* [PATCH 8/28] rt2x00: optimize mac/bssid writing
@ 2007-02-28 14:07 Ivo van Doorn
2007-02-28 17:36 ` Michael Buesch
0 siblings, 1 reply; 7+ messages in thread
From: Ivo van Doorn @ 2007-02-28 14:07 UTC (permalink / raw)
To: John W. Linville; +Cc: linux-wireless
Handling the mac and bssid configuration can be done much easier
by writing the passed data directly into the register instead
of moving it to a local variable first.
Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com>
---
diff --git a/drivers/net/wireless/mac80211/rt2x00/rt2400pci.c b/drivers/net/wireless/mac80211/rt2x00/rt2400pci.c
index 27e151d..b6bf9f3 100644
--- a/drivers/net/wireless/mac80211/rt2x00/rt2400pci.c
+++ b/drivers/net/wireless/mac80211/rt2x00/rt2400pci.c
@@ -319,14 +319,11 @@ static inline void rt2400pci_close_debugfs(struct rt2x00_dev *rt2x00dev){}
*/
static void rt2400pci_config_bssid(struct rt2x00_dev *rt2x00dev, u8 *bssid)
{
- u32 reg[2] = { 0, 0 };
-
/*
* The BSSID is passed to us as an array of bytes,
* that array is little endian, so no need for byte ordering.
*/
- memcpy(®, bssid, ETH_ALEN);
- rt2x00_register_multiwrite(rt2x00dev, CSR5, ®[0], sizeof(reg));
+ rt2x00_register_multiwrite(rt2x00dev, CSR5, (u32*)bssid, ETH_ALEN);
}
static void rt2400pci_config_promisc(struct rt2x00_dev *rt2x00dev, int promisc)
@@ -674,14 +671,11 @@ static void rt2400pci_config_phymode(struct rt2x00_dev *rt2x00dev,
static void rt2400pci_config_mac_address(struct rt2x00_dev *rt2x00dev,
void *addr)
{
- u32 reg[2] = { 0, 0 };
-
/*
* The MAC address is passed to us as an array of bytes,
* that array is little endian, so no need for byte ordering.
*/
- memcpy(®, addr, ETH_ALEN);
- rt2x00_register_multiwrite(rt2x00dev, CSR3, ®[0], sizeof(reg));
+ rt2x00_register_multiwrite(rt2x00dev, CSR3, (u32*)addr, ETH_ALEN);
}
/*
diff --git a/drivers/net/wireless/mac80211/rt2x00/rt2500pci.c b/drivers/net/wireless/mac80211/rt2x00/rt2500pci.c
index 6234d20..69ada0e 100644
--- a/drivers/net/wireless/mac80211/rt2x00/rt2500pci.c
+++ b/drivers/net/wireless/mac80211/rt2x00/rt2500pci.c
@@ -319,14 +319,11 @@ static inline void rt2500pci_close_debugfs(struct rt2x00_dev *rt2x00dev){}
*/
static void rt2500pci_config_bssid(struct rt2x00_dev *rt2x00dev, u8 *bssid)
{
- u32 reg[2] = { 0, 0 };
-
/*
* The BSSID is passed to us as an array of bytes,
* that array is little endian, so no need for byte ordering.
*/
- memcpy(®, bssid, ETH_ALEN);
- rt2x00_register_multiwrite(rt2x00dev, CSR5, ®[0], sizeof(reg));
+ rt2x00_register_multiwrite(rt2x00dev, CSR5, (u32*)bssid, ETH_ALEN);
}
static void rt2500pci_config_promisc(struct rt2x00_dev *rt2x00dev, int promisc)
@@ -749,14 +746,11 @@ static void rt2500pci_config_phymode(struct rt2x00_dev *rt2x00dev,
static void rt2500pci_config_mac_address(struct rt2x00_dev *rt2x00dev,
void *addr)
{
- u32 reg[2] = { 0, 0 };
-
/*
* The MAC address is passed to us as an array of bytes,
* that array is little endian, so no need for byte ordering.
*/
- memcpy(®, addr, ETH_ALEN);
- rt2x00_register_multiwrite(rt2x00dev, CSR3, ®[0], sizeof(reg));
+ rt2x00_register_multiwrite(rt2x00dev, CSR3, (u32*)addr, ETH_ALEN);
}
/*
diff --git a/drivers/net/wireless/mac80211/rt2x00/rt2500usb.c b/drivers/net/wireless/mac80211/rt2x00/rt2500usb.c
index c8c0f5a..86c06ac 100644
--- a/drivers/net/wireless/mac80211/rt2x00/rt2500usb.c
+++ b/drivers/net/wireless/mac80211/rt2x00/rt2500usb.c
@@ -373,14 +373,11 @@ static inline void rt2500usb_close_debugfs(struct rt2x00_dev *rt2x00dev){}
*/
static void rt2500usb_config_bssid(struct rt2x00_dev *rt2x00dev, u8 *bssid)
{
- u16 reg[3] = { 0, 0, 0 };
-
/*
* The BSSID is passed to us as an array of bytes,
* that array is little endian, so no need for byte ordering.
*/
- memcpy(®, bssid, ETH_ALEN);
- rt2x00_register_multiwrite(rt2x00dev, MAC_CSR5, ®[0], sizeof(reg));
+ rt2x00_register_multiwrite(rt2x00dev, MAC_CSR5, (u16*)bssid, ETH_ALEN);
}
static void rt2500usb_config_promisc(struct rt2x00_dev *rt2x00dev, int promisc)
@@ -781,14 +778,11 @@ static void rt2500usb_config_phymode(struct rt2x00_dev *rt2x00dev,
static void rt2500usb_config_mac_address(struct rt2x00_dev *rt2x00dev,
void *addr)
{
- u16 reg[3] = { 0, 0, 0 };
-
/*
* The MAC address is passed to us as an array of bytes,
* that array is little endian, so no need for byte ordering.
*/
- memcpy(®, addr, ETH_ALEN);
- rt2x00_register_multiwrite(rt2x00dev, MAC_CSR2, ®[0], sizeof(reg));
+ rt2x00_register_multiwrite(rt2x00dev, MAC_CSR2, (u16*)addr, ETH_ALEN);
}
/*
diff --git a/drivers/net/wireless/mac80211/rt2x00/rt61pci.c b/drivers/net/wireless/mac80211/rt2x00/rt61pci.c
index 4224a57..2f7d18e 100644
--- a/drivers/net/wireless/mac80211/rt2x00/rt61pci.c
+++ b/drivers/net/wireless/mac80211/rt2x00/rt61pci.c
@@ -351,16 +351,18 @@ static inline void rt61pci_close_debugfs(struct rt2x00_dev *rt2x00dev){}
*/
static void rt61pci_config_bssid(struct rt2x00_dev *rt2x00dev, u8 *bssid)
{
- u32 reg[2] = { 0, 0 };
+ u32 reg;
/*
* The BSSID is passed to us as an array of bytes,
* that array is little endian, so no need for byte ordering.
* We only need to set the BSS ID MASK at the correct offset.
*/
- memcpy(®, bssid, ETH_ALEN);
- rt2x00_set_field32(®[1], MAC_CSR5_BSS_ID_MASK, 3);
- rt2x00_register_multiwrite(rt2x00dev, MAC_CSR4, ®[0], sizeof(reg));
+ rt2x00_register_multiwrite(rt2x00dev, MAC_CSR4, (u32*)bssid, ETH_ALEN);
+
+ rt2x00_register_read(rt2x00dev, MAC_CSR5, ®);
+ rt2x00_set_field32(®, MAC_CSR5_BSS_ID_MASK, 3);
+ rt2x00_register_write(rt2x00dev, MAC_CSR5, reg);
}
static void rt61pci_config_promisc(struct rt2x00_dev *rt2x00dev, int promisc)
@@ -970,7 +972,7 @@ static void rt61pci_config_phymode(struct rt2x00_dev *rt2x00dev,
static void rt61pci_config_mac_address(struct rt2x00_dev *rt2x00dev,
void *addr)
{
- u32 reg[2] = { 0, 0 };
+ u32 reg;
/*
* The MAC address is passed to us as an array of bytes,
@@ -978,9 +980,11 @@ static void rt61pci_config_mac_address(struct rt2x00_dev *rt2x00dev,
* We only need to set the MAC_CSR3_UNICAST_TO_ME_MASK
* at the correct offset.
*/
- memcpy(®, addr, ETH_ALEN);
- rt2x00_set_field32(®[1], MAC_CSR3_UNICAST_TO_ME_MASK, 0xff);
- rt2x00_register_multiwrite(rt2x00dev, MAC_CSR2, ®[0], sizeof(reg));
+ rt2x00_register_multiwrite(rt2x00dev, MAC_CSR2, (u32*)addr, ETH_ALEN);
+
+ rt2x00_register_read(rt2x00dev, MAC_CSR3, ®);
+ rt2x00_set_field32(®, MAC_CSR3_UNICAST_TO_ME_MASK, 0xff);
+ rt2x00_register_write(rt2x00dev, MAC_CSR3, reg);
}
/*
diff --git a/drivers/net/wireless/mac80211/rt2x00/rt73usb.c b/drivers/net/wireless/mac80211/rt2x00/rt73usb.c
index ef16d82..f0fff49 100644
--- a/drivers/net/wireless/mac80211/rt2x00/rt73usb.c
+++ b/drivers/net/wireless/mac80211/rt2x00/rt73usb.c
@@ -358,16 +358,18 @@ static inline void rt73usb_close_debugfs(struct rt2x00_dev *rt2x00dev){}
*/
static void rt73usb_config_bssid(struct rt2x00_dev *rt2x00dev, u8 *bssid)
{
- u32 reg[2] = { 0, 0 };
+ u32 reg;
/*
* The BSSID is passed to us as an array of bytes,
* that array is little endian, so no need for byte ordering.
* We only need to set the BSS ID MASK at the correct offset.
*/
- memcpy(®, bssid, ETH_ALEN);
- rt2x00_set_field32(®[1], MAC_CSR5_BSS_ID_MASK, 3);
- rt2x00_register_multiwrite(rt2x00dev, MAC_CSR4, ®[0], sizeof(reg));
+ rt2x00_register_multiwrite(rt2x00dev, MAC_CSR4, (u32*)bssid, ETH_ALEN);
+
+ rt2x00_register_read(rt2x00dev, MAC_CSR5, ®);
+ rt2x00_set_field32(®, MAC_CSR5_BSS_ID_MASK, 3);
+ rt2x00_register_write(rt2x00dev, MAC_CSR5, reg);
}
static void rt73usb_config_promisc(struct rt2x00_dev *rt2x00dev, int promisc)
@@ -868,7 +870,7 @@ static void rt73usb_config_phymode(struct rt2x00_dev *rt2x00dev,
static void rt73usb_config_mac_address(struct rt2x00_dev *rt2x00dev,
void *addr)
{
- u32 reg[2] = { 0, 0 };
+ u32 reg;
/*
* The MAC address is passed to us as an array of bytes,
@@ -876,9 +878,11 @@ static void rt73usb_config_mac_address(struct rt2x00_dev *rt2x00dev,
* We only need to set the MAC_CSR3_UNICAST_TO_ME_MASK
* at the correct offset.
*/
- memcpy(®, addr, ETH_ALEN);
- rt2x00_set_field32(®[1], MAC_CSR3_UNICAST_TO_ME_MASK, 0xff);
- rt2x00_register_multiwrite(rt2x00dev, MAC_CSR2, ®[0], sizeof(reg));
+ rt2x00_register_multiwrite(rt2x00dev, MAC_CSR2, (u32*)addr, ETH_ALEN);
+
+ rt2x00_register_read(rt2x00dev, MAC_CSR3, ®);
+ rt2x00_set_field32(®, MAC_CSR3_UNICAST_TO_ME_MASK, 0xff);
+ rt2x00_register_write(rt2x00dev, MAC_CSR3, reg);
}
/*
^ permalink raw reply related [flat|nested] 7+ messages in thread* Re: [PATCH 8/28] rt2x00: optimize mac/bssid writing
2007-02-28 14:07 [PATCH 8/28] rt2x00: optimize mac/bssid writing Ivo van Doorn
@ 2007-02-28 17:36 ` Michael Buesch
2007-02-28 18:15 ` Ivo van Doorn
0 siblings, 1 reply; 7+ messages in thread
From: Michael Buesch @ 2007-02-28 17:36 UTC (permalink / raw)
To: Ivo van Doorn; +Cc: John W. Linville, linux-wireless
On Wednesday 28 February 2007 15:07, Ivo van Doorn wrote:
> Handling the mac and bssid configuration can be done much easier
> by writing the passed data directly into the register instead
> of moving it to a local variable first.
>
> Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com>
>
> ---
>
> diff --git a/drivers/net/wireless/mac80211/rt2x00/rt2400pci.c b/drivers/net/wireless/mac80211/rt2x00/rt2400pci.c
> index 27e151d..b6bf9f3 100644
> --- a/drivers/net/wireless/mac80211/rt2x00/rt2400pci.c
> +++ b/drivers/net/wireless/mac80211/rt2x00/rt2400pci.c
> @@ -319,14 +319,11 @@ static inline void rt2400pci_close_debugfs(struct rt2x00_dev *rt2x00dev){}
> */
> static void rt2400pci_config_bssid(struct rt2x00_dev *rt2x00dev, u8 *bssid)
> {
> - u32 reg[2] = { 0, 0 };
> -
> /*
> * The BSSID is passed to us as an array of bytes,
> * that array is little endian, so no need for byte ordering.
^^^^^^^^^^^^^^^^^^^^^^^^^^^
> */
> - memcpy(®, bssid, ETH_ALEN);
> - rt2x00_register_multiwrite(rt2x00dev, CSR5, ®[0], sizeof(reg));
> + rt2x00_register_multiwrite(rt2x00dev, CSR5, (u32*)bssid, ETH_ALEN);
^^^^^^
This doesn't break on BE machines?
[Snipped the rest of the patch. Same comment applies there, too]
--
Greetings Michael.
^ permalink raw reply [flat|nested] 7+ messages in thread* Re: [PATCH 8/28] rt2x00: optimize mac/bssid writing
2007-02-28 17:36 ` Michael Buesch
@ 2007-02-28 18:15 ` Ivo van Doorn
2007-02-28 18:24 ` Michael Buesch
0 siblings, 1 reply; 7+ messages in thread
From: Ivo van Doorn @ 2007-02-28 18:15 UTC (permalink / raw)
To: Michael Buesch; +Cc: John W. Linville, linux-wireless
On Wednesday 28 February 2007 18:36, Michael Buesch wrote:
> On Wednesday 28 February 2007 15:07, Ivo van Doorn wrote:
> > Handling the mac and bssid configuration can be done much easier
> > by writing the passed data directly into the register instead
> > of moving it to a local variable first.
> >
> > Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com>
> >
> > ---
> >
> > diff --git a/drivers/net/wireless/mac80211/rt2x00/rt2400pci.c b/drivers/net/wireless/mac80211/rt2x00/rt2400pci.c
> > index 27e151d..b6bf9f3 100644
> > --- a/drivers/net/wireless/mac80211/rt2x00/rt2400pci.c
> > +++ b/drivers/net/wireless/mac80211/rt2x00/rt2400pci.c
> > @@ -319,14 +319,11 @@ static inline void rt2400pci_close_debugfs(struct rt2x00_dev *rt2x00dev){}
> > */
> > static void rt2400pci_config_bssid(struct rt2x00_dev *rt2x00dev, u8 *bssid)
> > {
> > - u32 reg[2] = { 0, 0 };
> > -
> > /*
> > * The BSSID is passed to us as an array of bytes,
> > * that array is little endian, so no need for byte ordering.
> ^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > */
> > - memcpy(®, bssid, ETH_ALEN);
> > - rt2x00_register_multiwrite(rt2x00dev, CSR5, ®[0], sizeof(reg));
> > + rt2x00_register_multiwrite(rt2x00dev, CSR5, (u32*)bssid, ETH_ALEN);
> ^^^^^^
>
> This doesn't break on BE machines?
No, the multiwrite (just like multiread) does not perform byteordering,
so the device will receive each byte in the correct order.
Ivo
^ permalink raw reply [flat|nested] 7+ messages in thread* Re: [PATCH 8/28] rt2x00: optimize mac/bssid writing
2007-02-28 18:15 ` Ivo van Doorn
@ 2007-02-28 18:24 ` Michael Buesch
2007-02-28 18:30 ` Ivo van Doorn
0 siblings, 1 reply; 7+ messages in thread
From: Michael Buesch @ 2007-02-28 18:24 UTC (permalink / raw)
To: Ivo van Doorn; +Cc: John W. Linville, linux-wireless
On Wednesday 28 February 2007 19:15, Ivo van Doorn wrote:
> On Wednesday 28 February 2007 18:36, Michael Buesch wrote:
> > On Wednesday 28 February 2007 15:07, Ivo van Doorn wrote:
> > > Handling the mac and bssid configuration can be done much easier
> > > by writing the passed data directly into the register instead
> > > of moving it to a local variable first.
> > >
> > > Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com>
> > >
> > > ---
> > >
> > > diff --git a/drivers/net/wireless/mac80211/rt2x00/rt2400pci.c b/drivers/net/wireless/mac80211/rt2x00/rt2400pci.c
> > > index 27e151d..b6bf9f3 100644
> > > --- a/drivers/net/wireless/mac80211/rt2x00/rt2400pci.c
> > > +++ b/drivers/net/wireless/mac80211/rt2x00/rt2400pci.c
> > > @@ -319,14 +319,11 @@ static inline void rt2400pci_close_debugfs(struct rt2x00_dev *rt2x00dev){}
> > > */
> > > static void rt2400pci_config_bssid(struct rt2x00_dev *rt2x00dev, u8 *bssid)
> > > {
> > > - u32 reg[2] = { 0, 0 };
> > > -
> > > /*
> > > * The BSSID is passed to us as an array of bytes,
> > > * that array is little endian, so no need for byte ordering.
> > ^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > > */
> > > - memcpy(®, bssid, ETH_ALEN);
> > > - rt2x00_register_multiwrite(rt2x00dev, CSR5, ®[0], sizeof(reg));
> > > + rt2x00_register_multiwrite(rt2x00dev, CSR5, (u32*)bssid, ETH_ALEN);
> > ^^^^^^
> >
> > This doesn't break on BE machines?
>
> No, the multiwrite (just like multiread) does not perform byteordering,
> so the device will receive each byte in the correct order.
No wait. It's not about performing byteordering somewhere.
It's that casting a bytearray (which is little endian) into an
u32 array (which is CPU endian) is almost always wrong.
Do you writel this array in a loop inside of register_multiwrite?
If yes, it's broken. writel expects values in CPU endianess.
This array will always be little endian.
--
Greetings Michael.
^ permalink raw reply [flat|nested] 7+ messages in thread* Re: [PATCH 8/28] rt2x00: optimize mac/bssid writing
2007-02-28 18:24 ` Michael Buesch
@ 2007-02-28 18:30 ` Ivo van Doorn
2007-02-28 18:46 ` Michael Buesch
0 siblings, 1 reply; 7+ messages in thread
From: Ivo van Doorn @ 2007-02-28 18:30 UTC (permalink / raw)
To: Michael Buesch; +Cc: John W. Linville, linux-wireless
On Wednesday 28 February 2007 19:24, Michael Buesch wrote:
> On Wednesday 28 February 2007 19:15, Ivo van Doorn wrote:
> > On Wednesday 28 February 2007 18:36, Michael Buesch wrote:
> > > On Wednesday 28 February 2007 15:07, Ivo van Doorn wrote:
> > > > Handling the mac and bssid configuration can be done much easier
> > > > by writing the passed data directly into the register instead
> > > > of moving it to a local variable first.
> > > >
> > > > Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com>
> > > >
> > > > ---
> > > >
> > > > diff --git a/drivers/net/wireless/mac80211/rt2x00/rt2400pci.c b/drivers/net/wireless/mac80211/rt2x00/rt2400pci.c
> > > > index 27e151d..b6bf9f3 100644
> > > > --- a/drivers/net/wireless/mac80211/rt2x00/rt2400pci.c
> > > > +++ b/drivers/net/wireless/mac80211/rt2x00/rt2400pci.c
> > > > @@ -319,14 +319,11 @@ static inline void rt2400pci_close_debugfs(struct rt2x00_dev *rt2x00dev){}
> > > > */
> > > > static void rt2400pci_config_bssid(struct rt2x00_dev *rt2x00dev, u8 *bssid)
> > > > {
> > > > - u32 reg[2] = { 0, 0 };
> > > > -
> > > > /*
> > > > * The BSSID is passed to us as an array of bytes,
> > > > * that array is little endian, so no need for byte ordering.
> > > ^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > > > */
> > > > - memcpy(®, bssid, ETH_ALEN);
> > > > - rt2x00_register_multiwrite(rt2x00dev, CSR5, ®[0], sizeof(reg));
> > > > + rt2x00_register_multiwrite(rt2x00dev, CSR5, (u32*)bssid, ETH_ALEN);
> > > ^^^^^^
> > >
> > > This doesn't break on BE machines?
> >
> > No, the multiwrite (just like multiread) does not perform byteordering,
> > so the device will receive each byte in the correct order.
>
> No wait. It's not about performing byteordering somewhere.
> It's that casting a bytearray (which is little endian) into an
> u32 array (which is CPU endian) is almost always wrong.
>
> Do you writel this array in a loop inside of register_multiwrite?
> If yes, it's broken. writel expects values in CPU endianess.
> This array will always be little endian.
register_multiread uses memcpy_fromio(), and
register_multiwrite uses memcpy_toio().
So that should be safe to use right?
Otherwise would making register_multiwrite accept a void* pointer be better?
Ivo
^ permalink raw reply [flat|nested] 7+ messages in thread* Re: [PATCH 8/28] rt2x00: optimize mac/bssid writing
2007-02-28 18:30 ` Ivo van Doorn
@ 2007-02-28 18:46 ` Michael Buesch
2007-02-28 18:48 ` Ivo van Doorn
0 siblings, 1 reply; 7+ messages in thread
From: Michael Buesch @ 2007-02-28 18:46 UTC (permalink / raw)
To: Ivo van Doorn; +Cc: John W. Linville, linux-wireless
On Wednesday 28 February 2007 19:30, Ivo van Doorn wrote:
> On Wednesday 28 February 2007 19:24, Michael Buesch wrote:
> > On Wednesday 28 February 2007 19:15, Ivo van Doorn wrote:
> > > On Wednesday 28 February 2007 18:36, Michael Buesch wrote:
> > > > On Wednesday 28 February 2007 15:07, Ivo van Doorn wrote:
> > > > > Handling the mac and bssid configuration can be done much easier
> > > > > by writing the passed data directly into the register instead
> > > > > of moving it to a local variable first.
> > > > >
> > > > > Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com>
> > > > >
> > > > > ---
> > > > >
> > > > > diff --git a/drivers/net/wireless/mac80211/rt2x00/rt2400pci.c b/drivers/net/wireless/mac80211/rt2x00/rt2400pci.c
> > > > > index 27e151d..b6bf9f3 100644
> > > > > --- a/drivers/net/wireless/mac80211/rt2x00/rt2400pci.c
> > > > > +++ b/drivers/net/wireless/mac80211/rt2x00/rt2400pci.c
> > > > > @@ -319,14 +319,11 @@ static inline void rt2400pci_close_debugfs(struct rt2x00_dev *rt2x00dev){}
> > > > > */
> > > > > static void rt2400pci_config_bssid(struct rt2x00_dev *rt2x00dev, u8 *bssid)
> > > > > {
> > > > > - u32 reg[2] = { 0, 0 };
> > > > > -
> > > > > /*
> > > > > * The BSSID is passed to us as an array of bytes,
> > > > > * that array is little endian, so no need for byte ordering.
> > > > ^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > > > > */
> > > > > - memcpy(®, bssid, ETH_ALEN);
> > > > > - rt2x00_register_multiwrite(rt2x00dev, CSR5, ®[0], sizeof(reg));
> > > > > + rt2x00_register_multiwrite(rt2x00dev, CSR5, (u32*)bssid, ETH_ALEN);
> > > > ^^^^^^
> > > >
> > > > This doesn't break on BE machines?
> > >
> > > No, the multiwrite (just like multiread) does not perform byteordering,
> > > so the device will receive each byte in the correct order.
> >
> > No wait. It's not about performing byteordering somewhere.
> > It's that casting a bytearray (which is little endian) into an
> > u32 array (which is CPU endian) is almost always wrong.
> >
> > Do you writel this array in a loop inside of register_multiwrite?
> > If yes, it's broken. writel expects values in CPU endianess.
> > This array will always be little endian.
>
> register_multiread uses memcpy_fromio(), and
> register_multiwrite uses memcpy_toio().
Ah, ok.
> So that should be safe to use right?
Yes.
> Otherwise would making register_multiwrite accept a void* pointer be better?
Yeah it would. The argument type is wrong. If the function uses
memcpy_foo it expects an u8 array (or void array, whatever you prefer).
--
Greetings Michael.
^ permalink raw reply [flat|nested] 7+ messages in thread* Re: [PATCH 8/28] rt2x00: optimize mac/bssid writing
2007-02-28 18:46 ` Michael Buesch
@ 2007-02-28 18:48 ` Ivo van Doorn
0 siblings, 0 replies; 7+ messages in thread
From: Ivo van Doorn @ 2007-02-28 18:48 UTC (permalink / raw)
To: Michael Buesch; +Cc: John W. Linville, linux-wireless
> > > Do you writel this array in a loop inside of register_multiwrite?
> > > If yes, it's broken. writel expects values in CPU endianess.
> > > This array will always be little endian.
> >
> > register_multiread uses memcpy_fromio(), and
> > register_multiwrite uses memcpy_toio().
>
> Ah, ok.
>
> > So that should be safe to use right?
>
> Yes.
>
> > Otherwise would making register_multiwrite accept a void* pointer be better?
>
> Yeah it would. The argument type is wrong. If the function uses
> memcpy_foo it expects an u8 array (or void array, whatever you prefer).
Ok, thanks. I'll fix this as soon as possible.
Ivo
^ permalink raw reply [flat|nested] 7+ messages in thread
end of thread, other threads:[~2007-02-28 18:48 UTC | newest]
Thread overview: 7+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2007-02-28 14:07 [PATCH 8/28] rt2x00: optimize mac/bssid writing Ivo van Doorn
2007-02-28 17:36 ` Michael Buesch
2007-02-28 18:15 ` Ivo van Doorn
2007-02-28 18:24 ` Michael Buesch
2007-02-28 18:30 ` Ivo van Doorn
2007-02-28 18:46 ` Michael Buesch
2007-02-28 18:48 ` Ivo van Doorn
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).