linux/drivers/net/ethernet/micrel/ks8851_par.c
Marek Vasut 5c9fcac3c8 net: ks8851: Reinstate disabling of BHs around IRQ handler
If the driver executes ks8851_irq() AND a TX packet has been sent, then
the driver enables TX queue via netif_wake_queue() which schedules TX
softirq to queue packets for this device.

If CONFIG_PREEMPT_RT=y is set AND a packet has also been received by
the MAC, then ks8851_rx_pkts() calls netdev_alloc_skb_ip_align() to
allocate SKBs for the received packets. If netdev_alloc_skb_ip_align()
is called with BH enabled, then local_bh_enable() at the end of
netdev_alloc_skb_ip_align() will trigger the pending softirq processing,
which may ultimately call the .xmit callback ks8851_start_xmit_par().
The ks8851_start_xmit_par() will try to lock struct ks8851_net_par
.lock spinlock, which is already locked by ks8851_irq() from which
ks8851_start_xmit_par() was called. This leads to a deadlock, which
is reported by the kernel, including a trace listed below.

If CONFIG_PREEMPT_RT is not set, then since commit 0913ec336a
("net: ks8851: Fix deadlock with the SPI chip variant") the deadlock
can also be triggered without received packet in the RX FIFO. The
pending softirqs will be processed on return from
spin_unlock_bh(&ks->statelock) in ks8851_irq(), which triggers the
deadlock as well.

Fix the problem by disabling BH around critical sections, including the
IRQ handler, thus preventing the net_tx_action() softirq from triggering
during these critical sections. The net_tx_action() softirq is triggered
once BH are re-enabled and at the end of the IRQ handler, once all the
other IRQ handler actions have been completed.

 __schedule from schedule_rtlock+0x1c/0x34
 schedule_rtlock from rtlock_slowlock_locked+0x548/0x904
 rtlock_slowlock_locked from rt_spin_lock+0x60/0x9c
 rt_spin_lock from ks8851_start_xmit_par+0x74/0x1a8
 ks8851_start_xmit_par from netdev_start_xmit+0x20/0x44
 netdev_start_xmit from dev_hard_start_xmit+0xd0/0x188
 dev_hard_start_xmit from sch_direct_xmit+0xb8/0x25c
 sch_direct_xmit from __qdisc_run+0x1f8/0x4ec
 __qdisc_run from qdisc_run+0x1c/0x28
 qdisc_run from net_tx_action+0x1f0/0x268
 net_tx_action from handle_softirqs+0x1a4/0x270
 handle_softirqs from __local_bh_enable_ip+0xcc/0xe0
 __local_bh_enable_ip from __alloc_skb+0xd8/0x128
 __alloc_skb from __netdev_alloc_skb+0x3c/0x19c
 __netdev_alloc_skb from ks8851_irq+0x388/0x4d4
 ks8851_irq from irq_thread_fn+0x24/0x64
 irq_thread_fn from irq_thread+0x178/0x28c
 irq_thread from kthread+0x12c/0x138
 kthread from ret_from_fork+0x14/0x28

Reviewed-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Fixes: e0863634bf ("net: ks8851: Queue RX packets in IRQ handler instead of disabling BHs")
Cc: stable@vger.kernel.org
Signed-off-by: Marek Vasut <marex@nabladev.com>
Link: https://patch.msgid.link/20260415231020.455298-1-marex@nabladev.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2026-04-18 12:14:19 -07:00

344 lines
9.4 KiB
C

// SPDX-License-Identifier: GPL-2.0-only
/* drivers/net/ethernet/micrel/ks8851.c
*
* Copyright 2009 Simtec Electronics
* http://www.simtec.co.uk/
* Ben Dooks <ben@simtec.co.uk>
*/
#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
#include <linux/interrupt.h>
#include <linux/module.h>
#include <linux/kernel.h>
#include <linux/netdevice.h>
#include <linux/etherdevice.h>
#include <linux/ethtool.h>
#include <linux/iopoll.h>
#include <linux/mii.h>
#include <linux/platform_device.h>
#include <linux/of_net.h>
#include "ks8851.h"
static int msg_enable;
#define BE3 0x8000 /* Byte Enable 3 */
#define BE2 0x4000 /* Byte Enable 2 */
#define BE1 0x2000 /* Byte Enable 1 */
#define BE0 0x1000 /* Byte Enable 0 */
/**
* struct ks8851_net_par - KS8851 Parallel driver private data
* @ks8851: KS8851 driver common private data
* @lock: Lock to ensure that the device is not accessed when busy.
* @hw_addr : start address of data register.
* @hw_addr_cmd : start address of command register.
* @cmd_reg_cache : command register cached.
*
* The @lock ensures that the chip is protected when certain operations are
* in progress. When the read or write packet transfer is in progress, most
* of the chip registers are not accessible until the transfer is finished
* and the DMA has been de-asserted.
*/
struct ks8851_net_par {
struct ks8851_net ks8851;
spinlock_t lock;
void __iomem *hw_addr;
void __iomem *hw_addr_cmd;
u16 cmd_reg_cache;
};
#define to_ks8851_par(ks) container_of((ks), struct ks8851_net_par, ks8851)
/**
* ks8851_lock_par - register access lock
* @ks: The chip state
*
* Claim chip register access lock
*/
static void ks8851_lock_par(struct ks8851_net *ks)
{
struct ks8851_net_par *ksp = to_ks8851_par(ks);
spin_lock_bh(&ksp->lock);
}
/**
* ks8851_unlock_par - register access unlock
* @ks: The chip state
*
* Release chip register access lock
*/
static void ks8851_unlock_par(struct ks8851_net *ks)
{
struct ks8851_net_par *ksp = to_ks8851_par(ks);
spin_unlock_bh(&ksp->lock);
}
/**
* ks_check_endian - Check whether endianness of the bus is correct
* @ks : The chip information
*
* The KS8851-16MLL EESK pin allows selecting the endianness of the 16bit
* bus. To maintain optimum performance, the bus endianness should be set
* such that it matches the endianness of the CPU.
*/
static int ks_check_endian(struct ks8851_net *ks)
{
struct ks8851_net_par *ksp = to_ks8851_par(ks);
u16 cider;
/*
* Read CIDER register first, however read it the "wrong" way around.
* If the endian strap on the KS8851-16MLL in incorrect and the chip
* is operating in different endianness than the CPU, then the meaning
* of BE[3:0] byte-enable bits is also swapped such that:
* BE[3,2,1,0] becomes BE[1,0,3,2]
*
* Luckily for us, the byte-enable bits are the top four MSbits of
* the address register and the CIDER register is at offset 0xc0.
* Hence, by reading address 0xc0c0, which is not impacted by endian
* swapping, we assert either BE[3:2] or BE[1:0] while reading the
* CIDER register.
*
* If the bus configuration is correct, reading 0xc0c0 asserts
* BE[3:2] and this read returns 0x0000, because to read register
* with bottom two LSbits of address set to 0, BE[1:0] must be
* asserted.
*
* If the bus configuration is NOT correct, reading 0xc0c0 asserts
* BE[1:0] and this read returns non-zero 0x8872 value.
*/
iowrite16(BE3 | BE2 | KS_CIDER, ksp->hw_addr_cmd);
cider = ioread16(ksp->hw_addr);
if (!cider)
return 0;
netdev_err(ks->netdev, "incorrect EESK endian strap setting\n");
return -EINVAL;
}
/**
* ks8851_wrreg16_par - write 16bit register value to chip
* @ks: The chip state
* @reg: The register address
* @val: The value to write
*
* Issue a write to put the value @val into the register specified in @reg.
*/
static void ks8851_wrreg16_par(struct ks8851_net *ks, unsigned int reg,
unsigned int val)
{
struct ks8851_net_par *ksp = to_ks8851_par(ks);
ksp->cmd_reg_cache = (u16)reg | ((BE1 | BE0) << (reg & 0x02));
iowrite16(ksp->cmd_reg_cache, ksp->hw_addr_cmd);
iowrite16(val, ksp->hw_addr);
}
/**
* ks8851_rdreg16_par - read 16 bit register from chip
* @ks: The chip information
* @reg: The register address
*
* Read a 16bit register from the chip, returning the result
*/
static unsigned int ks8851_rdreg16_par(struct ks8851_net *ks, unsigned int reg)
{
struct ks8851_net_par *ksp = to_ks8851_par(ks);
ksp->cmd_reg_cache = (u16)reg | ((BE1 | BE0) << (reg & 0x02));
iowrite16(ksp->cmd_reg_cache, ksp->hw_addr_cmd);
return ioread16(ksp->hw_addr);
}
/**
* ks8851_rdfifo_par - read data from the receive fifo
* @ks: The device state.
* @buff: The buffer address
* @len: The length of the data to read
*
* Issue an RXQ FIFO read command and read the @len amount of data from
* the FIFO into the buffer specified by @buff.
*/
static void ks8851_rdfifo_par(struct ks8851_net *ks, u8 *buff, unsigned int len)
{
struct ks8851_net_par *ksp = to_ks8851_par(ks);
netif_dbg(ks, rx_status, ks->netdev,
"%s: %d@%p\n", __func__, len, buff);
ioread16_rep(ksp->hw_addr, (u16 *)buff + 1, len / 2);
}
/**
* ks8851_wrfifo_par - write packet to TX FIFO
* @ks: The device state.
* @txp: The sk_buff to transmit.
* @irq: IRQ on completion of the packet.
*
* Send the @txp to the chip. This means creating the relevant packet header
* specifying the length of the packet and the other information the chip
* needs, such as IRQ on completion. Send the header and the packet data to
* the device.
*/
static void ks8851_wrfifo_par(struct ks8851_net *ks, struct sk_buff *txp,
bool irq)
{
struct ks8851_net_par *ksp = to_ks8851_par(ks);
unsigned int len = ALIGN(txp->len, 4);
unsigned int fid = 0;
netif_dbg(ks, tx_queued, ks->netdev, "%s: skb %p, %d@%p, irq %d\n",
__func__, txp, txp->len, txp->data, irq);
fid = ks->fid++;
fid &= TXFR_TXFID_MASK;
if (irq)
fid |= TXFR_TXIC; /* irq on completion */
iowrite16(fid, ksp->hw_addr);
iowrite16(txp->len, ksp->hw_addr);
iowrite16_rep(ksp->hw_addr, txp->data, len / 2);
}
static unsigned int ks8851_rdreg16_par_txqcr(struct ks8851_net *ks)
{
return ks8851_rdreg16_par(ks, KS_TXQCR);
}
/**
* ks8851_start_xmit_par - transmit packet
* @skb: The buffer to transmit
* @dev: The device used to transmit the packet.
*
* Called by the network layer to transmit the @skb. Queue the packet for
* the device and schedule the necessary work to transmit the packet when
* it is free.
*
* We do this to firstly avoid sleeping with the network device locked,
* and secondly so we can round up more than one packet to transmit which
* means we can try and avoid generating too many transmit done interrupts.
*/
static netdev_tx_t ks8851_start_xmit_par(struct sk_buff *skb,
struct net_device *dev)
{
struct ks8851_net *ks = netdev_priv(dev);
netdev_tx_t ret = NETDEV_TX_OK;
unsigned int txqcr;
u16 txmir;
int err;
netif_dbg(ks, tx_queued, ks->netdev,
"%s: skb %p, %d@%p\n", __func__, skb, skb->len, skb->data);
ks8851_lock_par(ks);
txmir = ks8851_rdreg16_par(ks, KS_TXMIR) & 0x1fff;
if (likely(txmir >= skb->len + 12)) {
ks8851_wrreg16_par(ks, KS_RXQCR, ks->rc_rxqcr | RXQCR_SDA);
ks8851_wrfifo_par(ks, skb, false);
ks8851_wrreg16_par(ks, KS_RXQCR, ks->rc_rxqcr);
ks8851_wrreg16_par(ks, KS_TXQCR, TXQCR_METFE);
err = readx_poll_timeout_atomic(ks8851_rdreg16_par_txqcr, ks,
txqcr, !(txqcr & TXQCR_METFE),
5, 1000000);
if (err)
ret = NETDEV_TX_BUSY;
ks8851_done_tx(ks, skb);
} else {
ret = NETDEV_TX_BUSY;
}
ks8851_unlock_par(ks);
return ret;
}
static int ks8851_probe_par(struct platform_device *pdev)
{
struct device *dev = &pdev->dev;
struct ks8851_net_par *ksp;
struct net_device *netdev;
struct ks8851_net *ks;
int ret;
netdev = devm_alloc_etherdev(dev, sizeof(struct ks8851_net_par));
if (!netdev)
return -ENOMEM;
ks = netdev_priv(netdev);
ks->lock = ks8851_lock_par;
ks->unlock = ks8851_unlock_par;
ks->rdreg16 = ks8851_rdreg16_par;
ks->wrreg16 = ks8851_wrreg16_par;
ks->rdfifo = ks8851_rdfifo_par;
ks->wrfifo = ks8851_wrfifo_par;
ks->start_xmit = ks8851_start_xmit_par;
#define STD_IRQ (IRQ_LCI | /* Link Change */ \
IRQ_RXI | /* RX done */ \
IRQ_RXPSI) /* RX process stop */
ks->rc_ier = STD_IRQ;
ksp = to_ks8851_par(ks);
spin_lock_init(&ksp->lock);
ksp->hw_addr = devm_platform_ioremap_resource(pdev, 0);
if (IS_ERR(ksp->hw_addr))
return PTR_ERR(ksp->hw_addr);
ksp->hw_addr_cmd = devm_platform_ioremap_resource(pdev, 1);
if (IS_ERR(ksp->hw_addr_cmd))
return PTR_ERR(ksp->hw_addr_cmd);
ret = ks_check_endian(ks);
if (ret)
return ret;
netdev->irq = platform_get_irq(pdev, 0);
if (netdev->irq < 0)
return netdev->irq;
return ks8851_probe_common(netdev, dev, msg_enable);
}
static void ks8851_remove_par(struct platform_device *pdev)
{
ks8851_remove_common(&pdev->dev);
}
static const struct of_device_id ks8851_match_table[] = {
{ .compatible = "micrel,ks8851-mll" },
{ }
};
MODULE_DEVICE_TABLE(of, ks8851_match_table);
static struct platform_driver ks8851_driver = {
.driver = {
.name = "ks8851",
.of_match_table = ks8851_match_table,
.pm = &ks8851_pm_ops,
},
.probe = ks8851_probe_par,
.remove = ks8851_remove_par,
};
module_platform_driver(ks8851_driver);
MODULE_DESCRIPTION("KS8851 Network driver");
MODULE_AUTHOR("Ben Dooks <ben@simtec.co.uk>");
MODULE_LICENSE("GPL");
module_param_named(message, msg_enable, int, 0);
MODULE_PARM_DESC(message, "Message verbosity level (0=none, 31=all)");