Skip to contentSkip to conversation

p54usb: support LM87 firmwares

PATCH Superseded Christian Lamparter 4 participants 11 messages Sep 2, 2008 – Sep 7, 2008 Networking drivers (wireless)

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 p54usb: support LM87 firmwares

This patch adds the necessary changes to support LM87 firmwares.

---
John,

this one should go into wireless-next.
---

drivers/net/wireless/p54/p54common.c +5 -2
diff -Nurp a/drivers/net/wireless/p54/p54common.c b/drivers/net/wireless/p54/p54common.c
@@ -93,7 +93,10 @@ int p54_parse_firmware(struct ieee80211_
 		u32 code = le32_to_cpu(bootrec->code);
 		switch (code) {
 		case BR_CODE_COMPONENT_ID:
-			switch (be32_to_cpu(*(__be32 *)bootrec->data)) {
+			priv->fw_interface =
+				be32_to_cpu(*(__be32 *)bootrec->data);
+
+			switch (priv->fw_interface) {
 			case FW_FMAC:
 				printk(KERN_INFO "p54: FreeMAC firmware\n");
 				break;
@@ -104,7 +107,7 @@ int p54_parse_firmware(struct ieee80211_
 				printk(KERN_INFO "p54: LM86 firmware\n");
 				break;
 			case FW_LM87:
-				printk(KERN_INFO "p54: LM87 firmware - not supported yet!\n");
+				printk(KERN_INFO "p54: LM87 firmware\n");
 				break;
 			default:
 				printk(KERN_INFO "p54: unknown firmware\n");
drivers/net/wireless/p54/p54common.h +0 -5
diff -Nurp a/drivers/net/wireless/p54/p54common.h b/drivers/net/wireless/p54/p54common.h
@@ -50,11 +50,6 @@ struct bootrec_desc {
 #define BR_CODE_END_OF_BRA		0xFF0000FF
 #define LEGACY_BR_CODE_END_OF_BRA	0xFFFFFFFF
 
-#define FW_FMAC 0x464d4143
-#define FW_LM86 0x4c4d3836
-#define FW_LM87 0x4c4d3837
-#define FW_LM20 0x4c4d3230
-
 /* PDA defines are Copyright (C) 2005 Nokia Corporation (taken from islsm_pda.h) */
 
 struct pda_entry {
drivers/net/wireless/p54/p54.h +6 -0
diff -Nurp a/drivers/net/wireless/p54/p54.h b/drivers/net/wireless/p54/p54.h
@@ -42,6 +42,11 @@ struct p54_control_hdr {
 
 #define ISL38XX_DEV_FIRMWARE_ADDR 0x20000
 
+#define FW_FMAC 0x464d4143
+#define FW_LM86 0x4c4d3836
+#define FW_LM87 0x4c4d3837
+#define FW_LM20 0x4c4d3230
+
 struct p54_common {
 	u32 rx_start;
 	u32 rx_end;
@@ -68,6 +73,7 @@ struct p54_common {
 	unsigned int tx_hdr_len;
 	void *cached_vdcf;
 	unsigned int fw_var;
+	unsigned int fw_interface;
 	struct ieee80211_tx_queue_stats tx_stats[8];
 	void *eeprom;
 	struct completion eeprom_comp;
drivers/net/wireless/p54/p54usb.c +59 -13

103 lines not on this page · raw message

drivers/net/wireless/p54/p54usb.h +5 -0
diff -Nurp a/drivers/net/wireless/p54/p54usb.h b/drivers/net/wireless/p54/p54usb.h
@@ -72,6 +72,11 @@ struct net2280_tx_hdr {
 	u8 padding[8];
 } __attribute__((packed));
 
+struct lm87_tx_hdr {
+	__le32 device_addr;
+	__le32 chksum;
+} __attribute__((packed));
+
 /* Some flags for the isl hardware registers controlling DMA inside the
  * chip */
 #define ISL38XX_DMA_STATUS_DONE			0x00000001
#1 lore raw
Quoted 12 lines · Christian Lamparter

This patch adds the necessary changes to support LM87 firmwares.

---
John,

this one should go into wireless-next.
---
diff -Nurp a/drivers/net/wireless/p54/p54common.c b/drivers/net/wireless/p54/p54common.c

drivers/net/wireless/p54/p54common.c 2008-09-02 21:29:47.000000000 +0200 int p54_parse_firmware(struct ieee80211_ · line 97

Quoted 6 lines
drivers/net/wireless/p54/p54common.c 2008-09-02 21:29:47.000000000 +0200 renamed from drivers/net/wireless/p54/p54common.c 2008-09-01 22:25:35.000000000 +0200 +2 -1
@@ -93,7 +93,10 @@ int p54_parse_firmware(struct ieee80211_
               u32 code = le32_to_cpu(bootrec->code);
               switch (code) {
               case BR_CODE_COMPONENT_ID:
-                       switch (be32_to_cpu(*(__be32 *)bootrec->data)) {
+                       priv->fw_interface =
+                               be32_to_cpu(*(__be32 *)bootrec->data);

what about be32_to_cpup?
Tomas

#2 lore raw
Quoted 13 lines · On Wednesday 03 September 2008 01:14:12 Tomas Winkler wrote:

drivers/net/wireless/p54/p54common.c int p54_parse_firmware(struct ieee80211_ · line 97

Quoted 12 lines

diff -Nurp a/drivers/net/wireless/p54/p54common.c b/drivers/net/wireless/p54/p54common.c ---
a/drivers/net/wireless/p54/p54common.c 2008-09-01 22:25:35.000000000
+0200 +++ b/drivers/net/wireless/p54/p54common.c 2008-09-02
21:29:47.000000000 +0200 @@ -93,7 +93,10 @@ int p54_parse_firmware(struct ieee80211_

               u32 code = le32_to_cpu(bootrec->code);
               switch (code) {
               case BR_CODE_COMPONENT_ID:
drivers/net/wireless/p54/p54common.c +2 -1
-                       switch (be32_to_cpu(*(__be32 *)bootrec->data)) {
+                       priv->fw_interface =
+                               be32_to_cpu(*(__be32 *)bootrec->data);
Quoted 2 lines

what about be32_to_cpup?
Tomas

whats the main difference between be32_to_cpu vs be32_to_cpus vs be32_to_cpup? Does the extra p just stand for something like "pointer"?

hmm,
if yes, >>*<<(__be32 *)bootrec->data.
if no, please tell me more about it ;-).

Regards,
Chr

#3 lore raw
Quoted 2 lines · Christian Lamparter

whats the main difference between be32_to_cpu vs be32_to_cpus vs be32_to_cpup? Does the extra p just stand for something like "pointer"?

Yes. See include/linux/byteorder/generic.h

Quoted 4 lines

hmm,
if yes, >>*<<(__be32 *)bootrec->data.
if no, please tell me more about it ;-).

The pointer versions take the pointer. They do the cast and the dereference for you. The cast in your code should be unnecessary.

The in-situ version don't return anything; they put the result into the original variable.

Signature · 2 lines
Regards,
Pavel Roskin
#4 lore raw

PATCH v2 1/1 p54usb: support LM87 firmwares

This patch adds the necessary changes to support LM87 firmwares.

---
as you can see, the *(__be32 *) cast is now gone ;-)
---
diff -Nurp a/drivers/net/wireless/p54/p54common.c b/drivers/net/wireless/p54/p54common.c

drivers/net/wireless/p54/p54common.c +3 -2
@@ -93,7 +93,8 @@ int p54_parse_firmware(struct ieee80211_
 		u32 code = le32_to_cpu(bootrec->code);
 		switch (code) {
 		case BR_CODE_COMPONENT_ID:
-			switch (be32_to_cpu(*(__be32 *)bootrec->data)) {
+			priv->fw_interface = be32_to_cpup(bootrec->data);
+			switch (priv->fw_interface) {
 			case FW_FMAC:
 				printk(KERN_INFO "p54: FreeMAC firmware\n");
 				break;
@@ -104,7 +105,7 @@ int p54_parse_firmware(struct ieee80211_
 				printk(KERN_INFO "p54: LM86 firmware\n");
 				break;
 			case FW_LM87:
-				printk(KERN_INFO "p54: LM87 firmware - not supported yet!\n");
+				printk(KERN_INFO "p54: LM87 firmware\n");
 				break;
 			default:
 				printk(KERN_INFO "p54: unknown firmware\n");

diff -Nurp a/drivers/net/wireless/p54/p54common.h
b/drivers/net/wireless/p54/p54common.h

drivers/net/wireless/p54/p54common.h +0 -5
@@ -50,11 +50,6 @@ struct bootrec_desc {
 #define BR_CODE_END_OF_BRA		0xFF0000FF
 #define LEGACY_BR_CODE_END_OF_BRA	0xFFFFFFFF
 
-#define FW_FMAC 0x464d4143
-#define FW_LM86 0x4c4d3836
-#define FW_LM87 0x4c4d3837
-#define FW_LM20 0x4c4d3230
-
 /* PDA defines are Copyright (C) 2005 Nokia Corporation (taken from 

islsm_pda.h) */

struct pda_entry {

drivers/net/wireless/p54/p54.h +6 -0
diff -Nurp a/drivers/net/wireless/p54/p54.h b/drivers/net/wireless/p54/p54.h
@@ -42,6 +42,11 @@ struct p54_control_hdr {
 
 #define ISL38XX_DEV_FIRMWARE_ADDR 0x20000
 
+#define FW_FMAC 0x464d4143
+#define FW_LM86 0x4c4d3836
+#define FW_LM87 0x4c4d3837
+#define FW_LM20 0x4c4d3230
+
 struct p54_common {
 	u32 rx_start;
 	u32 rx_end;
@@ -68,6 +73,7 @@ struct p54_common {
 	unsigned int tx_hdr_len;
 	void *cached_vdcf;
 	unsigned int fw_var;
+	unsigned int fw_interface;
 	struct ieee80211_tx_queue_stats tx_stats[8];
 	void *eeprom;
 	struct completion eeprom_comp;

diff -Nurp a/drivers/net/wireless/p54/p54usb.c
b/drivers/net/wireless/p54/p54usb.c

drivers/net/wireless/p54/p54usb.c +49 -5

70 lines not on this page · raw message

*data,
size_t len, int free_on_tx)
{
@@ -776,21 +820,23 @@ static int __devinit p54u_probe(struct u
}
}
priv->common.open = p54u_open;
-
+ priv->common.stop = p54u_stop;
if (recognized_pipes < P54U_PIPE_NUMBER) {
priv->hw_type = P54U_3887;
- priv->common.tx = p54u_tx_3887;
+ err = p54u_upload_firmware_3887(dev);
+ if (priv->common.fw_interface == FW_LM87) {
+ dev->extra_tx_headroom += sizeof(struct lm87_tx_hdr);
+ priv->common.tx_hdr_len = sizeof(struct lm87_tx_hdr);
+ priv->common.tx = p54u_tx_lm87;
+ } else
+ priv->common.tx = p54u_tx_3887;
} else {
+ priv->hw_type = P54U_NET2280;
dev->extra_tx_headroom += sizeof(struct net2280_tx_hdr); priv->common.tx_hdr_len = sizeof(struct net2280_tx_hdr); priv->common.tx = p54u_tx_net2280;
- }
- priv->common.stop = p54u_stop;
-
- if (priv->hw_type)
- err = p54u_upload_firmware_3887(dev);
- else
err = p54u_upload_firmware_net2280(dev);
+ }
if (err)
goto err_free_dev;

diff -Nurp a/drivers/net/wireless/p54/p54usb.h
b/drivers/net/wireless/p54/p54usb.h

drivers/net/wireless/p54/p54usb.h +5 -0
@@ -72,6 +72,11 @@ struct net2280_tx_hdr {
 	u8 padding[8];
 } __attribute__((packed));
 
+struct lm87_tx_hdr {
+	__le32 device_addr;
+	__le32 chksum;
+} __attribute__((packed));
+
 /* Some flags for the isl hardware registers controlling DMA inside the
  * chip */
 #define ISL38XX_DMA_STATUS_DONE			0x00000001
#5 lore raw
Quoted 2 lines · Christian Lamparter

This patch adds the necessary changes to support LM87 firmwares.

Do you have a URL for the LM87 firmware? After this patch was added, I tried the 2.5.11.0.arm and 2.13.1.0.arm files, but only the old 2.5.8.0 works. That is the one with the 'ping -s 1400' problem.

Larry

author
#6 lore raw
Quoted 7 lines · On Sunday 07 September 2008 00:57:35 Larry Finger wrote:
Quoted 1 line · Christian Lamparter

This patch adds the necessary changes to support LM87 firmwares.

Do you have a URL for the LM87 firmware? After this patch was added, I tried the 2.5.11.0.arm and 2.13.1.0.arm files, but only the old 2.5.8.0 works. That is the one with the 'ping -s 1400' problem.

Yes

But, for newer firmwares you'll need "p54: better firmware support" (cc4b0cbf4ab) as well. This changes were merged about 25-26 hours ago into wireless-testing, the only thing that isn't in yet is 802.11a support....

the LM87 firmware is this one:
http://daemonizer.de/prism54/prism54-fw/fw-usb/2.13.1.0.arm.1

This link can be found on the driver's site too:
http://linuxwireless.org/en/users/Drivers/p54

Regards,
Chr

#7 lore raw
Quoted 9 lines · Chr

But, for newer firmwares you'll need "p54: better firmware support" (cc4b0cbf4ab) as well. This changes were merged about 25-26 hours ago into wireless-testing, the only thing that isn't in yet is 802.11a support....

the LM87 firmware is this one:
http://daemonizer.de/prism54/prism54-fw/fw-usb/2.13.1.0.arm.1

This link can be found on the driver's site too:
http://linuxwireless.org/en/users/Drivers/p54

Thanks. That one fixed the ping problem. I had the latest wireless-testing, and have added the patches you posted earlier today.

Thanks,

Larry

author
#8 lore raw
Quoted 6 lines · On Sunday 07 September 2008 01:33:29 Larry Finger wrote:
Quoted 5 lines · Chr

But, for newer firmwares you'll need "p54: better firmware support" (cc4b0cbf4ab) as well. This changes were merged about 25-26 hours ago into wireless-testing, the only thing that isn't in yet is 802.11a support....

Quoted 9 lines
Quoted 5 lines

the LM87 firmware is this one:
http://daemonizer.de/prism54/prism54-fw/fw-usb/2.13.1.0.arm.1

This link can be found on the driver's site too:
http://linuxwireless.org/en/users/Drivers/p54

Thanks. That one fixed the ping problem. I had the latest wireless-testing, and have added the patches you posted earlier today.

Well, since you already do some testing here. Want more? ;-)

I will be off for about 4-5 weeks and it would be very nice if someone with lots of broadcom hardware can test if the p54 works ADHOC & AP mode with them.

i've already (successfully) tested Intel's 2915 & 4965, several atheros chipsets and another prism54 on different operating systems.

But there could be problems with some broadcom chips.

Regards,
Chr
---

drivers/net/wireless/p54/p54common.c +239 -29
diff -Nurp a/drivers/net/wireless/p54/p54common.c b/drivers/net/wireless/p54/p54common.c
@@ -589,10 +589,19 @@ static void p54_rx_frame_sent(struct iee
 				else
 					info->status.excessive_retries = 1;
 			}
+			if (payload->status & 0x04)
+				info->flags |= IEEE80211_TX_STAT_TX_FILTERED;
+			info->status.retry_count = payload->retries - 1;
 			info->status.retry_count = payload->retries - 1;
 			info->status.ack_signal = p54_rssi_to_dbm(dev,
 					le16_to_cpu(payload->ack_rssi));
 			skb_pull(entry, sizeof(*hdr) + pad + sizeof(*entry_data));
+			if (info->flags & IEEE80211_TX_CTL_ASSIGN_SEQ) {
+				struct ieee80211_hdr *ieee80211hdr =
+					(struct ieee80211_hdr *)entry->data;
+				ieee80211hdr->seq_ctrl |=
+					cpu_to_le16(payload->seq);
+			}
 			ieee80211_tx_status_irqsafe(dev, entry);
 			goto out;
 		} else
@@ -803,25 +812,130 @@ free:
 }
 EXPORT_SYMBOL_GPL(p54_read_eeprom);
 
+static int p54_sta_unlock(struct ieee80211_hw *dev, u8 *addr)
+{
+	struct p54_common *priv = dev->priv;
+	struct p54_control_hdr *hdr;
+	struct p54_tx_control_sta_unlock *sta;
+
+	if (is_multicast_ether_addr(addr))
+		return 0;
+
+	hdr = kmalloc(sizeof(*hdr) + sizeof(*sta) +
+		      priv->tx_hdr_len, GFP_ATOMIC);
+	if (!hdr)
+		return -ENOMEM;
+
+	hdr = (void *)hdr + priv->tx_hdr_len;
+	hdr->magic1 = cpu_to_le16(0x8001);
+	hdr->len = cpu_to_le16(sizeof(*sta));
+	hdr->type = cpu_to_le16(P54_CONTROL_TYPE_PSM_STA_UNLOCK);
+	p54_assign_address(dev, NULL, hdr, sizeof(*hdr) + sizeof(*sta));
+
+	sta = (struct p54_tx_control_sta_unlock *) hdr->data;
+	memcpy(sta->addr, addr, ETH_ALEN);
+
+	priv->tx(dev, hdr, sizeof(*hdr) + sizeof(*sta), 1);
+
+	return 0;
+}
+
+static int p54_set_tim(struct ieee80211_hw *dev, int aid, int set)
+{
+	struct p54_common *priv = dev->priv;
+	struct p54_control_hdr *hdr;
+	struct p54_tx_control_tim *tim;
+
+	hdr = kmalloc(sizeof(*hdr) + sizeof(*tim) +
+		      priv->tx_hdr_len, GFP_KERNEL);
+	if (!hdr)
+		return -ENOMEM;
+
+	hdr = (void *)hdr + priv->tx_hdr_len;
+	hdr->magic1 = cpu_to_le16(0x8001);
+	hdr->len = cpu_to_le16(sizeof(*tim));
+	hdr->type = cpu_to_le16(P54_CONTROL_TYPE_TIM);
+	p54_assign_address(dev, NULL, hdr, sizeof(*hdr) + sizeof(*tim));
+
+	tim = (struct p54_tx_control_tim *) hdr->data;
+	tim->count = 1;
+	tim->entry[0] = cpu_to_le16(set ? (aid | 0x8000) : aid);
+
+	priv->tx(dev, hdr, sizeof(*hdr) + sizeof(*tim), 1);
+
+	return 0;
+}
+
+static int p54_tx_fill_header(struct ieee80211_hw *dev,
+	struct ieee80211_hdr *hdr, struct ieee80211_tx_info *info,
+	u8 *queue, u16 *flags, u16 *aid)
+{
+	struct p54_common *priv = dev->priv;
+	u16 fc = le16_to_cpu(hdr->frame_control);
+	int ret = 1;
+
+	if ((fc & IEEE80211_FCTL_FTYPE) == IEEE80211_FTYPE_MGMT)
+		switch (fc & IEEE80211_FCTL_STYPE) {
+		case IEEE80211_STYPE_PROBE_RESP:
+			*aid = 0;
+			*queue = 2;
+			*flags = 0x22;
+			return 0;
+		case IEEE80211_STYPE_BEACON:
+			*aid = 0;
+			*queue = 0;
+			*flags = 0x02;
+			return 0;
+		default:
+			*queue = 2;
+			ret = 0;
+		}
+	else
+		*queue += 4;
+
+	if (is_multicast_ether_addr(hdr->addr1)) {
+		*queue = 3;
+		*aid = 0;
+		return 0;
+	}
+
+	if (priv->mode != IEEE80211_IF_TYPE_STA) {
+		*aid = info->control.aid;
+		if (!info->control.aid)
+			*flags = 0x20;
+	} else
+		*aid = 1;
+
+	return ret;
+}
+
 static int p54_tx(struct ieee80211_hw *dev, struct sk_buff *skb)
 {
 	struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb);
 	struct ieee80211_tx_queue_stats *current_queue;
+	struct ieee80211_hdr *ieee80211hdr = (struct ieee80211_hdr *)skb->data;
 	struct p54_common *priv = dev->priv;
 	struct p54_control_hdr *hdr;
-	struct ieee80211_hdr *ieee80211hdr = (struct ieee80211_hdr *)skb->data;
 	struct p54_tx_control_allocdata *txhdr;
 	size_t padding, len;
-	u8 rate;
-	u8 cts_rate = 0x20;
+	u16 aid, hdr_flags = 0x0000;
+	u8 queue, rate, cts_rate = 0x20;
+
+	queue = skb_get_queue_mapping(skb);
 
-	current_queue = &priv->tx_stats[skb_get_queue_mapping(skb) + 4];
-	if (unlikely(current_queue->len > current_queue->limit))
-		return NETDEV_TX_BUSY;
-	current_queue->len++;
-	current_queue->count++;
-	if (current_queue->len == current_queue->limit)
-		ieee80211_stop_queue(dev, skb_get_queue_mapping(skb));
+	if (likely(p54_tx_fill_header(dev, ieee80211hdr, info,
+				      &queue, &hdr_flags, &aid))) {
+		current_queue = &priv->tx_stats[queue];
+		if (unlikely(current_queue->len > current_queue->limit))
+			return NETDEV_TX_BUSY;
+		current_queue->len++;
+		current_queue->count++;
+		if (current_queue->len == current_queue->limit)
+			ieee80211_stop_queue(dev, skb_get_queue_mapping(skb));
+	}
+
+	if (info->flags & IEEE80211_TX_CTL_CLEAR_PS_FILT)
+		p54_sta_unlock(dev, ieee80211hdr->addr1);
 
 	padding = (unsigned long)(skb->data - (sizeof(*hdr) + sizeof(*txhdr))) & 3;
 	len = skb->len;
@@ -831,13 +945,12 @@ static int p54_tx(struct ieee80211_hw *d
 	hdr = (struct p54_control_hdr *) skb_push(skb, sizeof(*hdr));
 
 	if (padding)
-		hdr->magic1 = cpu_to_le16(0x4010);
-	else
-		hdr->magic1 = cpu_to_le16(0x0010);
+		hdr_flags = 0x4000;
+
+	hdr->magic1 = cpu_to_le16(hdr_flags);
 	hdr->len = cpu_to_le16(len);
-	hdr->type = (info->flags & IEEE80211_TX_CTL_NO_ACK) ? 0 : cpu_to_le16(1);
+	hdr->type = cpu_to_le16(aid);
 	hdr->retry1 = hdr->retry2 = info->control.retry_limit;
-
 	/* TODO: add support for alternate retry TX rates */
 	rate = ieee80211_get_tx_rate(dev, info)->hw_value;
 	if (info->flags & IEEE80211_TX_CTL_SHORT_PREAMBLE) {
@@ -854,7 +967,7 @@ static int p54_tx(struct ieee80211_hw *d
 	memset(txhdr->rateset, rate, 8);
 	txhdr->key_type = 0;
 	txhdr->key_len = 0;
-	txhdr->hw_queue = skb_get_queue_mapping(skb) + 4;
+	txhdr->hw_queue = queue;
 	txhdr->tx_antenna = (info->antenna_sel_tx == 0) ?
 		2 : info->antenna_sel_tx - 1;
 	txhdr->output_power = priv->output_power;
@@ -870,17 +983,12 @@ static int p54_tx(struct ieee80211_hw *d
 	 * patch places the sequence number in the hardware state, which
 	 * limits us to a single virtual state.
 	 */
-	if (info->flags & IEEE80211_TX_CTL_ASSIGN_SEQ) {
-		if (info->flags & IEEE80211_TX_CTL_FIRST_FRAGMENT)
-			priv->seqno += 0x10;
-		ieee80211hdr->seq_ctrl &= cpu_to_le16(IEEE80211_SCTL_FRAG);
-		ieee80211hdr->seq_ctrl |= cpu_to_le16(priv->seqno);
-	}
 	/* modifies skb->cb and with it info, so must be last! */
 	p54_assign_address(dev, skb, hdr, skb->len);
 
 	priv->tx(dev, hdr, skb->len, 0);
 	return 0;
+
 }
 
 static int p54_set_filter(struct ieee80211_hw *dev, u16 filter_type,
@@ -1170,6 +1278,9 @@ static int p54_add_interface(struct ieee
 
 	switch (conf->type) {
 	case IEEE80211_IF_TYPE_STA:
+	case IEEE80211_IF_TYPE_AP:
+	case IEEE80211_IF_TYPE_MNTR:
+	case IEEE80211_IF_TYPE_IBSS:
 		priv->mode = conf->type;
 		break;
 	default:
@@ -1184,6 +1295,15 @@ static int p54_add_interface(struct ieee
 	case IEEE80211_IF_TYPE_STA:
 		p54_set_filter(dev, 1, NULL);
 		break;
+	case IEEE80211_IF_TYPE_IBSS:
+		p54_set_filter(dev, 0x02, NULL);
+		break;
+	case IEEE80211_IF_TYPE_MNTR:
+		p54_set_filter(dev, 0x10, NULL);
+		break;
+	case IEEE80211_IF_TYPE_AP:
+		p54_set_filter(dev, 0x04, NULL);
+		break;
 	default:
 		BUG();	/* impossible */
 		break;
@@ -1211,25 +1331,114 @@ static int p54_config(struct ieee80211_h
 	mutex_lock(&priv->conf_mutex);
 	priv->rx_antenna = (conf->antenna_sel_rx == 0) ?
 		2 : conf->antenna_sel_tx - 1;
-	priv->output_power = conf->power_level << 2;
+	priv->output_power = (conf->power_level << 2) & ~0x80;
 	ret = p54_set_freq(dev, cpu_to_le16(conf->channel->center_freq));
 	p54_set_vdcf(dev);
 	mutex_unlock(&priv->conf_mutex);
 	return ret;
 }
 
+static void p54_beacon_tim(struct sk_buff *skb)
+{
+	/*
+	 * the good excuse for this mess is ... the firmware.
+	 * The dummy TIM MUST be at the end of the beacon frame,
+	 * because it'll be overwritten!
+	 */
+
+	struct ieee80211_mgmt *mgmt = (void *)skb->data;
+	u8 *pos, *end;
+
+	if (skb->len <= sizeof(mgmt))
+		return ;
+
+	pos = (u8 *)mgmt->u.beacon.variable;
+	end = skb->data + skb->len;
+	while (pos < end) {
+		if (pos + 2 + pos[1] > end) {
+			printk(KERN_ERR "p54: parsing beacon failed\n");
+			return;
+		}
+
+		if (pos[0] == WLAN_EID_TIM) {
+			u8 dtim_len = pos[1];
+			u8 *next = pos + 2 + dtim_len;
+
+			memmove(pos, next, end - next);
+
+			if (dtim_len > 3)
+				skb_trim(skb, skb->len - (dtim_len - 3));
+			if (dtim_len < 3)
+				skb_put(skb, skb->len + (3 - dtim_len));
+
+			pos = end - (dtim_len + 2);
+
+			/* add the dummy at the end */
+			pos[0] = WLAN_EID_TIM;
+			pos[1] = 3;
+			pos[2] = 0;
+			pos[3] = 1; /* FIX MAC80211: get the real dtim period
+				       from the bss struct... */
+			pos[4] = 0;
+			return ;
+		}
+		pos += 2 + pos[1];
+	}
+}
+
+static int p54_beacon_update(struct ieee80211_hw *dev,
+			struct ieee80211_vif *vif)
+{
+	struct p54_common *priv = dev->priv;
+	struct sk_buff *beacon = ieee80211_beacon_get(dev, vif);
+
+	if (!beacon)
+		return -ENOMEM;
+
+	p54_beacon_tim(beacon);
+	p54_tx(dev, beacon);
+	priv->tsf_high32 = 0;
+	priv->tsf_low32 = 0;
+
+	return 0;
+}
+
 static int p54_config_interface(struct ieee80211_hw *dev,
 				struct ieee80211_vif *vif,
 				struct ieee80211_if_conf *conf)
 {
 	struct p54_common *priv = dev->priv;
+	int ret = 0;
 
-	mutex_lock(&priv->conf_mutex);
-	p54_set_filter(dev, 0, conf->bssid);
-	p54_set_leds(dev, 1, !is_multicast_ether_addr(conf->bssid), 0);
-	memcpy(priv->bssid, conf->bssid, ETH_ALEN);
-	mutex_unlock(&priv->conf_mutex);
-	return 0;
+	switch (priv->mode) {
+	case IEEE80211_IF_TYPE_STA:
+		mutex_lock(&priv->conf_mutex);
+		memcpy(priv->bssid, conf->bssid, ETH_ALEN);
+		p54_set_filter(dev, 0x1, conf->bssid);
+		p54_set_leds(dev, 1, !is_multicast_ether_addr(conf->bssid), 0);
+		mutex_unlock(&priv->conf_mutex);
+		break;
+	case IEEE80211_IF_TYPE_AP:
+	case IEEE80211_IF_TYPE_IBSS:
+		mutex_lock(&priv->conf_mutex);
+		memcpy(priv->bssid, conf->bssid, ETH_ALEN);
+		p54_set_freq(dev, dev->conf.channel->center_freq);
+		p54_set_filter(dev, priv->filter_type, priv->bssid);
+		mutex_unlock(&priv->conf_mutex);
+
+		if (conf->changed & IEEE80211_IFCC_BEACON) {
+			ret = p54_beacon_update(dev, vif);
+			if (ret)
+				return ret;
+
+			mutex_lock(&priv->conf_mutex);
+			p54_set_vdcf(dev);
+			mutex_unlock(&priv->conf_mutex);
+		}
+		break;
+	}
+
+	return ret;
 }
 
 static void p54_configure_filter(struct ieee80211_hw *dev,
@@ -1366,6 +1575,7 @@ static const struct ieee80211_ops p54_op
 	.config_interface	= p54_config_interface,
 	.configure_filter	= p54_configure_filter,
 	.conf_tx		= p54_conf_tx,
+	.set_tim		= p54_set_tim,
 	.get_stats		= p54_get_stats,
 	.get_tx_stats		= p54_get_tx_stats
 };
drivers/net/wireless/p54/p54common.h +11 -0
diff -Nurp a/drivers/net/wireless/p54/p54common.h b/drivers/net/wireless/p54/p54common.h
@@ -321,4 +321,15 @@ struct p54_tx_control_xbow_synth {
 	u32 padding[5];
 } __attribute__ ((packed));
 
+struct p54_tx_control_sta_unlock {
+	u8 addr[ETH_ALEN];
+	u16 padding;
+} __attribute__ ((packed));
+
+struct p54_tx_control_tim {
+	u8 count;
+	u8 unalloc0[3];
+	__le16 entry[8];
+} __attribute__ ((packed));
+
 #endif /* P54COMMON_H */
#9 lore raw
Quoted 17 lines · Chr
Quoted 6 lines · On Sunday 07 September 2008 01:33:29 Larry Finger wrote:
Quoted 5 lines · Chr

But, for newer firmwares you'll need "p54: better firmware support" (cc4b0cbf4ab) as well. This changes were merged about 25-26 hours ago into wireless-testing, the only thing that isn't in yet is 802.11a support....

Quoted 8 lines
Quoted 5 lines

the LM87 firmware is this one:
http://daemonizer.de/prism54/prism54-fw/fw-usb/2.13.1.0.arm.1

This link can be found on the driver's site too:
http://linuxwireless.org/en/users/Drivers/p54

Thanks. That one fixed the ping problem. I had the latest wireless-testing, and have added the patches you posted earlier today.

Well, since you already do some testing here. Want more? ;-)

Quoted 7 lines

I will be off for about 4-5 weeks and it would be very nice if someone with lots of broadcom hardware can test if the p54 works ADHOC & AP mode with them.

i've already (successfully) tested Intel's 2915 & 4965, several atheros chipsets and another prism54 on different operating systems.

But there could be problems with some broadcom chips.

I'll see what I can do.

After loading the new firmware, p54usb crapped out after the following message was logged:

phy9: not handling 0x02 type control frame

After this, I had to unload and reload the driver. I have not yet had time to check on it.

Larry

#10 lore raw
Quoted 30 lines · On Sunday 07 September 2008 02:32:56 Larry Finger wrote:
Quoted 18 lines · Chr
Quoted 6 lines · On Sunday 07 September 2008 01:33:29 Larry Finger wrote:
Quoted 5 lines · Chr

But, for newer firmwares you'll need "p54: better firmware support" (cc4b0cbf4ab) as well. This changes were merged about 25-26 hours ago into wireless-testing, the only thing that isn't in yet is 802.11a support....

Quoted 8 lines
Quoted 5 lines

the LM87 firmware is this one:
http://daemonizer.de/prism54/prism54-fw/fw-usb/2.13.1.0.arm.1

This link can be found on the driver's site too:
http://linuxwireless.org/en/users/Drivers/p54

Thanks. That one fixed the ping problem. I had the latest wireless-testing, and have added the patches you posted earlier today.

Well, since you already do some testing here. Want more? ;-)

Quoted 8 lines

I will be off for about 4-5 weeks and it would be very nice if someone with lots of broadcom hardware can test if the p54 works ADHOC & AP mode with them.

i've already (successfully) tested Intel's 2915 & 4965, several atheros chipsets and another prism54 on different operating systems.

But there could be problems with some broadcom chips.

I'll see what I can do.

Quoted 8 lines

After loading the new firmware, p54usb crapped out after the following message was logged:

phy9: not handling 0x02 type control frame

After this, I had to unload and reload the driver. I have not yet had time to check on it.

Do you have a big-endian machine by accident?

Anyway control flags with 0x02 is some sort of scanning trap... These packages can be produces easily by setting chan->flags = 2 (or 4,6?) and setting the dwell timers as well in p54_set_freq.

Since we don't do this, is there any chance you can hexdump the unknown/corrupted frames?

and as a last shot: try another firmware, maybe this one works better for you.

Regards,
Chr