linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: "Linux kernel regression tracking (Thorsten Leemhuis)"  <regressions@leemhuis.info>
To: Felix Fietkau <nbd@nbd.name>,
	Mikhail Gavrilov <mikhail.v.gavrilov@gmail.com>
Cc: lorenzo@kernel.org, sujuan.chen@mediatek.com,
	Linux List Kernel Mailing <linux-wireless@vger.kernel.org>,
	Linux List Kernel Mailing <linux-kernel@vger.kernel.org>,
	"David S. Miller" <davem@davemloft.net>,
	Eric Dumazet <edumazet@google.com>,
	Jakub Kicinski <kuba@kernel.org>, Paolo Abeni <pabeni@redhat.com>
Subject: Re: [6.2][regression] after commit cd372b8c99c5a5cf6a464acebb7e4a79af7ec8ae stopping working wifi mt7921e
Date: Mon, 9 Jan 2023 08:32:17 +0100	[thread overview]
Message-ID: <6538f66a-bc3e-6f16-180a-50ef37b02e98@leemhuis.info> (raw)
In-Reply-To: <bb480259-2c0f-deab-ab6e-5101e09eb189@leemhuis.info>

On 04.01.23 15:20, Thorsten Leemhuis wrote:
> Hi, this is your Linux kernel regression tracker. Top-posting for once,
> to make this easily accessible to everyone.
> 
> Felix, Lorenzo, did below fix for the regression

There is another report about an issue with mediatek wifi in 6.2-rc:
https://bugzilla.kernel.org/show_bug.cgi?id=216901
To me this looks like a duplicate of the report that started this thread.

(side note: there was another, earlier report that might be a dupe, too:
https://bugzilla.kernel.org/show_bug.cgi?id=216829 )

> Mikhail reported make
> any progress to get mainlined? It doesn't look like it from here, but I
> suspect I missed something, that's why I'm asking.

No reply. :-((

That lack of feedback is another reason why I'm CCing the network
maintainers now, as the mediatek wifi issues in 6.2-rc (this one) and
6.1 ([1]) are already hitting a nerve here because the fixes are
progressing so slowly. I known, it was holiday season, but seems quite a
few people ran into these regressions already, hence we IMHO should
really try to aim fixing both this week.

[1] see
https://lore.kernel.org/all/ac023262-c6cb-01ad-aeee-2dbf379f4c37@leemhuis.info/

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

#regzbot poke

> --
> 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.
> 
> #regzbot poke
> 
> On 21.12.22 18:17, Felix Fietkau wrote:
>> On 21.12.22 17:46, Mikhail Gavrilov wrote:
>>> On Wed, Dec 21, 2022 at 7:12 PM Felix Fietkau <nbd@nbd.name> wrote:
>>>>
>>>> Thanks! I guess I focused on the wrong part of your kernel log
>>>> initially. After more code review, I found that there is in fact a DMA
>>>> related bug in the commit that your bisection pointed to, which happened
>>>> to uncover and trigger the deadlock fixed by my other patch.
>>>>
>>>> So here's my fix for the DMA issue:
>>>> ---
>>> [cutted]
>>>>                 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);
>>>>
>>>
>>> Sorry for stupid question.
>>>
>>> Do you have a separate branch?
>>> I see that the code is differ between master branch and the patch.
>>>
>>> For example in patch the line:
>>> - if (mt76_dma_add_buf(dev, q, &qbuf, 1, 0, buf, t) < 0) {
>>> replaced by the line:
>>> + if (mt76_dma_add_rx_buf(dev, q, &qbuf, buf) < 0) {
>>>
>>> But in master branch
>>> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/net/wireless/mediatek/mt76/dma.c?id=b6bb9676f2165d518b35ba3bea5f1fcfc0d969bf#n604
>>> after line:
>>> qbuf.skip_unmap = false;
>>> followed the line:
>>> mt76_dma_add_buf(dev, q, &qbuf, 1, 0, buf, t);
>>> without if condition.
>>>
>>> So I'm stuck applying the patch :(
>> Sorry, I worked on a tree that had other pending fixes applied.
>> Please try this:
>>
>>
>> --- a/drivers/net/wireless/mediatek/mt76/dma.c
>> +++ b/drivers/net/wireless/mediatek/mt76/dma.c
>> @@ -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,
>> @@ -212,65 +258,51 @@ 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;
>> +
>> +    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;
>>  
>>          idx = q->head;
>> -        q->head = (q->head + 1) % q->ndesc;
>> +        next = (q->head + 1) % q->ndesc;
>>  
>>          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);
>> -            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[q->head].txwi = DMA_DUMMY_DATA;
>> -                q->entry[q->head].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));
>>          WRITE_ONCE(desc->ctrl, cpu_to_le32(ctrl));
>>  
>> +        q->head = next;
>>          q->queued++;
>>      }
>>  
>> @@ -577,17 +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;
>> @@ -601,7 +625,12 @@ 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_rx_buf(dev, q, &qbuf, buf) < 0) {
>> +            dma_unmap_single(dev->dma_dev, addr, len,
>> +                     DMA_FROM_DEVICE);
>> +            skb_free_frag(buf);
>> +            break;
>> +        }
>>          frames++;
>>      }
>>  
>>
>> --- a/drivers/net/wireless/mediatek/mt76/mt7915/mmio.c
>> +++ b/drivers/net/wireless/mediatek/mt76/mt7915/mmio.c
>> @@ -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++;
>>
>> --- a/drivers/net/wireless/mediatek/mt76/tx.c
>> +++ b/drivers/net/wireless/mediatek/mt76/tx.c
>> @@ -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);
>>

  reply	other threads:[~2023-01-09  7:32 UTC|newest]

Thread overview: 25+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-12-21  1:10 [6.2][regression] after commit cd372b8c99c5a5cf6a464acebb7e4a79af7ec8ae stopping working wifi mt7921e Mikhail Gavrilov
2022-12-21 10:45 ` Felix Fietkau
2022-12-21 11:26   ` Lorenzo Bianconi
2022-12-21 13:10   ` Mikhail Gavrilov
2022-12-21 14:12     ` Felix Fietkau
2022-12-21 16:07       ` Lorenzo Bianconi
2022-12-21 16:46       ` Mikhail Gavrilov
2022-12-21 17:17         ` Felix Fietkau
2022-12-22  6:47           ` Mikhail Gavrilov
2022-12-24  7:55             ` Thorsten Leemhuis
2022-12-26 10:59               ` Thorsten Leemhuis
2023-01-04 14:20           ` Thorsten Leemhuis
2023-01-09  7:32             ` Linux kernel regression tracking (Thorsten Leemhuis) [this message]
2023-01-10  7:16               ` Linux kernel regression tracking (Thorsten Leemhuis)
2023-01-10  8:00                 ` Felix Fietkau
2023-01-10  8:41                   ` Linux kernel regression tracking (Thorsten Leemhuis)
2023-01-13 14:11                     ` Kalle Valo
2023-01-10 21:52                   ` Mikhail Gavrilov
2022-12-22 12:36 ` [6.2][regression] after commit cd372b8c99c5a5cf6a464acebb7e4a79af7ec8ae stopping working wifi mt7921e #forregzbot Thorsten Leemhuis
2023-01-27 11:36   ` Linux kernel regression tracking (#update)
2023-01-17  0:33 [6.2][regression] after commit cd372b8c99c5a5cf6a464acebb7e4a79af7ec8ae stopping working wifi mt7921e Mike Lothian
2023-01-17  5:42 ` Mikhail Gavrilov
2023-01-17  6:42   ` Kalle Valo
2023-01-17 13:06   ` Mike Lothian
2023-01-17 13:13     ` Mikhail Gavrilov

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=6538f66a-bc3e-6f16-180a-50ef37b02e98@leemhuis.info \
    --to=regressions@leemhuis.info \
    --cc=davem@davemloft.net \
    --cc=edumazet@google.com \
    --cc=kuba@kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-wireless@vger.kernel.org \
    --cc=lorenzo@kernel.org \
    --cc=mikhail.v.gavrilov@gmail.com \
    --cc=nbd@nbd.name \
    --cc=pabeni@redhat.com \
    --cc=regressions@lists.linux.dev \
    --cc=sujuan.chen@mediatek.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).