]> git.ipfire.org Git - thirdparty/linux.git/commit
net: sfp: initialize i2c_block_size at adapter configure time
authorJonas Jelonek <jelonek.jonas@gmail.com>
Thu, 28 May 2026 20:52:40 +0000 (20:52 +0000)
committerJakub Kicinski <kuba@kernel.org>
Thu, 4 Jun 2026 00:57:31 +0000 (17:57 -0700)
commit56d0885514491e5ed8f7593400879ab77c52504c
treebc297b7f89226f2afc94c65a48d935481707bd46
parent22ba97ea9cc1f63a0d0244fae38057ed452b6ac7
net: sfp: initialize i2c_block_size at adapter configure time

sfp->i2c_block_size is only assigned in sfp_sm_mod_probe(), which runs
from the state machine timer after SFP_F_PRESENT has been set. Between
those two points, sfp_module_eeprom() (the ethtool -m callback) gates
only on SFP_F_PRESENT and can be entered with i2c_block_size still at
its kzalloc'd value of 0.

On a pure-I2C adapter, sfp_i2c_read() then issues an i2c_transfer()
with msgs[1].len = 0 inside a loop that subtracts this_len from len
each iteration; on adapters that succeed a zero-length read the loop
never advances, spinning while holding rtnl_lock.

This was previously addressed by initializing i2c_block_size in
sfp_alloc() (commit 813c2dd78618), but the initialization was dropped
when i2c_block_size was split from i2c_max_block_size.

Initialize sfp->i2c_block_size from sfp->i2c_max_block_size in
sfp_i2c_configure(), so the field is valid as soon as the adapter is
known. sfp_sm_mod_probe() still reassigns it on each module insertion
to recover from a per-module clamp to 1 (sfp_id_needs_byte_io).

Fixes: 7662abf4db94 ("net: phy: sfp: Add support for SMBus module access")
Cc: stable@vger.kernel.org
Signed-off-by: Jonas Jelonek <jelonek.jonas@gmail.com>
Link: https://patch.msgid.link/20260528205242.971410-2-jelonek.jonas@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
drivers/net/phy/sfp.c