Skip to contentSkip to conversation

pci/switchtec: fix stream_open.cocci warnings (fwd)

PATCH Kirill Smelkov 5 participants 6 messages Jun 19, 2019 – Jun 21, 2019 1 message not in the archive PCI

lore
Share & tools
lore

A message that is not in the archive (sent off-list or not yet received): alpine.DEB.2.20.1906191227430.3726@hadrien

#1 lore raw

Hi Julia, everyone.

Quoted 4 lines · Julia Lawall

Hi,

Can you forward this patch to the people below if you think it is appropriate?

Yes, this patch is appropriate. It was actually part of git.kernel.org/linus/c5bf68fe0c86 . It should be safe, (and desirable as it removes a chance for deadlock) to apply it.

Sebastian, Kurt, Logan, Bjorn, please consider picking it up.

Quoted 2 lines

Could I tell the kbuild people to add you to the CC list for this semantic patch?

Yes, sure. Please feel free to add me to CC list for stream_open.cocci related patches.

Kirill

Quoted 50 lines

thanks,
julia

---------- Forwarded message ----------
Date: Wed, 19 Jun 2019 18:23:18 +0800
From: kbuild test robot <[email protected]>
To: [email protected]
Cc: Julia Lawall <[email protected]>
Subject: [PATCH] pci/switchtec: fix stream_open.cocci warnings

CC: [email protected]
TO: Sebastian Andrzej Siewior <[email protected]>
CC: Kurt Schwemmer <[email protected]>
CC: Logan Gunthorpe <[email protected]>
CC: Bjorn Helgaas <[email protected]>
CC: [email protected]
CC: [email protected]

From: kbuild test robot <[email protected]>

drivers/pci/switch/switchtec.c:395:1-17: ERROR: switchtec_fops: .read() can deadlock .write(); change nonseekable_open -> stream_open to fix.

Generated by: scripts/coccinelle/api/stream_open.cocci

Fixes a3a1e895d4fa ("pci/switchtec: Don't use completion's wait queue")

---

tree:   https://git.kernel.org/pub/scm/linux/kernel/git/rt/linux-rt-devel.git linux-5.0.y-rt-rebase
head:   31cc76d5590f5e60c2f26f029e40bc7d0441d93f
commit: a3a1e895d4fa0508e11ac9107ace883a5b2a4d3b [171/305] pci/switchtec: Don't use completion's wait queue

:::::: branch date: 6 days ago
:::::: commit date: 6 days ago

Please take the patch only if it's a positive warning. Thanks!

 switchtec.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/pci/switch/switchtec.c +1 -1
@@ -392,7 +392,7 @@ static int switchtec_dev_open(struct ino
 		return PTR_ERR(stuser);
 
 	filp->private_data = stuser;
-	nonseekable_open(inode, filp);
+	stream_open(inode, filp);
 
 	dev_dbg(&stdev->dev, "%s: %p\n", __func__, stuser);
 
#2 lore raw
Quoted 19 lines · Kirill Smelkov

Hi Julia, everyone.

Quoted 16 lines · Julia Lawall

Hi,

Can you forward this patch to the people below if you think it is appropriate?

From: kbuild test robot <[email protected]>

drivers/pci/switch/switchtec.c:395:1-17: ERROR: switchtec_fops: .read() can deadlock .write(); change nonseekable_open -> stream_open to fix.

Generated by: scripts/coccinelle/api/stream_open.cocci

Fixes a3a1e895d4fa ("pci/switchtec: Don't use completion's wait queue")

---

tree: https://git.kernel.org/pub/scm/linux/kernel/git/rt/linux-rt-devel.git linux-5.0.y-rt-rebase

This is for the RT tree? The patch in the fixes tag isn't in upstream and I don't understand how this is related to that patch at all. It just looks like the RT tree hasn't picked up the patch which made this change in upstream.

I feel like I've seen the change in this patch a bunch of times already and it appears to be correct in rc5 at least...

Logan

Quoted 21 lines
Quoted 21 lines
head:   31cc76d5590f5e60c2f26f029e40bc7d0441d93f
commit: a3a1e895d4fa0508e11ac9107ace883a5b2a4d3b [171/305] pci/switchtec: Don't use completion's wait queue

:::::: branch date: 6 days ago
:::::: commit date: 6 days ago

Please take the patch only if it's a positive warning. Thanks!

 switchtec.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/pci/switch/switchtec.c +1 -1
@@ -392,7 +392,7 @@ static int switchtec_dev_open(struct ino
 		return PTR_ERR(stuser);
 
 	filp->private_data = stuser;
-	nonseekable_open(inode, filp);
+	stream_open(inode, filp);
 
 	dev_dbg(&stdev->dev, "%s: %p\n", __func__, stuser);
 
#3 lore raw
In reply to Kirill Smelkov
Quoted 5 lines · Kirill Smelkov

Hi Julia, everyone.

Quoted 2 lines · Julia Lawall

Hi,

Quoted 8 lines
Quoted 2 lines

Can you forward this patch to the people below if you think it is appropriate?

Yes, this patch is appropriate. It was actually part of git.kernel.org/linus/c5bf68fe0c86 . It should be safe, (and desirable as it removes a chance for deadlock) to apply it.

Sebastian, Kurt, Logan, Bjorn, please consider picking it up.

I don't get it. This appeared in v5.2-rc1 as c5bf68fe0c86 ("*: convert stream-like files from nonseekable_open -> stream_open"), so it looks like this is already done. What would you like me to do with it?

Quoted 59 lines
Quoted 2 lines

Could I tell the kbuild people to add you to the CC list for this semantic patch?

Yes, sure. Please feel free to add me to CC list for stream_open.cocci related patches.

Kirill

Quoted 50 lines

thanks,
julia

---------- Forwarded message ----------
Date: Wed, 19 Jun 2019 18:23:18 +0800
From: kbuild test robot <[email protected]>
To: [email protected]
Cc: Julia Lawall <[email protected]>
Subject: [PATCH] pci/switchtec: fix stream_open.cocci warnings

CC: [email protected]
TO: Sebastian Andrzej Siewior <[email protected]>
CC: Kurt Schwemmer <[email protected]>
CC: Logan Gunthorpe <[email protected]>
CC: Bjorn Helgaas <[email protected]>
CC: [email protected]
CC: [email protected]

From: kbuild test robot <[email protected]>

drivers/pci/switch/switchtec.c:395:1-17: ERROR: switchtec_fops: .read() can deadlock .write(); change nonseekable_open -> stream_open to fix.

Generated by: scripts/coccinelle/api/stream_open.cocci

Fixes a3a1e895d4fa ("pci/switchtec: Don't use completion's wait queue")

---

tree:   https://git.kernel.org/pub/scm/linux/kernel/git/rt/linux-rt-devel.git linux-5.0.y-rt-rebase
head:   31cc76d5590f5e60c2f26f029e40bc7d0441d93f
commit: a3a1e895d4fa0508e11ac9107ace883a5b2a4d3b [171/305] pci/switchtec: Don't use completion's wait queue

:::::: branch date: 6 days ago
:::::: commit date: 6 days ago

Please take the patch only if it's a positive warning. Thanks!

 switchtec.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/pci/switch/switchtec.c +1 -1
@@ -392,7 +392,7 @@ static int switchtec_dev_open(struct ino
 		return PTR_ERR(stuser);
 
 	filp->private_data = stuser;
-	nonseekable_open(inode, filp);
+	stream_open(inode, filp);
 
 	dev_dbg(&stdev->dev, "%s: %p\n", __func__, stuser);
 
#4 lore raw
Quoted 13 lines · Bjorn Helgaas
Quoted 12 lines · Kirill Smelkov

Hi Julia, everyone.

Quoted 4 lines · Julia Lawall

Hi,

Can you forward this patch to the people below if you think it is appropriate?

Yes, this patch is appropriate. It was actually part of git.kernel.org/linus/c5bf68fe0c86 . It should be safe, (and desirable as it removes a chance for deadlock) to apply it.

Quoted 5 lines
Quoted 1 line

Sebastian, Kurt, Logan, Bjorn, please consider picking it up.

I don't get it. This appeared in v5.2-rc1 as c5bf68fe0c86 ("*: convert stream-like files from nonseekable_open -> stream_open"), so it looks like this is already done. What would you like me to do with it?

Somehow 0-day got a hold of it... If there is nothing to do, just ignore it.

thanks,
julia

Quoted 61 lines
Quoted 60 lines
Quoted 3 lines

Could I tell the kbuild people to add you to the CC list for this semantic patch?

Yes, sure. Please feel free to add me to CC list for stream_open.cocci related patches.

Kirill

Quoted 50 lines

thanks,
julia

---------- Forwarded message ----------
Date: Wed, 19 Jun 2019 18:23:18 +0800
From: kbuild test robot <[email protected]>
To: [email protected]
Cc: Julia Lawall <[email protected]>
Subject: [PATCH] pci/switchtec: fix stream_open.cocci warnings

CC: [email protected]
TO: Sebastian Andrzej Siewior <[email protected]>
CC: Kurt Schwemmer <[email protected]>
CC: Logan Gunthorpe <[email protected]>
CC: Bjorn Helgaas <[email protected]>
CC: [email protected]
CC: [email protected]

From: kbuild test robot <[email protected]>

drivers/pci/switch/switchtec.c:395:1-17: ERROR: switchtec_fops: .read() can deadlock .write(); change nonseekable_open -> stream_open to fix.

Generated by: scripts/coccinelle/api/stream_open.cocci

Fixes a3a1e895d4fa ("pci/switchtec: Don't use completion's wait queue")

---

tree:   https://git.kernel.org/pub/scm/linux/kernel/git/rt/linux-rt-devel.git linux-5.0.y-rt-rebase
head:   31cc76d5590f5e60c2f26f029e40bc7d0441d93f
commit: a3a1e895d4fa0508e11ac9107ace883a5b2a4d3b [171/305] pci/switchtec: Don't use completion's wait queue

:::::: branch date: 6 days ago
:::::: commit date: 6 days ago

Please take the patch only if it's a positive warning. Thanks!

 switchtec.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/pci/switch/switchtec.c +1 -1
@@ -392,7 +392,7 @@ static int switchtec_dev_open(struct ino
 		return PTR_ERR(stuser);
 
 	filp->private_data = stuser;
-	nonseekable_open(inode, filp);
+	stream_open(inode, filp);
 
 	dev_dbg(&stdev->dev, "%s: %p\n", __func__, stuser);
 
#5 lore raw
In reply to Bjorn Helgaas

Bjorn Helgaas писал 2019-06-19 23:19:

Quoted 14 lines · Kirill Smelkov
Quoted 13 lines

Hi Julia, everyone.

Quoted 4 lines · Julia Lawall

Hi,

Can you forward this patch to the people below if you think it is appropriate?

Yes, this patch is appropriate. It was actually part of git.kernel.org/linus/c5bf68fe0c86 . It should be safe, (and desirable as
it removes a chance for deadlock) to apply it.

Quoted 6 lines
Quoted 1 line

Sebastian, Kurt, Logan, Bjorn, please consider picking it up.

I don't get it. This appeared in v5.2-rc1 as c5bf68fe0c86 ("*: convert stream-like files from nonseekable_open -> stream_open"), so it looks like
this is already done. What would you like me to do with it?

I meant just that it was ok to pick this change into 5.0-RT tree as kbuild robot was suggesting. Sorry for not being clear.

Kirill

Quoted 59 lines
Quoted 59 lines
Quoted 2 lines

Could I tell the kbuild people to add you to the CC list for this semantic patch?

Yes, sure. Please feel free to add me to CC list for stream_open.cocci related patches.

Kirill

Quoted 50 lines

thanks,
julia

---------- Forwarded message ----------
Date: Wed, 19 Jun 2019 18:23:18 +0800
From: kbuild test robot <[email protected]>
To: [email protected]
Cc: Julia Lawall <[email protected]>
Subject: [PATCH] pci/switchtec: fix stream_open.cocci warnings

CC: [email protected]
TO: Sebastian Andrzej Siewior <[email protected]>
CC: Kurt Schwemmer <[email protected]>
CC: Logan Gunthorpe <[email protected]>
CC: Bjorn Helgaas <[email protected]>
CC: [email protected]
CC: [email protected]

From: kbuild test robot <[email protected]>

drivers/pci/switch/switchtec.c:395:1-17: ERROR: switchtec_fops: .read() can deadlock .write(); change nonseekable_open -> stream_open to fix.

Generated by: scripts/coccinelle/api/stream_open.cocci

Fixes a3a1e895d4fa ("pci/switchtec: Don't use completion's wait queue")

---

tree:   https://git.kernel.org/pub/scm/linux/kernel/git/rt/linux-rt-devel.git linux-5.0.y-rt-rebase
head:   31cc76d5590f5e60c2f26f029e40bc7d0441d93f
commit: a3a1e895d4fa0508e11ac9107ace883a5b2a4d3b [171/305] pci/switchtec: Don't use completion's wait queue

:::::: branch date: 6 days ago
:::::: commit date: 6 days ago

Please take the patch only if it's a positive warning. Thanks!

 switchtec.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/pci/switch/switchtec.c +1 -1
@@ -392,7 +392,7 @@ static int switchtec_dev_open(struct ino
 		return PTR_ERR(stuser);
 
 	filp->private_data = stuser;
-	nonseekable_open(inode, filp);
+	stream_open(inode, filp);
 
 	dev_dbg(&stdev->dev, "%s: %p\n", __func__, stuser);
 
#6 lore raw
Quoted 2 lines · On 2019-06-20 07:01:43 [+0000], [email protected] wrote:

I meant just that it was ok to pick this change into 5.0-RT tree as kbuild robot was suggesting. Sorry for not being clear.

Okay, I was under the impression that this affect non-RT as well. But if this is RT-only then let me pick it up…

Quoted 1 line

Kirill

Sebastian