Skip to contentSkip to conversation

wifi: mt76: dma: do not increment queue head if mt76_dma_add_buf fails

PATCH In review Felix Fietkau 3 participants 7 messages Jan 13, 2023 – Jan 16, 2023

Subsystems (3)
lore
Share & tools
lore

Felix Fietkau

author maintainer
#0 lore raw

PATCH 1/3 wifi: mt76: dma: do not increment queue head if mt76_dma_add_buf fails

From: Lorenzo Bianconi <[email protected]>

Do not increment queue head if mt76_dma_add_buf fails for Wireless Ethernet Dispatcher rx queues.

Fixes cd372b8c99c5 ("wifi: mt76: add WED RX support to mt76_dma_{add,get}_buf")
(cherry picked from commit fe13dad8992be0b26c1be390bcd111acf9892c17)
---
 drivers/net/wireless/mediatek/mt76/dma.c | 9 +++++----
 1 file changed, 5 insertions(+), 4 deletions(-)
drivers/net/wireless/mediatek/mt76/dma.c +5 -4
index f795548562f5..c058a02e9ba5 100644
@@ -212,14 +212,14 @@ mt76_dma_add_buf(struct mt76_dev *dev, struct mt76_queue *q,
 {
 	struct mt76_queue_entry *entry;
 	struct mt76_desc *desc;
-	u32 ctrl;
 	int i, idx = -1;
+	u32 ctrl, next;
 
 	for (i = 0; i < nbufs; i += 2, buf += 2) {
 		u32 buf0 = buf[0].addr, buf1 = 0;
 
 		idx = q->head;
-		q->head = (q->head + 1) % q->ndesc;
+		next = (q->head + 1) % q->ndesc;
 
 		desc = &q->desc[idx];
 		entry = &q->entry[idx];
@@ -239,8 +239,8 @@ mt76_dma_add_buf(struct mt76_dev *dev, struct mt76_queue *q,
 			       MT_DMA_CTL_TO_HOST;
 		} else {
 			if (txwi) {
-				q->entry[q->head].txwi = DMA_DUMMY_DATA;
-				q->entry[q->head].skip_buf0 = true;
+				q->entry[next].txwi = DMA_DUMMY_DATA;
+				q->entry[next].skip_buf0 = true;
 			}
 
 			if (buf[0].skip_unmap)
@@ -271,6 +271,7 @@ mt76_dma_add_buf(struct mt76_dev *dev, struct mt76_queue *q,
 		WRITE_ONCE(desc->info, cpu_to_le32(info));
 		WRITE_ONCE(desc->ctrl, cpu_to_le32(ctrl));
 
+		q->head = next;
 		q->queued++;
 	}
 
Signature · 1 line
2.39.0
author maintainer
#1 lore raw

PATCH 2/3 wifi: mt76: handle possible mt76_rx_token_consume failures

From: Lorenzo Bianconi <[email protected]>

Take into account possible error conditions of mt76_rx_token_consume routine in mt7915_mmio_wed_init_rx_buf() and mt76_dma_add_buf()

Fixes cd372b8c99c5 ("wifi: mt76: add WED RX support to mt76_dma_{add,get}_buf")
Fixes 4f831d18d12d ("wifi: mt76: mt7915: enable WED RX support")
(cherry picked from commit 96f134dc19645be4994e89a2f68fa89309becbee)
---
 drivers/net/wireless/mediatek/mt76/dma.c         | 10 +++++++++-
 drivers/net/wireless/mediatek/mt76/mt7915/mmio.c |  7 +++++++
 drivers/net/wireless/mediatek/mt76/tx.c          |  7 ++++---
 3 files changed, 20 insertions(+), 4 deletions(-)
drivers/net/wireless/mediatek/mt76/dma.c +9 -1
index c058a02e9ba5..61a8ab9e1db5 100644
@@ -234,6 +234,9 @@ mt76_dma_add_buf(struct mt76_dev *dev, struct mt76_queue *q,
 
 			rx_token = mt76_rx_token_consume(dev, (void *)skb, t,
 							 buf[0].addr);
+			if (rx_token < 0)
+				return -ENOMEM;
+
 			buf1 |= FIELD_PREP(MT_DMA_CTL_TOKEN, rx_token);
 			ctrl = FIELD_PREP(MT_DMA_CTL_SD_LEN0, buf[0].len) |
 			       MT_DMA_CTL_TO_HOST;
@@ -602,7 +605,12 @@ mt76_dma_rx_fill(struct mt76_dev *dev, struct mt76_queue *q)
 		qbuf.addr = addr + offset;
 		qbuf.len = len - offset;
 		qbuf.skip_unmap = false;
-		mt76_dma_add_buf(dev, q, &qbuf, 1, 0, buf, t);
+		if (mt76_dma_add_buf(dev, q, &qbuf, 1, 0, buf, t) < 0) {
+			dma_unmap_single(dev->dma_dev, addr, len,
+					 DMA_FROM_DEVICE);
+			skb_free_frag(buf);
+			break;
+		}
 		frames++;
 	}
 
drivers/net/wireless/mediatek/mt76/mt7915/mmio.c +7 -0
index 0a95c3da241b..8388e2a65853 100644
@@ -653,6 +653,13 @@ static u32 mt7915_mmio_wed_init_rx_buf(struct mtk_wed_device *wed, int size)
 
 		desc->buf0 = cpu_to_le32(phy_addr);
 		token = mt76_rx_token_consume(&dev->mt76, ptr, t, phy_addr);
+		if (token < 0) {
+			dma_unmap_single(dev->mt76.dma_dev, phy_addr,
+					 wed->wlan.rx_size, DMA_TO_DEVICE);
+			skb_free_frag(ptr);
+			goto unmap;
+		}
+
 		desc->token |= cpu_to_le32(FIELD_PREP(MT_DMA_CTL_TOKEN,
 						      token));
 		desc++;
drivers/net/wireless/mediatek/mt76/tx.c +4 -3
index 24568b98ed9d..1f309d05380a 100644
@@ -764,11 +764,12 @@ int mt76_rx_token_consume(struct mt76_dev *dev, void *ptr,
 	spin_lock_bh(&dev->rx_token_lock);
 	token = idr_alloc(&dev->rx_token, t, 0, dev->rx_token_size,
 			  GFP_ATOMIC);
+	if (token >= 0) {
+		t->ptr = ptr;
+		t->dma_addr = phys;
+	}
 	spin_unlock_bh(&dev->rx_token_lock);
 
-	t->ptr = ptr;
-	t->dma_addr = phys;
-
 	return token;
 }
 EXPORT_SYMBOL_GPL(mt76_rx_token_consume);
Signature · 1 line
2.39.0
author maintainer
#2 lore raw

PATCH 3/3 wifi: mt76: dma: fix a regression in adding rx buffers

In reply to Felix Fietkau

When adding WED support, mt76_dma_add_buf was accidentally changed to set the skip_buf0 flag for tx buffers on the wrong queue descriptor entry. Additionally, there is a rxwi leak when rx buffer allocation fails.

Fix this and make the code more readable by adding a separate function for adding rx buffers.

Fixes cd372b8c99c5 ("wifi: mt76: add WED RX support to mt76_dma_{add,get}_buf")
---
 drivers/net/wireless/mediatek/mt76/dma.c | 124 +++++++++++++----------
 1 file changed, 72 insertions(+), 52 deletions(-)
drivers/net/wireless/mediatek/mt76/dma.c +72 -52
index 61a8ab9e1db5..06161815c180 100644
@@ -205,6 +205,52 @@ mt76_dma_queue_reset(struct mt76_dev *dev, struct mt76_queue *q)
 	mt76_dma_sync_idx(dev, q);
 }
 
+static int
+mt76_dma_add_rx_buf(struct mt76_dev *dev, struct mt76_queue *q,
+		    struct mt76_queue_buf *buf, void *data)
+{
+	struct mt76_desc *desc = &q->desc[q->head];
+	struct mt76_queue_entry *entry = &q->entry[q->head];
+	struct mt76_txwi_cache *txwi = NULL;
+	u32 buf1 = 0, ctrl;
+	int idx = q->head;
+	int rx_token;
+
+	ctrl = FIELD_PREP(MT_DMA_CTL_SD_LEN0, buf[0].len);
+
+	if ((q->flags & MT_QFLAG_WED) &&
+	    FIELD_GET(MT_QFLAG_WED_TYPE, q->flags) == MT76_WED_Q_RX) {
+		txwi = mt76_get_rxwi(dev);
+		if (!txwi)
+			return -ENOMEM;
+
+		rx_token = mt76_rx_token_consume(dev, data, txwi, buf->addr);
+		if (rx_token < 0) {
+			mt76_put_rxwi(dev, txwi);
+			return -ENOMEM;
+		}
+
+		buf1 |= FIELD_PREP(MT_DMA_CTL_TOKEN, rx_token);
+		ctrl |= MT_DMA_CTL_TO_HOST;
+	}
+
+	WRITE_ONCE(desc->buf0, cpu_to_le32(buf->addr));
+	WRITE_ONCE(desc->buf1, cpu_to_le32(buf1));
+	WRITE_ONCE(desc->ctrl, cpu_to_le32(ctrl));
+	WRITE_ONCE(desc->info, 0);
+
+	entry->dma_addr[0] = buf->addr;
+	entry->dma_len[0] = buf->len;
+	entry->txwi = txwi;
+	entry->buf = data;
+	entry->wcid = 0xffff;
+	entry->skip_buf1 = true;
+	q->head = (q->head + 1) % q->ndesc;
+	q->queued++;
+
+	return idx;
+}
+
 static int
 mt76_dma_add_buf(struct mt76_dev *dev, struct mt76_queue *q,
 		 struct mt76_queue_buf *buf, int nbufs, u32 info,
@@ -215,6 +261,11 @@ mt76_dma_add_buf(struct mt76_dev *dev, struct mt76_queue *q,
 	int i, idx = -1;
 	u32 ctrl, next;
 
+	if (txwi) {
+		q->entry[q->head].txwi = DMA_DUMMY_DATA;
+		q->entry[q->head].skip_buf0 = true;
+	}
+
 	for (i = 0; i < nbufs; i += 2, buf += 2) {
 		u32 buf0 = buf[0].addr, buf1 = 0;
 
@@ -224,51 +275,28 @@ mt76_dma_add_buf(struct mt76_dev *dev, struct mt76_queue *q,
 		desc = &q->desc[idx];
 		entry = &q->entry[idx];
 
-		if ((q->flags & MT_QFLAG_WED) &&
-		    FIELD_GET(MT_QFLAG_WED_TYPE, q->flags) == MT76_WED_Q_RX) {
-			struct mt76_txwi_cache *t = txwi;
-			int rx_token;
-
-			if (!t)
-				return -ENOMEM;
-
-			rx_token = mt76_rx_token_consume(dev, (void *)skb, t,
-							 buf[0].addr);
-			if (rx_token < 0)
-				return -ENOMEM;
-
-			buf1 |= FIELD_PREP(MT_DMA_CTL_TOKEN, rx_token);
-			ctrl = FIELD_PREP(MT_DMA_CTL_SD_LEN0, buf[0].len) |
-			       MT_DMA_CTL_TO_HOST;
-		} else {
-			if (txwi) {
-				q->entry[next].txwi = DMA_DUMMY_DATA;
-				q->entry[next].skip_buf0 = true;
-			}
-
-			if (buf[0].skip_unmap)
-				entry->skip_buf0 = true;
-			entry->skip_buf1 = i == nbufs - 1;
-
-			entry->dma_addr[0] = buf[0].addr;
-			entry->dma_len[0] = buf[0].len;
-
-			ctrl = FIELD_PREP(MT_DMA_CTL_SD_LEN0, buf[0].len);
-			if (i < nbufs - 1) {
-				entry->dma_addr[1] = buf[1].addr;
-				entry->dma_len[1] = buf[1].len;
-				buf1 = buf[1].addr;
-				ctrl |= FIELD_PREP(MT_DMA_CTL_SD_LEN1, buf[1].len);
-				if (buf[1].skip_unmap)
-					entry->skip_buf1 = true;
-			}
-
-			if (i == nbufs - 1)
-				ctrl |= MT_DMA_CTL_LAST_SEC0;
-			else if (i == nbufs - 2)
-				ctrl |= MT_DMA_CTL_LAST_SEC1;
+		if (buf[0].skip_unmap)
+			entry->skip_buf0 = true;
+		entry->skip_buf1 = i == nbufs - 1;
+
+		entry->dma_addr[0] = buf[0].addr;
+		entry->dma_len[0] = buf[0].len;
+
+		ctrl = FIELD_PREP(MT_DMA_CTL_SD_LEN0, buf[0].len);
+		if (i < nbufs - 1) {
+			entry->dma_addr[1] = buf[1].addr;
+			entry->dma_len[1] = buf[1].len;
+			buf1 = buf[1].addr;
+			ctrl |= FIELD_PREP(MT_DMA_CTL_SD_LEN1, buf[1].len);
+			if (buf[1].skip_unmap)
+				entry->skip_buf1 = true;
 		}
 
+		if (i == nbufs - 1)
+			ctrl |= MT_DMA_CTL_LAST_SEC0;
+		else if (i == nbufs - 2)
+			ctrl |= MT_DMA_CTL_LAST_SEC1;
+
 		WRITE_ONCE(desc->buf0, cpu_to_le32(buf0));
 		WRITE_ONCE(desc->buf1, cpu_to_le32(buf1));
 		WRITE_ONCE(desc->info, cpu_to_le32(info));
@@ -581,17 +609,9 @@ mt76_dma_rx_fill(struct mt76_dev *dev, struct mt76_queue *q)
 	spin_lock_bh(&q->lock);
 
 	while (q->queued < q->ndesc - 1) {
-		struct mt76_txwi_cache *t = NULL;
 		struct mt76_queue_buf qbuf;
 		void *buf = NULL;
 
-		if ((q->flags & MT_QFLAG_WED) &&
-		    FIELD_GET(MT_QFLAG_WED_TYPE, q->flags) == MT76_WED_Q_RX) {
-			t = mt76_get_rxwi(dev);
-			if (!t)
-				break;
-		}
-
 		buf = page_frag_alloc(rx_page, q->buf_size, GFP_ATOMIC);
 		if (!buf)
 			break;
@@ -605,7 +625,7 @@ mt76_dma_rx_fill(struct mt76_dev *dev, struct mt76_queue *q)
 		qbuf.addr = addr + offset;
 		qbuf.len = len - offset;
 		qbuf.skip_unmap = false;
-		if (mt76_dma_add_buf(dev, q, &qbuf, 1, 0, buf, t) < 0) {
+		if (mt76_dma_add_rx_buf(dev, q, &qbuf, buf) < 0) {
 			dma_unmap_single(dev->dma_dev, addr, len,
 					 DMA_FROM_DEVICE);
 			skb_free_frag(buf);
Signature · 1 line
2.39.0
#3 lore raw

Re: [PATCH wireless 3/3] wifi: mt76: dma: fix a regression in adding rx buffers

Quoted 3 lines · Felix Fietkau

When adding WED support, mt76_dma_add_buf was accidentally changed to set the skip_buf0 flag for tx buffers on the wrong queue descriptor entry. Additionally, there is a rxwi leak when rx buffer allocation fails.

thx for working on this

Quoted 5 lines

Fix this and make the code more readable by adding a separate function for adding rx buffers.

Many thx for taking care of this. There is one small thing to improve, please add the following tags here to make things easier for future code archaeologists and give proper credit:

Link:
https://lore.kernel.org/r/CABXGCsMEnQd=gYKTd1knRsWuxCb=Etv5nAre%[email protected]/

Link:
https://lore.kernel.org/lkml/[email protected]/

To explain: Linus[1] and others considered proper link tags important in cases like this, as they allow anyone to look into the backstory of a fix weeks or years later. That's nothing new, the documentation[2] for some time says to place such tags in cases like this. I care personally (and made it a bit more explicit in the docs a while ago), because these tags make my regression tracking efforts a whole lot easier, as they allow my tracking bot 'regzbot' to automatically connect reports with patches posted or committed to fix tracked regressions.

Apropos regzbot, let me tell regzbot to monitor this thread:

#regzbot ^backmonitor:
https://lore.kernel.org/r/CABXGCsMEnQd=gYKTd1knRsWuxCb=Etv5nAre%[email protected]/

Quoted 1 line

[...]

Ciao, Thorsten (wearing his 'the Linux kernel's regression tracker' hat)

[1] for details, see:
https://lore.kernel.org/all/CAHk-=wjMmSZzMJ3Xnskdg4+GGz=5p5p+GSYyFBTh0f-DgvdBWg@mail.gmail.com/
https://lore.kernel.org/all/CAHk-=wgs38ZrfPvy=nOwVkVzjpM3VFU1zobP37Fwd_h9iAD5JQ@mail.gmail.com/
https://lore.kernel.org/all/CAHk-=wjxzafG-=J8oT30s7upn4RhBs6TX-uVFZ5rME+L5_DoJA@mail.gmail.com/

[2] see Documentation/process/submitting-patches.rst
(http://docs.kernel.org/process/submitting-patches.html) and Documentation/process/5.Posting.rst
(https://docs.kernel.org/process/5.Posting.html)

--
Everything you wanna know about Linux kernel regression tracking:
https://linux-regtracking.leemhuis.info/about/#tldr
If I did something stupid, please tell me, as explained on that page.

#4 lore raw

Re: [PATCH wireless 3/3] wifi: mt76: dma: fix a regression in adding rx buffers

"Linux kernel regression tracking (Thorsten Leemhuis)"

Quoted 10 lines · <[email protected]> writes:
Quoted 3 lines · Felix Fietkau

When adding WED support, mt76_dma_add_buf was accidentally changed to set the skip_buf0 flag for tx buffers on the wrong queue descriptor entry. Additionally, there is a rxwi leak when rx buffer allocation fails.

thx for working on this

Quoted 3 lines

Fix this and make the code more readable by adding a separate function for adding rx buffers.

Quoted 6 lines
Quoted 2 lines

Many thx for taking care of this. There is one small thing to improve, please add the following tags here to make things easier for future code archaeologists and give proper credit:

I can add those.

Quoted 5 lines

Link:
https://lore.kernel.org/r/CABXGCsMEnQd=gYKTd1knRsWuxCb=Etv5nAre%[email protected]/

But the email is missing for AngeloGioacchino.

Signature · 3 lines
https://patchwork.kernel.org/project/linux-wireless/list/

https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches
#5 lore raw

Re: [PATCH wireless 3/3] wifi: mt76: dma: fix a regression in adding rx buffers

Quoted 13 lines · Kalle Valo

"Linux kernel regression tracking (Thorsten Leemhuis)"

Quoted 10 lines · <[email protected]> writes:
Quoted 3 lines · Felix Fietkau

When adding WED support, mt76_dma_add_buf was accidentally changed to set the skip_buf0 flag for tx buffers on the wrong queue descriptor entry. Additionally, there is a rxwi leak when rx buffer allocation fails.

thx for working on this

Quoted 3 lines

Fix this and make the code more readable by adding a separate function for adding rx buffers.

Quoted 8 lines
Quoted 6 lines
Quoted 2 lines

Many thx for taking care of this. There is one small thing to improve, please add the following tags here to make things easier for future code archaeologists and give proper credit:

I can add those.

Great, many thx!

Quoted 7 lines
Quoted 5 lines

Link:
https://lore.kernel.org/r/CABXGCsMEnQd=gYKTd1knRsWuxCb=Etv5nAre%[email protected]/

But the email is missing for AngeloGioacchino.

/me wonders how that happened

Sorry for that. Here it is again:

Reported-by: AngeloGioacchino Del Regno
<[email protected]>
Ciao, Thorsten

#6 lore raw
In reply to Felix Fietkau
Quoted 9 lines · Felix Fietkau

From: Lorenzo Bianconi <[email protected]>

Do not increment queue head if mt76_dma_add_buf fails for Wireless Ethernet Dispatcher rx queues.

Fixes cd372b8c99c5 ("wifi: mt76: add WED RX support to mt76_dma_{add,get}_buf")
(cherry picked from commit fe13dad8992be0b26c1be390bcd111acf9892c17)

3 patches applied to wireless.git, thanks.

1132d1c834d6 wifi: mt76: dma: do not increment queue head if mt76_dma_add_buf fails e5c3ac895750 wifi: mt76: handle possible mt76_rx_token_consume failures 953519b35227 wifi: mt76: dma: fix a regression in adding rx buffers

Signature · 3 lines
https://patchwork.kernel.org/project/linux-wireless/patch/[email protected]/

https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches