Re: [Intel-wired-lan] [PATCH net-next v2] drivers/net/ethernet: check return value of e1e_rphy()

On Mon, Aug 29, 2022 at 11:03 AM Neftin, Sasha [off-list ref] wrote:

On 8/26/2022 09:56, Li Zhong wrote:

quoted

e1e_rphy() could return error value, which needs to be checked and
reported for debugging and diagnose.

Signed-off-by: Li Zhong <redacted>
---
drivers/net/ethernet/intel/e1000e/phy.c | 20 +++++++++++++++++---
1 file changed, 17 insertions(+), 3 deletions(-)

diff --git a/drivers/net/ethernet/intel/e1000e/phy.c b/drivers/net/ethernet/intel/e1000e/phy.c
index fd07c3679bb1..060b263348ce 100644
--- a/drivers/net/ethernet/intel/e1000e/phy.c
+++ b/drivers/net/ethernet/intel/e1000e/phy.c

@@ -2697,9 +2697,14 @@ static s32 e1000_access_phy_wakeup_reg_bm(struct e1000_hw *hw, u32 offset,

void e1000_power_up_phy_copper(struct e1000_hw *hw)
{
u16 mii_reg = 0;
+ int ret;

/* The PHY will retain its settings across a power down/up cycle */
- e1e_rphy(hw, MII_BMCR, &mii_reg);
+ ret = e1e_rphy(hw, MII_BMCR, &mii_reg);
+ if (ret) {
+ e_dbg("Error reading PHY register\n");
+ return;
+ }
mii_reg &= ~BMCR_PDOWN;
e1e_wphy(hw, MII_BMCR, mii_reg);
}

@@ -2715,9 +2720,14 @@ void e1000_power_up_phy_copper(struct e1000_hw *hw)

void e1000_power_down_phy_copper(struct e1000_hw *hw)
{
u16 mii_reg = 0;
+ int ret;

/* The PHY will retain its settings across a power down/up cycle */
- e1e_rphy(hw, MII_BMCR, &mii_reg);
+ ret = e1e_rphy(hw, MII_BMCR, &mii_reg);
+ if (ret) {
+ e_dbg("Error reading PHY register\n");
+ return;
+ }
mii_reg |= BMCR_PDOWN;
e1e_wphy(hw, MII_BMCR, mii_reg);
usleep_range(1000, 2000);

@@ -3037,7 +3047,11 @@ s32 e1000_link_stall_workaround_hv(struct e1000_hw *hw)

return 0;

/* Do not apply workaround if in PHY loopback bit 14 set */
- e1e_rphy(hw, MII_BMCR, &data);
+ ret_val = e1e_rphy(hw, MII_BMCR, &data);
+ if (ret_val) {
+ e_dbg("Error reading PHY register\n");
+ return ret_val;
+ }
if (data & BMCR_LOOPBACK)
return 0;

Generally, I am good with this patch. One question - it is old HW, any
idea how to check it? (82577/82578 GbE LOM - from 2008)
Li, how did you check it manually?

These bugs are detected by static analysis. Therefore it's not run-time
checked. However, since currently there is no error handling after
e1e_rphy(), so I think it's necessary to at least check it and expect
there is a chance to fail.