* [PATCH] PCI: access.c: Piggyback user config access on pci_read/write_*()
@ 2018-06-04 15:48 Alexandru Gagniuc
2018-06-04 16:12 ` Keith Busch
0 siblings, 1 reply; 4+ messages in thread
From: Alexandru Gagniuc @ 2018-06-04 15:48 UTC (permalink / raw)
To: bhelgaas
Cc: alex_gagniuc, austin_bolen, shyam_iyer, keith.busch,
Alexandru Gagniuc, linux-pci, linux-kernel
pci_read/write_config*() functions have several safeguards to prevent
stallling accesses when a device is removed. However, their "_user_"
counterparts use a different code path.
To make sure that safeguards are used for userspace PCI config
accesses, piggyback the "_user_" functions on the in-kernel
pci_read/write_config*().
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
---
drivers/pci/access.c | 16 ++--------------
1 file changed, 2 insertions(+), 14 deletions(-)
diff --git a/drivers/pci/access.c b/drivers/pci/access.c
index a3ad2fe185b9..6db2a8713c85 100644
--- a/drivers/pci/access.c
+++ b/drivers/pci/access.c
@@ -223,16 +223,9 @@ int pci_user_read_config_##size \
(struct pci_dev *dev, int pos, type *val) \
{ \
int ret = PCIBIOS_SUCCESSFUL; \
- u32 data = -1; \
if (PCI_##size##_BAD) \
return -EINVAL; \
- raw_spin_lock_irq(&pci_lock); \
- if (unlikely(dev->block_cfg_access)) \
- pci_wait_cfg(dev); \
- ret = dev->bus->ops->read(dev->bus, dev->devfn, \
- pos, sizeof(type), &data); \
- raw_spin_unlock_irq(&pci_lock); \
- *val = (type)data; \
+ ret = pci_read_config_##size(dev, pos, val); \
return pcibios_err_to_errno(ret); \
} \
EXPORT_SYMBOL_GPL(pci_user_read_config_##size);
@@ -245,12 +238,7 @@ int pci_user_write_config_##size \
int ret = PCIBIOS_SUCCESSFUL; \
if (PCI_##size##_BAD) \
return -EINVAL; \
- raw_spin_lock_irq(&pci_lock); \
- if (unlikely(dev->block_cfg_access)) \
- pci_wait_cfg(dev); \
- ret = dev->bus->ops->write(dev->bus, dev->devfn, \
- pos, sizeof(type), val); \
- raw_spin_unlock_irq(&pci_lock); \
+ ret = pci_write_config_##size(dev, pos, val); \
return pcibios_err_to_errno(ret); \
} \
EXPORT_SYMBOL_GPL(pci_user_write_config_##size);
--
2.14.4
^ permalink raw reply related [flat|nested] 4+ messages in thread
* Re: [PATCH] PCI: access.c: Piggyback user config access on pci_read/write_*()
2018-06-04 15:48 [PATCH] PCI: access.c: Piggyback user config access on pci_read/write_*() Alexandru Gagniuc
@ 2018-06-04 16:12 ` Keith Busch
2018-06-04 16:14 ` Alex_Gagniuc
0 siblings, 1 reply; 4+ messages in thread
From: Keith Busch @ 2018-06-04 16:12 UTC (permalink / raw)
To: Alexandru Gagniuc
Cc: bhelgaas, alex_gagniuc, austin_bolen, shyam_iyer, linux-pci,
linux-kernel
On Mon, Jun 04, 2018 at 10:48:02AM -0500, Alexandru Gagniuc wrote:
> +++ b/drivers/pci/access.c
> @@ -223,16 +223,9 @@ int pci_user_read_config_##size \
> (struct pci_dev *dev, int pos, type *val) \
> { \
> int ret = PCIBIOS_SUCCESSFUL; \
> - u32 data = -1; \
> if (PCI_##size##_BAD) \
> return -EINVAL; \
> - raw_spin_lock_irq(&pci_lock); \
> - if (unlikely(dev->block_cfg_access)) \
> - pci_wait_cfg(dev); \
> - ret = dev->bus->ops->read(dev->bus, dev->devfn, \
> - pos, sizeof(type), &data); \
> - raw_spin_unlock_irq(&pci_lock); \
> - *val = (type)data; \
> + ret = pci_read_config_##size(dev, pos, val); \
> return pcibios_err_to_errno(ret); \
> } \
If it wasn't for the block_cfg_access check for user access, this would
have been a nice code reuse cleanup.
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH] PCI: access.c: Piggyback user config access on pci_read/write_*()
2018-06-04 16:12 ` Keith Busch
@ 2018-06-04 16:14 ` Alex_Gagniuc
0 siblings, 0 replies; 4+ messages in thread
From: Alex_Gagniuc @ 2018-06-04 16:14 UTC (permalink / raw)
To: keith.busch, mr.nuke.me
Cc: bhelgaas, Austin.Bolen, Shyam.Iyer, linux-pci, linux-kernel
On 6/4/2018 11:09 AM, Keith Busch wrote:
> On Mon, Jun 04, 2018 at 10:48:02AM -0500, Alexandru Gagniuc wrote:
>> +++ b/drivers/pci/access.c
>> @@ -223,16 +223,9 @@ int pci_user_read_config_##size \
>> (struct pci_dev *dev, int pos, type *val) \
>> { \
>> int ret = PCIBIOS_SUCCESSFUL; \
>> - u32 data = -1; \
>> if (PCI_##size##_BAD) \
>> return -EINVAL; \
>> - raw_spin_lock_irq(&pci_lock); \
>> - if (unlikely(dev->block_cfg_access)) \
>> - pci_wait_cfg(dev); \
>> - ret = dev->bus->ops->read(dev->bus, dev->devfn, \
>> - pos, sizeof(type), &data); \
>> - raw_spin_unlock_irq(&pci_lock); \
>> - *val = (type)data; \
>> + ret = pci_read_config_##size(dev, pos, val); \
>> return pcibios_err_to_errno(ret); \
>> } \
>
> If it wasn't for the block_cfg_access check for user access, this would
> have been a nice code reuse cleanup.
Great fresh catch! I'll get that fixed in v2.
Alex
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH] PCI: access.c: Piggyback user config access on pci_read/write_*()
@ 2018-06-04 16:14 ` Alex_Gagniuc
0 siblings, 0 replies; 4+ messages in thread
From: Alex_Gagniuc @ 2018-06-04 16:14 UTC (permalink / raw)
To: keith.busch, mr.nuke.me
Cc: bhelgaas, Austin.Bolen, Shyam.Iyer, linux-pci, linux-kernel
On 6/4/2018 11:09 AM, Keith Busch wrote:=0A=
> On Mon, Jun 04, 2018 at 10:48:02AM -0500, Alexandru Gagniuc wrote:=0A=
>> +++ b/drivers/pci/access.c=0A=
>> @@ -223,16 +223,9 @@ int pci_user_read_config_##size \=0A=
>> (struct pci_dev *dev, int pos, type *val) \=0A=
>> { \=0A=
>> int ret =3D PCIBIOS_SUCCESSFUL; \=0A=
>> - u32 data =3D -1; \=0A=
>> if (PCI_##size##_BAD) \=0A=
>> return -EINVAL; \=0A=
>> - raw_spin_lock_irq(&pci_lock); \=0A=
>> - if (unlikely(dev->block_cfg_access)) \=0A=
>> - pci_wait_cfg(dev); \=0A=
>> - ret =3D dev->bus->ops->read(dev->bus, dev->devfn, \=0A=
>> - pos, sizeof(type), &data); \=0A=
>> - raw_spin_unlock_irq(&pci_lock); \=0A=
>> - *val =3D (type)data; \=0A=
>> + ret =3D pci_read_config_##size(dev, pos, val); \=0A=
>> return pcibios_err_to_errno(ret); \=0A=
>> } \=0A=
> =0A=
> If it wasn't for the block_cfg_access check for user access, this would=
=0A=
> have been a nice code reuse cleanup.=0A=
=0A=
Great fresh catch! I'll get that fixed in v2.=0A=
=0A=
Alex=0A=
=0A=
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2018-06-04 16:14 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2018-06-04 15:48 [PATCH] PCI: access.c: Piggyback user config access on pci_read/write_*() Alexandru Gagniuc
2018-06-04 16:12 ` Keith Busch
2018-06-04 16:14 ` Alex_Gagniuc
2018-06-04 16:14 ` Alex_Gagniuc
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.