lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <20220111171424.862764-12-Jerome.Pouiller@silabs.com>
Date:   Tue, 11 Jan 2022 18:14:11 +0100
From:   Jerome Pouiller <Jerome.Pouiller@...abs.com>
To:     linux-wireless@...r.kernel.org, netdev@...r.kernel.org,
        Kalle Valo <kvalo@...eaurora.org>
Cc:     devel@...verdev.osuosl.org, linux-kernel@...r.kernel.org,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        "David S . Miller" <davem@...emloft.net>,
        devicetree@...r.kernel.org, Rob Herring <robh+dt@...nel.org>,
        linux-mmc@...r.kernel.org,
        Pali Rohár <pali@...nel.org>,
        Ulf Hansson <ulf.hansson@...aro.org>,
        Jérôme Pouiller 
        <jerome.pouiller@...abs.com>
Subject: [PATCH v9 11/24] wfx: add bh.c/bh.h

From: Jérôme Pouiller <jerome.pouiller@...abs.com>

Signed-off-by: Jérôme Pouiller <jerome.pouiller@...abs.com>
---
 drivers/net/wireless/silabs/wfx/bh.c | 323 +++++++++++++++++++++++++++
 drivers/net/wireless/silabs/wfx/bh.h |  34 +++
 2 files changed, 357 insertions(+)
 create mode 100644 drivers/net/wireless/silabs/wfx/bh.c
 create mode 100644 drivers/net/wireless/silabs/wfx/bh.h

diff --git a/drivers/net/wireless/silabs/wfx/bh.c b/drivers/net/wireless/silabs/wfx/bh.c
new file mode 100644
index 000000000000..4c6ba9c342a6
--- /dev/null
+++ b/drivers/net/wireless/silabs/wfx/bh.c
@@ -0,0 +1,323 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Interrupt bottom half (BH).
+ *
+ * Copyright (c) 2017-2020, Silicon Laboratories, Inc.
+ * Copyright (c) 2010, ST-Ericsson
+ */
+#include <linux/gpio/consumer.h>
+#include <net/mac80211.h>
+
+#include "bh.h"
+#include "wfx.h"
+#include "hwio.h"
+#include "traces.h"
+#include "hif_rx.h"
+#include "hif_api_cmd.h"
+
+static void device_wakeup(struct wfx_dev *wdev)
+{
+	int max_retry = 3;
+
+	if (!wdev->pdata.gpio_wakeup)
+		return;
+	if (gpiod_get_value_cansleep(wdev->pdata.gpio_wakeup) > 0)
+		return;
+
+	if (wfx_api_older_than(wdev, 1, 4)) {
+		gpiod_set_value_cansleep(wdev->pdata.gpio_wakeup, 1);
+		if (!completion_done(&wdev->hif.ctrl_ready))
+			usleep_range(2000, 2500);
+		return;
+	}
+	for (;;) {
+		gpiod_set_value_cansleep(wdev->pdata.gpio_wakeup, 1);
+		/* completion.h does not provide any function to wait completion without consume it
+		 * (a kind of wait_for_completion_done_timeout()). So we have to emulate it.
+		 */
+		if (wait_for_completion_timeout(&wdev->hif.ctrl_ready, msecs_to_jiffies(2))) {
+			complete(&wdev->hif.ctrl_ready);
+			return;
+		} else if (max_retry-- > 0) {
+			/* Older firmwares have a race in sleep/wake-up process.  Redo the process
+			 * is sufficient to unfreeze the chip.
+			 */
+			dev_err(wdev->dev, "timeout while wake up chip\n");
+			gpiod_set_value_cansleep(wdev->pdata.gpio_wakeup, 0);
+			usleep_range(2000, 2500);
+		} else {
+			dev_err(wdev->dev, "max wake-up retries reached\n");
+			return;
+		}
+	}
+}
+
+static void device_release(struct wfx_dev *wdev)
+{
+	if (!wdev->pdata.gpio_wakeup)
+		return;
+
+	gpiod_set_value_cansleep(wdev->pdata.gpio_wakeup, 0);
+}
+
+static int rx_helper(struct wfx_dev *wdev, size_t read_len, int *is_cnf)
+{
+	struct sk_buff *skb;
+	struct wfx_hif_msg *hif;
+	size_t alloc_len;
+	size_t computed_len;
+	int release_count;
+	int piggyback = 0;
+
+	WARN(read_len > round_down(0xFFF, 2) * sizeof(u16), "request exceed the chip capability");
+
+	/* Add 2 to take into account piggyback size */
+	alloc_len = wdev->hwbus_ops->align_size(wdev->hwbus_priv, read_len + 2);
+	skb = dev_alloc_skb(alloc_len);
+	if (!skb)
+		return -ENOMEM;
+
+	if (wfx_data_read(wdev, skb->data, alloc_len))
+		goto err;
+
+	piggyback = le16_to_cpup((__le16 *)(skb->data + alloc_len - 2));
+	_trace_piggyback(piggyback, false);
+
+	hif = (struct wfx_hif_msg *)skb->data;
+	WARN(hif->encrypted & 0x3, "encryption is unsupported");
+	if (WARN(read_len < sizeof(struct wfx_hif_msg), "corrupted read"))
+		goto err;
+	computed_len = le16_to_cpu(hif->len);
+	computed_len = round_up(computed_len, 2);
+	if (computed_len != read_len) {
+		dev_err(wdev->dev, "inconsistent message length: %zu != %zu\n",
+			computed_len, read_len);
+		print_hex_dump(KERN_INFO, "hif: ", DUMP_PREFIX_OFFSET, 16, 1,
+			       hif, read_len, true);
+		goto err;
+	}
+
+	if (!(hif->id & HIF_ID_IS_INDICATION)) {
+		(*is_cnf)++;
+		if (hif->id == HIF_CNF_ID_MULTI_TRANSMIT)
+			release_count =
+				((struct wfx_hif_cnf_multi_transmit *)hif->body)->num_tx_confs;
+		else
+			release_count = 1;
+		WARN(wdev->hif.tx_buffers_used < release_count, "corrupted buffer counter");
+		wdev->hif.tx_buffers_used -= release_count;
+	}
+	_trace_hif_recv(hif, wdev->hif.tx_buffers_used);
+
+	if (hif->id != HIF_IND_ID_EXCEPTION && hif->id != HIF_IND_ID_ERROR) {
+		if (hif->seqnum != wdev->hif.rx_seqnum)
+			dev_warn(wdev->dev, "wrong message sequence: %d != %d\n",
+				 hif->seqnum, wdev->hif.rx_seqnum);
+		wdev->hif.rx_seqnum = (hif->seqnum + 1) % (HIF_COUNTER_MAX + 1);
+	}
+
+	skb_put(skb, le16_to_cpu(hif->len));
+	/* wfx_handle_rx takes care on SKB livetime */
+	wfx_handle_rx(wdev, skb);
+	if (!wdev->hif.tx_buffers_used)
+		wake_up(&wdev->hif.tx_buffers_empty);
+
+	return piggyback;
+
+err:
+	if (skb)
+		dev_kfree_skb(skb);
+	return -EIO;
+}
+
+static int bh_work_rx(struct wfx_dev *wdev, int max_msg, int *num_cnf)
+{
+	size_t len;
+	int i;
+	int ctrl_reg, piggyback;
+
+	piggyback = 0;
+	for (i = 0; i < max_msg; i++) {
+		if (piggyback & CTRL_NEXT_LEN_MASK)
+			ctrl_reg = piggyback;
+		else if (try_wait_for_completion(&wdev->hif.ctrl_ready))
+			ctrl_reg = atomic_xchg(&wdev->hif.ctrl_reg, 0);
+		else
+			ctrl_reg = 0;
+		if (!(ctrl_reg & CTRL_NEXT_LEN_MASK))
+			return i;
+		/* ctrl_reg units are 16bits words */
+		len = (ctrl_reg & CTRL_NEXT_LEN_MASK) * 2;
+		piggyback = rx_helper(wdev, len, num_cnf);
+		if (piggyback < 0)
+			return i;
+		if (!(piggyback & CTRL_WLAN_READY))
+			dev_err(wdev->dev, "unexpected piggyback value: ready bit not set: %04x\n",
+				piggyback);
+	}
+	if (piggyback & CTRL_NEXT_LEN_MASK) {
+		ctrl_reg = atomic_xchg(&wdev->hif.ctrl_reg, piggyback);
+		complete(&wdev->hif.ctrl_ready);
+		if (ctrl_reg)
+			dev_err(wdev->dev, "unexpected IRQ happened: %04x/%04x\n",
+				ctrl_reg, piggyback);
+	}
+	return i;
+}
+
+static void tx_helper(struct wfx_dev *wdev, struct wfx_hif_msg *hif)
+{
+	int ret;
+	void *data;
+	bool is_encrypted = false;
+	size_t len = le16_to_cpu(hif->len);
+
+	WARN(len < sizeof(*hif), "try to send corrupted data");
+
+	hif->seqnum = wdev->hif.tx_seqnum;
+	wdev->hif.tx_seqnum = (wdev->hif.tx_seqnum + 1) % (HIF_COUNTER_MAX + 1);
+
+	data = hif;
+	WARN(len > le16_to_cpu(wdev->hw_caps.size_inp_ch_buf),
+	     "request exceed the chip capability: %zu > %d\n",
+	     len, le16_to_cpu(wdev->hw_caps.size_inp_ch_buf));
+	len = wdev->hwbus_ops->align_size(wdev->hwbus_priv, len);
+	ret = wfx_data_write(wdev, data, len);
+	if (ret)
+		goto end;
+
+	wdev->hif.tx_buffers_used++;
+	_trace_hif_send(hif, wdev->hif.tx_buffers_used);
+end:
+	if (is_encrypted)
+		kfree(data);
+}
+
+static int bh_work_tx(struct wfx_dev *wdev, int max_msg)
+{
+	struct wfx_hif_msg *hif;
+	int i;
+
+	for (i = 0; i < max_msg; i++) {
+		hif = NULL;
+		if (wdev->hif.tx_buffers_used < le16_to_cpu(wdev->hw_caps.num_inp_ch_bufs)) {
+			if (try_wait_for_completion(&wdev->hif_cmd.ready)) {
+				WARN(!mutex_is_locked(&wdev->hif_cmd.lock), "data locking error");
+				hif = wdev->hif_cmd.buf_send;
+			} else {
+				hif = wfx_tx_queues_get(wdev);
+			}
+		}
+		if (!hif)
+			return i;
+		tx_helper(wdev, hif);
+	}
+	return i;
+}
+
+/* In SDIO mode, it is necessary to make an access to a register to acknowledge last received
+ * message. It could be possible to restrict this acknowledge to SDIO mode and only if last
+ * operation was rx.
+ */
+static void ack_sdio_data(struct wfx_dev *wdev)
+{
+	u32 cfg_reg;
+
+	wfx_config_reg_read(wdev, &cfg_reg);
+	if (cfg_reg & 0xFF) {
+		dev_warn(wdev->dev, "chip reports errors: %02x\n", cfg_reg & 0xFF);
+		wfx_config_reg_write_bits(wdev, 0xFF, 0x00);
+	}
+}
+
+static void bh_work(struct work_struct *work)
+{
+	struct wfx_dev *wdev = container_of(work, struct wfx_dev, hif.bh);
+	int stats_req = 0, stats_cnf = 0, stats_ind = 0;
+	bool release_chip = false, last_op_is_rx = false;
+	int num_tx, num_rx;
+
+	device_wakeup(wdev);
+	do {
+		num_tx = bh_work_tx(wdev, 32);
+		stats_req += num_tx;
+		if (num_tx)
+			last_op_is_rx = false;
+		num_rx = bh_work_rx(wdev, 32, &stats_cnf);
+		stats_ind += num_rx;
+		if (num_rx)
+			last_op_is_rx = true;
+	} while (num_rx || num_tx);
+	stats_ind -= stats_cnf;
+
+	if (last_op_is_rx)
+		ack_sdio_data(wdev);
+	if (!wdev->hif.tx_buffers_used && !work_pending(work)) {
+		device_release(wdev);
+		release_chip = true;
+	}
+	_trace_bh_stats(stats_ind, stats_req, stats_cnf, wdev->hif.tx_buffers_used, release_chip);
+}
+
+/* An IRQ from chip did occur */
+void wfx_bh_request_rx(struct wfx_dev *wdev)
+{
+	u32 cur, prev;
+
+	wfx_control_reg_read(wdev, &cur);
+	prev = atomic_xchg(&wdev->hif.ctrl_reg, cur);
+	complete(&wdev->hif.ctrl_ready);
+	queue_work(system_highpri_wq, &wdev->hif.bh);
+
+	if (!(cur & CTRL_NEXT_LEN_MASK))
+		dev_err(wdev->dev, "unexpected control register value: length field is 0: %04x\n",
+			cur);
+	if (prev != 0)
+		dev_err(wdev->dev, "received IRQ but previous data was not (yet) read: %04x/%04x\n",
+			prev, cur);
+}
+
+/* Driver want to send data */
+void wfx_bh_request_tx(struct wfx_dev *wdev)
+{
+	queue_work(system_highpri_wq, &wdev->hif.bh);
+}
+
+/* If IRQ is not available, this function allow to manually poll the control register and simulate
+ * an IRQ ahen an event happened.
+ *
+ * Note that the device has a bug: If an IRQ raise while host read control register, the IRQ is
+ * lost. So, use this function carefully (only duing device initialisation).
+ */
+void wfx_bh_poll_irq(struct wfx_dev *wdev)
+{
+	ktime_t now, start;
+	u32 reg;
+
+	WARN(!wdev->poll_irq, "unexpected IRQ polling can mask IRQ");
+	start = ktime_get();
+	for (;;) {
+		wfx_control_reg_read(wdev, &reg);
+		now = ktime_get();
+		if (reg & 0xFFF)
+			break;
+		if (ktime_after(now, ktime_add_ms(start, 1000))) {
+			dev_err(wdev->dev, "time out while polling control register\n");
+			return;
+		}
+		udelay(200);
+	}
+	wfx_bh_request_rx(wdev);
+}
+
+void wfx_bh_register(struct wfx_dev *wdev)
+{
+	INIT_WORK(&wdev->hif.bh, bh_work);
+	init_completion(&wdev->hif.ctrl_ready);
+	init_waitqueue_head(&wdev->hif.tx_buffers_empty);
+}
+
+void wfx_bh_unregister(struct wfx_dev *wdev)
+{
+	flush_work(&wdev->hif.bh);
+}
diff --git a/drivers/net/wireless/silabs/wfx/bh.h b/drivers/net/wireless/silabs/wfx/bh.h
new file mode 100644
index 000000000000..a44c8b421b7c
--- /dev/null
+++ b/drivers/net/wireless/silabs/wfx/bh.h
@@ -0,0 +1,34 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * Interrupt bottom half (BH).
+ *
+ * Copyright (c) 2017-2020, Silicon Laboratories, Inc.
+ * Copyright (c) 2010, ST-Ericsson
+ */
+#ifndef WFX_BH_H
+#define WFX_BH_H
+
+#include <linux/atomic.h>
+#include <linux/wait.h>
+#include <linux/completion.h>
+#include <linux/workqueue.h>
+
+struct wfx_dev;
+
+struct wfx_hif {
+	struct work_struct bh;
+	struct completion ctrl_ready;
+	wait_queue_head_t tx_buffers_empty;
+	atomic_t ctrl_reg;
+	int rx_seqnum;
+	int tx_seqnum;
+	int tx_buffers_used;
+};
+
+void wfx_bh_register(struct wfx_dev *wdev);
+void wfx_bh_unregister(struct wfx_dev *wdev);
+void wfx_bh_request_rx(struct wfx_dev *wdev);
+void wfx_bh_request_tx(struct wfx_dev *wdev);
+void wfx_bh_poll_irq(struct wfx_dev *wdev);
+
+#endif
-- 
2.34.1

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ