > address the two typos above, regardless: s/Unless/If/? > Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com> Thanks for the review!
On 12/11/23 15:35, Tobias Waldekranz wrote: > On mon, dec 11, 2023 at 15:03, Florian Fainelli <f.fainelli@gmail.com> wrote: >> On 12/11/23 14:33, Tobias Waldekranz wrote: >>> Chips in this family only has one set of histogram counters, which can >>> be used to count ingressing and/or egressing traffic. mv88e6xxx has, >>> up until this point, kept the hardware default of counting both >>> directions. >> >> s/has/have/ >> >>> >>> In the mean time, standard counter group support has been added to >>> ethtool. Via that interface, drivers may report ingress-only and >>> egress-only histograms separately - but not combined. >>> >>> In order for mv88e6xxx to maximalize amount of diagnostic information >>> that can be exported via standard interfaces, we opt to limit the >>> histogram counters to ingress traffic only. Which will allow us to >>> export them via the standard "rmon" group in an upcoming commit. >> >> s/maximalize/maximize/ >> >>> >>> The reason for choosing ingress-only over egress-only, is to be >>> compatible with RFC2819 (RMON MIB). >>> >>> Signed-off-by: Tobias Waldekranz <tobias@waldekranz.com> >> >> Out of curiosity: does this commit and the next one need to be swapped >> in order to surprises if someone happened to be bisecting across this >> patch series? > > I'm not sure I follow. This commit only changes the behavior of the > existing counters (ethtool -S). If it was swapped with the next one, > then there would be one commit in the history in which the "rmon" > histogram counters would report the wrong values (the bug pointed out by > Vladimir on v2) Right, somehow I thought there was provision for reporting the TX histograms through the rmon interface which your subsequent patch would do, but it does not appear so, never mind then. > >> Unless there is something else that needs to be addressed, please >> address the two typos above, regardless: > > s/Unless/If/? Yes, the latter!
On Mon, Dec 11, 2023 at 11:33:44PM +0100, Tobias Waldekranz wrote: > Chips in this family only has one set of histogram counters, which can > be used to count ingressing and/or egressing traffic. mv88e6xxx has, > up until this point, kept the hardware default of counting both > directions. > > In the mean time, standard counter group support has been added to > ethtool. Via that interface, drivers may report ingress-only and > egress-only histograms separately - but not combined. > > In order for mv88e6xxx to maximalize amount of diagnostic information > that can be exported via standard interfaces, we opt to limit the > histogram counters to ingress traffic only. Which will allow us to > export them via the standard "rmon" group in an upcoming commit. > > The reason for choosing ingress-only over egress-only, is to be > compatible with RFC2819 (RMON MIB). > > Signed-off-by: Tobias Waldekranz <tobias@waldekranz.com> > --- I guess this needs an ack from Andrew.
On Mon, Dec 11, 2023 at 11:33:44PM +0100, Tobias Waldekranz wrote: > Chips in this family only has one set of histogram counters, which can > be used to count ingressing and/or egressing traffic. mv88e6xxx has, > up until this point, kept the hardware default of counting both > directions. > > In the mean time, standard counter group support has been added to > ethtool. Via that interface, drivers may report ingress-only and > egress-only histograms separately - but not combined. > > In order for mv88e6xxx to maximalize amount of diagnostic information > that can be exported via standard interfaces, we opt to limit the > histogram counters to ingress traffic only. Which will allow us to > export them via the standard "rmon" group in an upcoming commit. > > The reason for choosing ingress-only over egress-only, is to be > compatible with RFC2819 (RMON MIB). > > Signed-off-by: Tobias Waldekranz <tobias@waldekranz.com> Keeping the hardware defaults is pretty arbitrary. So i don't see why we cannot change it to fulfil standard RMON. Reviewed-by: Andrew Lunn <andrew@lunn.ch> Andrew
diff --git a/drivers/net/dsa/mv88e6xxx/chip.c b/drivers/net/dsa/mv88e6xxx/chip.c index 6f1f71cb0de5..a3bd12ceba7b 100644 --- a/drivers/net/dsa/mv88e6xxx/chip.c +++ b/drivers/net/dsa/mv88e6xxx/chip.c @@ -1221,7 +1221,7 @@ static size_t mv88e6095_stats_get_stat(struct mv88e6xxx_chip *chip, int port, return 0; *data = _mv88e6xxx_get_ethtool_stat(chip, stat, port, 0, - MV88E6XXX_G1_STATS_OP_HIST_RX_TX); + MV88E6XXX_G1_STATS_OP_HIST_RX); return 1; } @@ -1233,7 +1233,7 @@ static size_t mv88e6250_stats_get_stat(struct mv88e6xxx_chip *chip, int port, return 0; *data = _mv88e6xxx_get_ethtool_stat(chip, stat, port, 0, - MV88E6XXX_G1_STATS_OP_HIST_RX_TX); + MV88E6XXX_G1_STATS_OP_HIST_RX); return 1; } @@ -1246,7 +1246,7 @@ static size_t mv88e6320_stats_get_stat(struct mv88e6xxx_chip *chip, int port, *data = _mv88e6xxx_get_ethtool_stat(chip, stat, port, MV88E6XXX_G1_STATS_OP_BANK_1_BIT_9, - MV88E6XXX_G1_STATS_OP_HIST_RX_TX); + MV88E6XXX_G1_STATS_OP_HIST_RX); return 1; } diff --git a/drivers/net/dsa/mv88e6xxx/global1.c b/drivers/net/dsa/mv88e6xxx/global1.c index 174c773b38c2..49444a72ff09 100644 --- a/drivers/net/dsa/mv88e6xxx/global1.c +++ b/drivers/net/dsa/mv88e6xxx/global1.c @@ -462,8 +462,7 @@ int mv88e6390_g1_rmu_disable(struct mv88e6xxx_chip *chip) int mv88e6390_g1_stats_set_histogram(struct mv88e6xxx_chip *chip) { return mv88e6xxx_g1_ctl2_mask(chip, MV88E6390_G1_CTL2_HIST_MODE_MASK, - MV88E6390_G1_CTL2_HIST_MODE_RX | - MV88E6390_G1_CTL2_HIST_MODE_TX); + MV88E6390_G1_CTL2_HIST_MODE_RX); } int mv88e6xxx_g1_set_device_number(struct mv88e6xxx_chip *chip, int index) @@ -491,7 +490,7 @@ int mv88e6095_g1_stats_set_histogram(struct mv88e6xxx_chip *chip) if (err) return err; - val |= MV88E6XXX_G1_STATS_OP_HIST_RX_TX; + val |= MV88E6XXX_G1_STATS_OP_HIST_RX; err = mv88e6xxx_g1_write(chip, MV88E6XXX_G1_STATS_OP, val); @@ -506,7 +505,7 @@ int mv88e6xxx_g1_stats_snapshot(struct mv88e6xxx_chip *chip, int port) err = mv88e6xxx_g1_write(chip, MV88E6XXX_G1_STATS_OP, MV88E6XXX_G1_STATS_OP_BUSY | MV88E6XXX_G1_STATS_OP_CAPTURE_PORT | - MV88E6XXX_G1_STATS_OP_HIST_RX_TX | port); + MV88E6XXX_G1_STATS_OP_HIST_RX | port); if (err) return err;
Chips in this family only has one set of histogram counters, which can be used to count ingressing and/or egressing traffic. mv88e6xxx has, up until this point, kept the hardware default of counting both directions. In the mean time, standard counter group support has been added to ethtool. Via that interface, drivers may report ingress-only and egress-only histograms separately - but not combined. In order for mv88e6xxx to maximalize amount of diagnostic information that can be exported via standard interfaces, we opt to limit the histogram counters to ingress traffic only. Which will allow us to export them via the standard "rmon" group in an upcoming commit. The reason for choosing ingress-only over egress-only, is to be compatible with RFC2819 (RMON MIB). Signed-off-by: Tobias Waldekranz <tobias@waldekranz.com> --- drivers/net/dsa/mv88e6xxx/chip.c | 6 +++--- drivers/net/dsa/mv88e6xxx/global1.c | 7 +++---- 2 files changed, 6 insertions(+), 7 deletions(-)