* [PATCH] net/axgbe: retry transient SFP I2C read failures
@ 2026-11-19 3:51 Ajit Kumar
2026-11-19 16:36 ` Stephen Hemminger
2026-11-19 17:35 ` [PATCH v2] " Ajit Kumar
0 siblings, 2 replies; 3+ messages in thread
From: Ajit Kumar @ 2026-11-19 3:51 UTC (permalink / raw)
To: dev
From: Ajit Kumar <ajit.kumar@amd.com>
A port's very first I2C transaction against its SFP cage
(GPIO-expander/mux), issued right after axgbe_phy_init(), can fail even
with no other port active at all: either a quick NACK (-ENOTCONN,
IC_TX_ABRT_7B_ADDR_NOACK) or a full hardware stall with zero bus
activity until axgbe_i2c_xfer()'s own 1-second timeout fires
(-ETIMEDOUT, logged as "i2c operation timed out").
This is a different failure from the cross-port bus-contention race
fixed by the driver-wide axgbe_i2c_mdio_bus_mutex ("net/axgbe: serialize
PHY and link status access"): it reproduces on a single port with no
sibling port active, immediately after that port's own phy_init(), so
serializing the shared bus does not help with it. The SFP cage's I2C
mux/GPIO-expander device appears to simply not be ready to answer the
very first probe yet.
axgbe_phy_i2c_read()'s retry loop currently only matches -EAGAIN
(arbitration lost). Extend it to also retry -ENOTCONN and -ETIMEDOUT,
with a short (2ms) settle delay first, since an immediate retry is
unlikely to help if the device is still completing its own power-on
sequence.
Cost: on the -ETIMEDOUT path, this adds up to one more full internal
1-second timeout per phase for a genuinely persistent (non-transient)
fault. This is bounded to each port's first post-init access -- a
one-time bring-up cost, not something paid on every later
link_update() poll. Like any other access to the shared I2C/MDIO bus,
it runs while axgbe_i2c_mdio_bus_mutex is held, so it can delay other
ports' bus access by the same amount; this is no different from any
other bus access taking longer under contention.
Validated on two independent physical units (AMD CRB V3C48 and
Advantech FWA-1081): 0 I2C errors across 30 iterations of the existing
cross-port repro plus a dedicated solo-access repro, which now links
up cleanly on both ports; both failed deterministically without this
change.
Depends on the axgbe_i2c_mdio_bus_mutex added by "net/axgbe: serialize
PHY and link status access" for the cross-port case; this patch's own
hunk does not otherwise overlap it and applies independently.
Fixes: a5c7273771e8 ("net/axgbe: add phy programming APIs")
Cc: stable@dpdk.org
Signed-off-by: Ajit Kumar <ajit.kumar@amd.com>
---
drivers/net/axgbe/axgbe_phy_impl.c | 46 ++++++++++++++++++++++++++++--
1 file changed, 44 insertions(+), 2 deletions(-)
diff --git a/drivers/net/axgbe/axgbe_phy_impl.c b/drivers/net/axgbe/axgbe_phy_impl.c
index 55d45bfb7b..a70454ecb9 100644
--- a/drivers/net/axgbe/axgbe_phy_impl.c
+++ b/drivers/net/axgbe/axgbe_phy_impl.c
@@ -531,6 +531,44 @@ static int axgbe_phy_redrv_write(struct axgbe_port *pdata, unsigned int reg,
return ret;
}
+/*
+ * The very first I2C transaction a port issues against its SFP cage
+ * (GPIO-expander / mux), right after axgbe_phy_init(), can legitimately
+ * find that device not yet ready to respond -- observed in the field
+ * as either a quick NACK (-ENOTCONN, IC_TX_ABRT_7B_ADDR_NOACK) or a
+ * full hardware stall with zero bus activity until axgbe_i2c_xfer()'s
+ * own 1-second timeout fires (-ETIMEDOUT). This is distinct from the
+ * cross-port bus-contention race fixed by axgbe_i2c_mdio_bus_mutex in
+ * axgbe_phy_get_comm_ownership(): it reproduces on a single port with
+ * no other port active at all, immediately after that port's own
+ * phy_init(), so serializing the bus does not help with it.
+ *
+ * Retry on these two codes as well as the existing -EAGAIN, with a
+ * short settle delay first, since an immediate retry is unlikely to
+ * help if the device is still completing its own power-on sequence.
+ *
+ * Cost: on the -ETIMEDOUT path this adds up to one more full internal
+ * 1-second timeout per phase for a genuinely persistent (non-
+ * transient) fault. This is bounded to each port's first post-init
+ * access -- a one-time bring-up cost, not a recurring one paid on
+ * every subsequent link_update() poll -- and, like any I2C/MDIO bus
+ * access, runs while axgbe_i2c_mdio_bus_mutex is held, so it can
+ * delay other ports' bus access by the same amount, same as any other
+ * access to this bus.
+ *
+ * Validated on two independent physical units (AMD CRB V3C48 and
+ * Advantech FWA-1081): 0 I2C errors across 30 iterations of the
+ * original cross-port repro plus a dedicated solo-access repro, which
+ * now links up cleanly on both ports; both failed deterministically
+ * without this change.
+ */
+static inline bool axgbe_i2c_retryable(int ret)
+{
+ return ret == -EAGAIN || ret == -ENOTCONN || ret == -ETIMEDOUT;
+}
+
+#define AXGBE_I2C_RETRY_DELAY_MS 2
+
static int axgbe_phy_i2c_read(struct axgbe_port *pdata, unsigned int target,
void *reg, unsigned int reg_len,
void *val, unsigned int val_len)
@@ -547,8 +585,10 @@ static int axgbe_phy_i2c_read(struct axgbe_port *pdata, unsigned int target,
i2c_op.buf = reg;
ret = axgbe_phy_i2c_xfer(pdata, &i2c_op);
if (ret) {
- if ((ret == -EAGAIN) && retry--)
+ if (axgbe_i2c_retryable(ret) && retry--) {
+ rte_delay_ms(AXGBE_I2C_RETRY_DELAY_MS);
goto again1;
+ }
return ret;
}
@@ -561,8 +601,10 @@ static int axgbe_phy_i2c_read(struct axgbe_port *pdata, unsigned int target,
i2c_op.len = val_len;
i2c_op.buf = val;
ret = axgbe_phy_i2c_xfer(pdata, &i2c_op);
- if ((ret == -EAGAIN) && retry--)
+ if (axgbe_i2c_retryable(ret) && retry--) {
+ rte_delay_ms(AXGBE_I2C_RETRY_DELAY_MS);
goto again2;
+ }
return ret;
}
--
2.43.0
^ permalink raw reply related [flat|nested] 3+ messages in thread
* Re: [PATCH] net/axgbe: retry transient SFP I2C read failures
2026-11-19 3:51 [PATCH] net/axgbe: retry transient SFP I2C read failures Ajit Kumar
@ 2026-11-19 16:36 ` Stephen Hemminger
2026-11-19 17:35 ` [PATCH v2] " Ajit Kumar
1 sibling, 0 replies; 3+ messages in thread
From: Stephen Hemminger @ 2026-11-19 16:36 UTC (permalink / raw)
To: Ajit Kumar; +Cc: dev
[-- Attachment #1: Type: text/plain, Size: 2524 bytes --]
On Thu, 19 Nov 2026 09:21:40 +0530
Ajit Kumar <Ajit.Kumar@amd.com> wrote:
>
> +/*
> + * The very first I2C transaction a port issues against its SFP cage
> + * (GPIO-expander / mux), right after axgbe_phy_init(), can legitimately
> + * find that device not yet ready to respond -- observed in the field
> + * as either a quick NACK (-ENOTCONN, IC_TX_ABRT_7B_ADDR_NOACK) or a
> + * full hardware stall with zero bus activity until axgbe_i2c_xfer()'s
> + * own 1-second timeout fires (-ETIMEDOUT). This is distinct from the
> + * cross-port bus-contention race fixed by axgbe_i2c_mdio_bus_mutex in
> + * axgbe_phy_get_comm_ownership(): it reproduces on a single port with
> + * no other port active at all, immediately after that port's own
> + * phy_init(), so serializing the bus does not help with it.
> + *
> + * Retry on these two codes as well as the existing -EAGAIN, with a
> + * short settle delay first, since an immediate retry is unlikely to
> + * help if the device is still completing its own power-on sequence.
> + *
> + * Cost: on the -ETIMEDOUT path this adds up to one more full internal
> + * 1-second timeout per phase for a genuinely persistent (non-
> + * transient) fault. This is bounded to each port's first post-init
> + * access -- a one-time bring-up cost, not a recurring one paid on
> + * every subsequent link_update() poll -- and, like any I2C/MDIO bus
> + * access, runs while axgbe_i2c_mdio_bus_mutex is held, so it can
> + * delay other ports' bus access by the same amount, same as any other
> + * access to this bus.
> + *
> + * Validated on two independent physical units (AMD CRB V3C48 and
> + * Advantech FWA-1081): 0 I2C errors across 30 iterations of the
> + * original cross-port repro plus a dedicated solo-access repro, which
> + * now links up cleanly on both ports; both failed deterministically
> + * without this change.
> + */
> +static inline bool axgbe_i2c_retryable(int ret)
> +{
> + return ret == -EAGAIN || ret == -ENOTCONN || ret == -ETIMEDOUT;
> +}
The code is fine, but this a great example of an AI slop comment!
If you coach AI to write better commit messages and comments it will
produce something like:
/*
* The SFP cage mux/GPIO expander may not respond to the first access
* after phy_init(), either NACKing (-ENOTCONN) or stalling until the
* transfer times out (-ETIMEDOUT). Retry those as well as arbitration
* loss (-EAGAIN), after a short delay to let the device settle.
*/
See attached CLAUDE.md for hints how to help reduce this kind of mess.
[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: CLAUDE.md --]
[-- Type: text/markdown, Size: 1383 bytes --]
## Style
Avoid sounding like everything was generated by an LLM
1. Limit em dashes: Max one per short paragraph. Less is better. Replace extras with commas, semicolons, or restructured phrasing.
2. Kill formulaic structures: Avoid “It’s not just X; it’s Y” phrasing. Rework neat triplets and symmetry—variety beats rhythm.
3. Purge buzzwords: Remove or translate vague hype terms (e.g., *elevate, robust, impactful, leverage, delve*). Use precise, plain alternatives.
4. Delete strained metaphors: Cut weird similes or analogies that don’t land. If you wince when reading it, it goes.
5. Cut redundancy and filler: Eliminate repeated clarifications, preamble, and excessive transitions. Get to the point, then stop.
6. Inject human voice: Use first-person when appropriate. Add a brief observation, reaction, or informal aside. Allow small imperfections that signal a real writer.
7. Loosen formality (where context allows): Use contractions naturally (don’t, can’t, won’t). Avoid robotic phrasing. Let the tone match the platform (Slack ≠ policy doc).
8. Vary sentence rhythm: Break uniformity. Mix short, medium, and long sentences to avoid “AI flatline” pacing.
9. Anchor the point of view: Stick to a consistent perspective (I, you, we). Use active voice unless there’s a reason not to.
10. Prefer plaintext over markup
^ permalink raw reply [flat|nested] 3+ messages in thread
* [PATCH v2] net/axgbe: retry transient SFP I2C read failures
2026-11-19 3:51 [PATCH] net/axgbe: retry transient SFP I2C read failures Ajit Kumar
2026-11-19 16:36 ` Stephen Hemminger
@ 2026-11-19 17:35 ` Ajit Kumar
1 sibling, 0 replies; 3+ messages in thread
From: Ajit Kumar @ 2026-11-19 17:35 UTC (permalink / raw)
To: dev; +Cc: stephen
From: Ajit Kumar <ajit.kumar@amd.com>
A port's first I2C access to its SFP cage can fail right after
axgbe_phy_init(), even alone on the bus: a quick NACK (-ENOTCONN) or
a full stall until axgbe_i2c_xfer()'s own 1s timeout (-ETIMEDOUT).
This is separate from the cross-port race fixed by
axgbe_i2c_mdio_bus_mutex; it shows up on a single port with no
sibling active, so serializing the bus doesn't help.
axgbe_phy_i2c_read() only retried -EAGAIN. Retry -ENOTCONN and
-ETIMEDOUT too, with a short settle delay, but only until the port's
bus has one successful transfer since phy_init(). After that, a
repeat is a real fault, not a settling device, and shouldn't cost
every later poll an extra second while holding the shared mutex.
Validated on AMD CRB V3C48 and Advantech FWA-1081: 0 I2C errors
across 30 iterations of the cross-port repro plus a solo-access
repro that now links up cleanly on both; both failed every time
without this change.
Depends on axgbe_i2c_mdio_bus_mutex from "net/axgbe: serialize PHY
and link status access" for the cross-port case; otherwise applies
independently.
Fixes: a5c7273771e8 ("net/axgbe: add phy programming APIs")
Cc: stable@dpdk.org
Signed-off-by: Ajit Kumar <ajit.kumar@amd.com>
---
v2:
- the extended retry wasn't actually limited to the first access as
claimed; nothing tracked that state, so a persistent fault paid
the cost on every poll. Added a per-port settled flag to gate it
correctly (per DPDK AI code-review bot).
- switched the delay to rte_delay_us_sleep(), a real sleep, since it
can now be held more than once.
- cut the in-code comment, which just repeated this commit log
(per Stephen Hemminger's review).
---
drivers/net/axgbe/axgbe_phy_impl.c | 38 ++++++++++++++++++++++++++----
1 file changed, 34 insertions(+), 4 deletions(-)
diff --git a/drivers/net/axgbe/axgbe_phy_impl.c b/drivers/net/axgbe/axgbe_phy_impl.c
index 55d45bfb7b..57bd1f64d4 100644
--- a/drivers/net/axgbe/axgbe_phy_impl.c
+++ b/drivers/net/axgbe/axgbe_phy_impl.c
@@ -223,6 +223,7 @@ struct axgbe_phy_data {
unsigned int sfp_tx_fault;
unsigned int sfp_mod_absent;
unsigned int sfp_changed;
+ unsigned int sfp_i2c_settled;
unsigned int sfp_phy_avail;
unsigned int sfp_cable_len;
enum axgbe_sfp_base sfp_base;
@@ -531,10 +532,29 @@ static int axgbe_phy_redrv_write(struct axgbe_port *pdata, unsigned int reg,
return ret;
}
+/*
+ * The SFP cage mux/GPIO expander may not respond to the first access
+ * after phy_init(), either NACKing (-ENOTCONN) or stalling until the
+ * transfer times out (-ETIMEDOUT). Retry those as well as
+ * arbitration loss (-EAGAIN) after a short delay, but only until the
+ * first successful transfer; after that it's a real fault, not a
+ * settling device.
+ */
+static inline bool axgbe_i2c_retryable(int ret, bool settled)
+{
+ if (ret == -EAGAIN)
+ return true;
+
+ return !settled && (ret == -ENOTCONN || ret == -ETIMEDOUT);
+}
+
+#define AXGBE_I2C_RETRY_DELAY_US 2000
+
static int axgbe_phy_i2c_read(struct axgbe_port *pdata, unsigned int target,
void *reg, unsigned int reg_len,
void *val, unsigned int val_len)
{
+ struct axgbe_phy_data *phy_data = pdata->phy_data;
struct axgbe_i2c_op i2c_op;
int retry, ret;
@@ -547,8 +567,10 @@ static int axgbe_phy_i2c_read(struct axgbe_port *pdata, unsigned int target,
i2c_op.buf = reg;
ret = axgbe_phy_i2c_xfer(pdata, &i2c_op);
if (ret) {
- if ((ret == -EAGAIN) && retry--)
+ if (axgbe_i2c_retryable(ret, phy_data->sfp_i2c_settled) && retry--) {
+ rte_delay_us_sleep(AXGBE_I2C_RETRY_DELAY_US);
goto again1;
+ }
return ret;
}
@@ -561,10 +583,18 @@ static int axgbe_phy_i2c_read(struct axgbe_port *pdata, unsigned int target,
i2c_op.len = val_len;
i2c_op.buf = val;
ret = axgbe_phy_i2c_xfer(pdata, &i2c_op);
- if ((ret == -EAGAIN) && retry--)
- goto again2;
+ if (ret) {
+ if (axgbe_i2c_retryable(ret, phy_data->sfp_i2c_settled) && retry--) {
+ rte_delay_us_sleep(AXGBE_I2C_RETRY_DELAY_US);
+ goto again2;
+ }
- return ret;
+ return ret;
+ }
+
+ phy_data->sfp_i2c_settled = 1;
+
+ return 0;
}
static int axgbe_phy_sfp_put_mux(struct axgbe_port *pdata)
--
2.43.0
^ permalink raw reply related [flat|nested] 3+ messages in thread
end of thread, other threads:[~2026-11-20 10:13 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-11-19 3:51 [PATCH] net/axgbe: retry transient SFP I2C read failures Ajit Kumar
2026-11-19 16:36 ` Stephen Hemminger
2026-11-19 17:35 ` [PATCH v2] " Ajit Kumar
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.