Re: Linux 2.4.28-rc1

Re: Linux 2.4.28-rc1

[O.Sezer]

There are many lost/forgotten patches posted here on lkml. Since 2.4.28
is near and 2.4 is going into "deep maintainance" mode soon, I gathered
a short list of some of them.  There, sure, are many more of them,  but
here it goes.
I think they deserve a re-review and re-consideration for inclusion.

Regards,
O. Sezer

The "list":
- Dave Jones:  AMD K7 MCE changes backported from 2.6.
   http://marc.theaimsgroup.com/?l=linux-kernel&m=106521456014393&w=2

- David Vrabel: TI CardBus PCI interrupt routing fix
   http://marc.theaimsgroup.com/?l=linux-kernel&m=108446444125446&w=2

- Michael Mueller: opti-viper pci-chipset support
   (have an updated-for-2.4.23+ patch for this)
   http://marc.theaimsgroup.com/?t=106698970100002&r=1&w=2
   http://marc.theaimsgroup.com/?l=linux-kernel&m=106698965700864&w=2

- Michael Frank: Highmem user-friendliness, Shutdown kernel on zone-
   alignment failure
   (have an updated patch)
   http://lkml.org/lkml/2004/2/7/51
   http://marc.theaimsgroup.com/?t=107619437300052&r=1&w=2
   http://marc.theaimsgroup.com/?l=linux-kernel&m=107619342911564&w=2

- Terry Hardie: 8 port SIIG serial card support
   http://marc.theaimsgroup.com/?l=linux-kernel&m=107765546507508&w=2

- Mauricio Martinez/Corey Minyard: fix a problem (multiple reads of
   the same data) while reading from a CDU31 SONY CD-ROM drive
   http://marc.theaimsgroup.com/?l=linux-kernel&m=106824345717317&w=2

- Roger Luethi: via-rhine, fix force media
   http://marc.theaimsgroup.com/?l=linux-kernel&m=108507431710317&w=2

- Robert White: usbserial hangup on disconnect
   http://marc.theaimsgroup.com/?t=108114071200002&r=1&w=2
   http://marc.theaimsgroup.com/?l=linux-kernel&m=108114073600529&w=2

- V Ganesh: ipaq, hangup tty on usb disconnect
   http://marc.theaimsgroup.com/?l=linux-usb-devel&m=109049411609590&w=2

- David M. Wilson: sis900 Wake-on-LAN support
   http://marc.theaimsgroup.com/?l=linux-kernel&m=105835662823748&w=2

- Thomas Gleixner: sis5513 fix for SiS962 chipset
   http://marc.theaimsgroup.com/?t=109482706500001&r=1&w=2
   http://marc.theaimsgroup.com/?l=linux-kernel&m=109482716300929&w=2

- Eric Sandeen: fix for large direct I/O
   http://marc.theaimsgroup.com/?l=linux-kernel&m=108197617129880&w=2

- Geert Uytterhoeven: smb_ops_unix compiler warning
   http://marc.theaimsgroup.com/?l=linux-kernel&m=107659039710361&w=2

- David A. Lethe: scsi_scan.c, look for LUNs on XYRATEX RAID subsystems
   http://marc.theaimsgroup.com/?l=linux-kernel&m=105534062611620&w=2

- Andrey Borzenkov: devfs deadlock on concurrent lookups on
   non-existent entry
   http://marc.theaimsgroup.com/?l=linux-kernel&m=105630542714518&w=2

- Jim Carter: apm.c, Dell Inspiron, limit rate of power status calls
   (without the star to the asm code)
   http://marc.theaimsgroup.com/?l=linux-kernel&m=106049225722612&w=2

- Eric Uhrhane: ATP867X PCI IDE driver: driver for the Acard/Artop PCI
   ATA/SATA cards (6885[LP]/6896[S]) based on the ATP867{A,B} chips.
   http://marc.theaimsgroup.com/?l=linux-kernel&m=108198418515134&w=2

- Jakub Bogusz: missing include in farsync WAN driver
   http://marc.theaimsgroup.com/?l=linux-kernel&m=109376793014054&w=2

- Willy Tarreau: MTU fix for tulip driver
   http://marc.theaimsgroup.com/?l=linux-kernel&m=109130863303540&w=2

- Ivan Kokshaysky: alpha, make bootimage and make bootpfile failure,
   boot failure
   http://marc.theaimsgroup.com/?t=109760337800003&r=1&w=2
   http://marc.theaimsgroup.com/?l=linux-kernel&m=109820176212217&w=2

- Sam King: usbserial, down function call being made from an interrupt
   handler
   http://marc.theaimsgroup.com/?t=109639065100005&r=1&w=2
   http://marc.theaimsgroup.com/?l=linux-usb-devel&m=109639053122263&w=2

- Wolfgang Mues: auerswald-usb, kernel oops at disconnect or reconnect
   time because of an endless urb resubmit
   http://marc.theaimsgroup.com/?l=linux-usb-devel&m=108465864428213&w=2

- Hilko Bengen: minor error in /proc/isapnp output
   http://marc.theaimsgroup.com/?l=linux-kernel&m=107607982001162&w=2

- Joshua Kwan:  scripts: Support output of new ld
   http://marc.theaimsgroup.com/?t=109549085600003&r=1&w=2

- Joshua Kwan: kbuild: use infobox instead of msgbox and 'sleep 5'
   http://marc.theaimsgroup.com/?l=linux-kernel&m=109549111519324&w=2

- Andre Hedrick: ide updates for 2.4.25
   http://www.kernel.org/pub/linux/kernel/people/hedrick/ide-2.4.25/

Re: Linux 2.4.28-rc1

[Marcelo Tosatti]

Hi,

If you have been suddenly CC'ed to this message please search
your name below - there is something which concerns you.

Replying only to the list, myself and O.Sezer is appreciated.

On Tue, Oct 26, 2004 at 05:02:44PM +0300, O.Sezer wrote:
> There are many lost/forgotten patches posted here on lkml. Since 2.4.28
> is near and 2.4 is going into "deep maintainance" mode soon, I gathered
> a short list of some of them. 

Oh it is hard to bookkeep all of this. I hope people check and resend, but
they dont do that always.

> There, sure, are many more of them,  but here it goes.

Please send'em all. I really appreciate your efforts.

> I think they deserve a re-review and re-consideration for inclusion.
> 
> Regards,
> O. Sezer
> 
> The "list":
> - Dave Jones:  AMD K7 MCE changes backported from 2.6.
>   http://marc.theaimsgroup.com/?l=linux-kernel&m=106521456014393&w=2

Should be merged - Dave?

> - David Vrabel: TI CardBus PCI interrupt routing fix
>   http://marc.theaimsgroup.com/?l=linux-kernel&m=108446444125446&w=2

Looks OK to me who dont have a clue about PCMCIA (it only tries
to handle failure case, pretty safe).

rmk, can you take a look at this patch please?

> - Michael Mueller: opti-viper pci-chipset support
>   (have an updated-for-2.4.23+ patch for this)
>   http://marc.theaimsgroup.com/?t=106698970100002&r=1&w=2
>   http://marc.theaimsgroup.com/?l=linux-kernel&m=106698965700864&w=2

Should be applied - v2.6 also lacks it AFAICS.

> - Michael Frank: Highmem user-friendliness, Shutdown kernel on zone-
>   alignment failure
>   (have an updated patch)
>   http://lkml.org/lkml/2004/2/7/51
>   http://marc.theaimsgroup.com/?t=107619437300052&r=1&w=2
>   http://marc.theaimsgroup.com/?l=linux-kernel&m=107619342911564&w=2

This is not really critical _and_ is the BUG is quite rare, thats
why I haven't applied it.

> - Terry Hardie: 8 port SIIG serial card support
>   http://marc.theaimsgroup.com/?l=linux-kernel&m=107765546507508&w=2

This one looks fine - I indeed missed it. Will apply to 2.4.29pre.

Is it present in v2.6 already?

> - Mauricio Martinez/Corey Minyard: fix a problem (multiple reads of
>   the same data) while reading from a CDU31 SONY CD-ROM drive
>   http://marc.theaimsgroup.com/?l=linux-kernel&m=106824345717317&w=2

I dont know this code at all and I'm not confident this is safe.

Maybe Jens can take a look at it?

What about v2.6?

> - Roger Luethi: via-rhine, fix force media
>   http://marc.theaimsgroup.com/?l=linux-kernel&m=108507431710317&w=2

Seems fine - Roger?

> - Robert White: usbserial hangup on disconnect
>   http://marc.theaimsgroup.com/?t=108114071200002&r=1&w=2
>   http://marc.theaimsgroup.com/?l=linux-kernel&m=108114073600529&w=2
> 
> - V Ganesh: ipaq, hangup tty on usb disconnect
>   http://marc.theaimsgroup.com/?l=linux-usb-devel&m=109049411609590&w=2

Pete, can you take a look at these?

> - David M. Wilson: sis900 Wake-on-LAN support
>   http://marc.theaimsgroup.com/?l=linux-kernel&m=105835662823748&w=2

Jeff?

> - Thomas Gleixner: sis5513 fix for SiS962 chipset
>   http://marc.theaimsgroup.com/?t=109482706500001&r=1&w=2
>   http://marc.theaimsgroup.com/?l=linux-kernel&m=109482716300929&w=2

Thomas?

> - Eric Sandeen: fix for large direct I/O
>   http://marc.theaimsgroup.com/?l=linux-kernel&m=108197617129880&w=2

Ouch - missed that. Applied. 

> - Geert Uytterhoeven: smb_ops_unix compiler warning
>   http://marc.theaimsgroup.com/?l=linux-kernel&m=107659039710361&w=2

Will apply to 2.4.29pre.

> - David A. Lethe: scsi_scan.c, look for LUNs on XYRATEX RAID subsystems
>   http://marc.theaimsgroup.com/?l=linux-kernel&m=105534062611620&w=2

[marcelo@logos scsi]$ grep XYRA *
scsi_scan.c:    {"XYRATEX", "RS", "*", BLIST_SPARSELUN | BLIST_LARGELUN},

Seems to be present already.

> - Andrey Borzenkov: devfs deadlock on concurrent lookups on
>   non-existent entry
>   http://marc.theaimsgroup.com/?l=linux-kernel&m=105630542714518&w=2

Looks OK, will look again during 2.4.29pre. 

> - Jim Carter: apm.c, Dell Inspiron, limit rate of power status calls
>   (without the star to the asm code)
>   http://marc.theaimsgroup.com/?l=linux-kernel&m=106049225722612&w=2

Dont know the code at all - seems to change generic code.

> - Eric Uhrhane: ATP867X PCI IDE driver: driver for the Acard/Artop PCI
>   ATA/SATA cards (6885[LP]/6896[S]) based on the ATP867{A,B} chips.
>   http://marc.theaimsgroup.com/?l=linux-kernel&m=108198418515134&w=2

Its a new driver - looks OK. 

If the maintainer really cares about why didnt he resend me? v2.6
already has this driver?

> - Jakub Bogusz: missing include in farsync WAN driver
>   http://marc.theaimsgroup.com/?l=linux-kernel&m=109376793014054&w=2

Applied.

> - Willy Tarreau: MTU fix for tulip driver
>   http://marc.theaimsgroup.com/?l=linux-kernel&m=109130863303540&w=2

Jeff?

> - Ivan Kokshaysky: alpha, make bootimage and make bootpfile failure,
>   boot failure
>   http://marc.theaimsgroup.com/?t=109760337800003&r=1&w=2
>   http://marc.theaimsgroup.com/?l=linux-kernel&m=109820176212217&w=2

Ivan, can you please resend me this?

> - Sam King: usbserial, down function call being made from an interrupt
>   handler
>   http://marc.theaimsgroup.com/?t=109639065100005&r=1&w=2
>   http://marc.theaimsgroup.com/?l=linux-usb-devel&m=109639053122263&w=2
> 
> - Wolfgang Mues: auerswald-usb, kernel oops at disconnect or reconnect
>   time because of an endless urb resubmit
>   http://marc.theaimsgroup.com/?l=linux-usb-devel&m=108465864428213&w=2

Pete, can you please a look.

> - Hilko Bengen: minor error in /proc/isapnp output
>   http://marc.theaimsgroup.com/?l=linux-kernel&m=107607982001162&w=2
>
> - Joshua Kwan:  scripts: Support output of new ld
>   http://marc.theaimsgroup.com/?t=109549085600003&r=1&w=2
> 
> - Joshua Kwan: kbuild: use infobox instead of msgbox and 'sleep 5'
>   http://marc.theaimsgroup.com/?l=linux-kernel&m=109549111519324&w=2

2.4.29pre, all three.

> - Andre Hedrick: ide updates for 2.4.25
>   http://www.kernel.org/pub/linux/kernel/people/hedrick/ide-2.4.25/

This I really dont know - I'm a complete IDE ignorant. 

Alan, Bart maybe?

Re: Linux 2.4.28-rc1

[Dave Jones]

On Tue, Oct 26, 2004 at 06:33:34PM -0200, Marcelo Tosatti wrote:
 
 > > - Dave Jones:  AMD K7 MCE changes backported from 2.6.
 > >   http://marc.theaimsgroup.com/?l=linux-kernel&m=106521456014393&w=2
 > 
 > Should be merged - Dave?

yep.

		Dave

Re: Linux 2.4.28-rc1

[O.Sezer]

[-- Attachment #1: Type: text/plain, Size: 1084 bytes --]

Marcelo Tosatti wrote:
> Hi,
> 
> If you have been suddenly CC'ed to this message please search
> your name below - there is something which concerns you.
> 
> Replying only to the list, myself and O.Sezer is appreciated.
> 
> On Tue, Oct 26, 2004 at 05:02:44PM +0300, O.Sezer wrote:
> 
>>There are many lost/forgotten patches posted here on lkml. Since 2.4.28
>>is near and 2.4 is going into "deep maintainance" mode soon, I gathered
>>a short list of some of them. 
> 
> 
> Oh it is hard to bookkeep all of this. I hope people check and resend, but
> they dont do that always.
> 
> 
>>There, sure, are many more of them,  but here it goes.
> 
> 
> Please send'em all. I really appreciate your efforts.
[...]
>>- Michael Mueller: opti-viper pci-chipset support
>>  (have an updated-for-2.4.23+ patch for this)
>>  http://marc.theaimsgroup.com/?t=106698970100002&r=1&w=2
>>  http://marc.theaimsgroup.com/?l=linux-kernel&m=106698965700864&w=2
> 
> 
> Should be applied - v2.6 also lacks it AFAICS.

Attached is a one that's supposed to apply cleanly to and work
with 2.4.23+ kernels.


[-- Attachment #2: opti-viper-2.4.23-pci-chipset.patch --]
[-- Type: text/plain, Size: 2218 bytes --]

--- 23/arch/i386/kernel/pci-irq.c~
+++ 23/arch/i386/kernel/pci-irq.c
@@ -241,18 +241,56 @@
 }
 
 static int pirq_opti_set(struct pci_dev *router, struct pci_dev *dev, int pirq, int irq)
 {
 	write_config_nybble(router, 0xb8, pirq >> 4, irq);
 	return 1;
 }
 
 /*
+ * OPTI Viper-M/N+: Bit field with 3 bits per entry.
+ * Due to the lack of a specification the information about this chipset
+ * was taken from the NetBSD source code.
+ */
+static int pirq_viper_get(struct pci_dev *router, struct pci_dev *dev, int pirq)
+{
+	static const int viper_irq_decode[] = { 0, 5, 9, 10, 11, 12, 14, 15 };
+	u32 irq;
+
+	pci_read_config_dword(router, 0x40, &irq);
+	irq >>= (pirq-1)*3;
+	irq &= 7;
+
+	return viper_irq_decode[irq];
+}
+
+static int pirq_viper_set(struct pci_dev *router, struct pci_dev *dev, int pirq, int irq)
+{
+	static const int viper_irq_map[] = { -1, -1, -1, -1, -1, 1, -1, -1, -1, 2, 3, 4, 5, -1, 6, 7 };
+	int newval = viper_irq_map[irq];
+	u32 val;
+	u32 mask = 7 << (3*(pirq-1));
+#if 0
+	mask |= 0x10000UL << (pirq-1);	/* edge triggered */
+#endif
+
+	if ( newval == -1 )
+		return 0;
+	
+	pci_read_config_dword(router, 0x40, &val);
+	val &= ~mask;
+	val |= newval << (3*(pirq-1));
+	pci_write_config_dword(router, 0x40, val);
+
+	return 1;
+}
+
+/*
  * Cyrix: nibble offset 0x5C
  */
 static int pirq_cyrix_get(struct pci_dev *router, struct pci_dev *dev, int pirq)
 {
 	return read_config_nybble(router, 0x5C, (pirq-1)^1);
 }
 
 static int pirq_cyrix_set(struct pci_dev *router, struct pci_dev *dev, int pirq, int irq)
 {
@@ -707,21 +745,28 @@
 
 static __init int opti_router_probe(struct irq_router *r, struct pci_dev *router, u16 device)
 {
 	switch(device)
 	{
 		case PCI_DEVICE_ID_OPTI_82C700:
 			r->name = "OPTI";
 			r->get = pirq_opti_get;
 			r->set = pirq_opti_set;
-			return 1;
+			break;
+		case PCI_DEVICE_ID_OPTI_82C558:
+			r->name = "OPTI VIPER";
+			r->get = pirq_viper_get;
+			r->set = pirq_viper_set;
+			break;
+		default:
+			return 0;
 	}
-	return 0;
+	return 1;
 }
 
 static __init int ite_router_probe(struct irq_router *r, struct pci_dev *router, u16 device)
 {
 	switch(device)
 	{
 		case PCI_DEVICE_ID_ITE_IT8330G_0:
 			r->name = "ITE";
 			r->get = pirq_ite_get;

Re: Linux 2.4.28-rc1

[Dave Jones]

On Wed, Oct 27, 2004 at 03:40:17PM +0300, O.Sezer wrote:
 >  /*
 > + * OPTI Viper-M/N+: Bit field with 3 bits per entry.
 > + * Due to the lack of a specification the information about this chipset
 > + * was taken from the NetBSD source code.
 > + */

I still have a lot of olde opti stuff downloaded from their ftp
site about 6 years ago, including Viper docs. If you (or
anyone else) is interested let me know, and I'll send them on.

		Dave

Re: Linux 2.4.28-rc1

[O.Sezer]

Dave Jones wrote:
> On Wed, Oct 27, 2004 at 03:40:17PM +0300, O.Sezer wrote:
>  >  /*
>  > + * OPTI Viper-M/N+: Bit field with 3 bits per entry.
>  > + * Due to the lack of a specification the information about this chipset
>  > + * was taken from the NetBSD source code.
>  > + */
> 
> I still have a lot of olde opti stuff downloaded from their ftp
> site about 6 years ago, including Viper docs. If you (or
> anyone else) is interested let me know, and I'll send them on.
> 
> 		Dave
> 

Would be nice to have around. (They're not online anymore, I assume?)

Ozkan.

Re: Linux 2.4.28-rc1

[Dave Jones]

On Wed, Oct 27, 2004 at 06:29:37PM +0300, O.Sezer wrote:

 > >I still have a lot of olde opti stuff downloaded from their ftp
 > >site about 6 years ago, including Viper docs. If you (or
 > >anyone else) is interested let me know, and I'll send them on.
 > 
 > Would be nice to have around. (They're not online anymore, I assume?)
 
not afaik, ftp.opti.com.tw doesn't seem to be around anymore,
and their .com site doesn't answer ftp. (and www is really nondescript)

I'll tar them up and send them your way.

		Dave

Re: Linux 2.4.28-rc1

[Marcelo Tosatti]

On Wed, Oct 27, 2004 at 03:40:17PM +0300, O.Sezer wrote:
> Marcelo Tosatti wrote:
> >Hi,
> >
> >If you have been suddenly CC'ed to this message please search
> >your name below - there is something which concerns you.
> >
> >Replying only to the list, myself and O.Sezer is appreciated.
> >
> >On Tue, Oct 26, 2004 at 05:02:44PM +0300, O.Sezer wrote:
> >
> >>There are many lost/forgotten patches posted here on lkml. Since 2.4.28
> >>is near and 2.4 is going into "deep maintainance" mode soon, I gathered
> >>a short list of some of them. 
> >
> >
> >Oh it is hard to bookkeep all of this. I hope people check and resend, but
> >they dont do that always.
> >
> >
> >>There, sure, are many more of them,  but here it goes.
> >
> >
> >Please send'em all. I really appreciate your efforts.
> [...]
> >>- Michael Mueller: opti-viper pci-chipset support
> >> (have an updated-for-2.4.23+ patch for this)
> >> http://marc.theaimsgroup.com/?t=106698970100002&r=1&w=2
> >> http://marc.theaimsgroup.com/?l=linux-kernel&m=106698965700864&w=2
> >
> >
> >Should be applied - v2.6 also lacks it AFAICS.
> 
> Attached is a one that's supposed to apply cleanly to and work
> with 2.4.23+ kernels.

Ozkan,

Someone needs to check v2.6.

Can you or Michael do that please?

I'll save it to 2.4.29pre. 

Re: Linux 2.4.28-rc1

[O.Sezer]

Marcelo Tosatti wrote:
> On Wed, Oct 27, 2004 at 03:40:17PM +0300, O.Sezer wrote:
> 
>>Marcelo Tosatti wrote:
>>
>>>Hi,
>>>
>>>If you have been suddenly CC'ed to this message please search
>>>your name below - there is something which concerns you.
>>>
>>>Replying only to the list, myself and O.Sezer is appreciated.
>>>
>>>On Tue, Oct 26, 2004 at 05:02:44PM +0300, O.Sezer wrote:
>>>
>>>
>>>>There are many lost/forgotten patches posted here on lkml. Since 2.4.28
>>>>is near and 2.4 is going into "deep maintainance" mode soon, I gathered
>>>>a short list of some of them. 
>>>
>>>
>>>Oh it is hard to bookkeep all of this. I hope people check and resend, but
>>>they dont do that always.
>>>
>>>
>>>
>>>>There, sure, are many more of them,  but here it goes.
>>>
>>>
>>>Please send'em all. I really appreciate your efforts.
>>
>>[...]
>>
>>>>- Michael Mueller: opti-viper pci-chipset support
>>>>(have an updated-for-2.4.23+ patch for this)
>>>>http://marc.theaimsgroup.com/?t=106698970100002&r=1&w=2
>>>>http://marc.theaimsgroup.com/?l=linux-kernel&m=106698965700864&w=2
>>>
>>>
>>>Should be applied - v2.6 also lacks it AFAICS.
>>
>>Attached is a one that's supposed to apply cleanly to and work
>>with 2.4.23+ kernels.
> 
> 
> Ozkan,
> 
> Someone needs to check v2.6.
> 
> Can you or Michael do that please?
> 
> I'll save it to 2.4.29pre. 
> 

2.6 doesn't have it but I don't know if it needs it (it should, but...)
I don't have the hardware anymore, so Michael can look after it, I'm
sure.

Re: Linux 2.4.28-rc1

[O.Sezer]

[-- Attachment #1: Type: text/plain, Size: 1899 bytes --]

O.Sezer wrote:
> Marcelo Tosatti wrote:
> 
>> On Wed, Oct 27, 2004 at 03:40:17PM +0300, O.Sezer wrote:
>>
>>> Marcelo Tosatti wrote:
>>>
>>>> Hi,
>>>>
>>>> If you have been suddenly CC'ed to this message please search
>>>> your name below - there is something which concerns you.
>>>>
>>>> Replying only to the list, myself and O.Sezer is appreciated.
>>>>
>>>> On Tue, Oct 26, 2004 at 05:02:44PM +0300, O.Sezer wrote:
>>>>
>>>>
>>>>> There are many lost/forgotten patches posted here on lkml. Since 
>>>>> 2.4.28
>>>>> is near and 2.4 is going into "deep maintainance" mode soon, I 
>>>>> gathered
>>>>> a short list of some of them. 
>>>>
>>>>
>>>>
>>>> Oh it is hard to bookkeep all of this. I hope people check and 
>>>> resend, but
>>>> they dont do that always.
>>>>
>>>>
>>>>
>>>>> There, sure, are many more of them,  but here it goes.
>>>>
>>>>
>>>>
>>>> Please send'em all. I really appreciate your efforts.
>>>
>>>
>>> [...]
>>>
>>>>> - Michael Mueller: opti-viper pci-chipset support
>>>>> (have an updated-for-2.4.23+ patch for this)
>>>>> http://marc.theaimsgroup.com/?t=106698970100002&r=1&w=2
>>>>> http://marc.theaimsgroup.com/?l=linux-kernel&m=106698965700864&w=2
>>>>
>>>>
>>>>
>>>> Should be applied - v2.6 also lacks it AFAICS.
>>>
>>>
>>> Attached is a one that's supposed to apply cleanly to and work
>>> with 2.4.23+ kernels.
>>
>>
>>
>> Ozkan,
>>
>> Someone needs to check v2.6.
>>
>> Can you or Michael do that please?
>>
>> I'll save it to 2.4.29pre.
> 
> 
> 2.6 doesn't have it but I don't know if it needs it (it should, but...)
> I don't have the hardware anymore, so Michael can look after it, I'm
> sure.
> 

OK, out of curiosity, I did a quick re-diff of the patch, applied onto
the correct file, and did a real quick compile test (and nothing more).
Nothing went bad. Can't say anything about functionality (Michael?).
It is attached. Whom to send it? Linus, akpm?

[-- Attachment #2: optiviper_26_test.diff --]
[-- Type: text/plain, Size: 1635 bytes --]

--- linux-2.6.9/arch/i386/pci/irq.c~
+++ linux-2.6.9/arch/i386/pci/irq.c
@@ -250,6 +250,44 @@
 }
 
 /*
+ * OPTI Viper-M/N+: Bit field with 3 bits per entry.
+ * Due to the lack of a specification the information about this chipset
+ * was taken from the NetBSD source code.
+ */
+static int pirq_viper_get(struct pci_dev *router, struct pci_dev *dev, int pirq)
+{
+	static const int viper_irq_decode[] = { 0, 5, 9, 10, 11, 12, 14, 15 };
+	u32 irq;
+
+	pci_read_config_dword(router, 0x40, &irq);
+	irq >>= (pirq-1)*3;
+	irq &= 7;
+
+	return viper_irq_decode[irq];
+}
+
+static int pirq_viper_set(struct pci_dev *router, struct pci_dev *dev, int pirq, int irq)
+{
+	static const int viper_irq_map[] = { -1, -1, -1, -1, -1, 1, -1, -1, -1, 2, 3, 4, 5, -1, 6, 7 };
+	int newval = viper_irq_map[irq];
+	u32 val;
+	u32 mask = 7 << (3*(pirq-1));
+#if 0
+	mask |= 0x10000UL << (pirq-1);	/* edge triggered */
+#endif
+
+	if ( newval == -1 )
+		return 0;
+	
+	pci_read_config_dword(router, 0x40, &val);
+	val &= ~mask;
+	val |= newval << (3*(pirq-1));
+	pci_write_config_dword(router, 0x40, val);
+
+	return 1;
+}
+
+/*
  * Cyrix: nibble offset 0x5C
  * 0x5C bits 7:4 is INTB bits 3:0 is INTA 
  * 0x5D bits 7:4 is INTD bits 3:0 is INTC
@@ -567,9 +605,16 @@
 			r->name = "OPTI";
 			r->get = pirq_opti_get;
 			r->set = pirq_opti_set;
-			return 1;
+			break;
+		case PCI_DEVICE_ID_OPTI_82C558:
+			r->name = "OPTI VIPER";
+			r->get = pirq_viper_get;
+			r->set = pirq_viper_set;
+			break;
+		default:
+			return 0;
 	}
-	return 0;
+	return 1;
 }
 
 static __init int ite_router_probe(struct irq_router *r, struct pci_dev *router, u16 device)

Re: Linux 2.4.28-rc1

[Pete Zaitcev]

On Tue, 26 Oct 2004 18:33:34 -0200, Marcelo Tosatti <marcelo.tosatti@cyclades.com> wrote:

> > - Robert White: usbserial hangup on disconnect
> >   http://marc.theaimsgroup.com/?t=108114071200002&r=1&w=2
> >   http://marc.theaimsgroup.com/?l=linux-kernel&m=108114073600529&w=2
> > 
> > - V Ganesh: ipaq, hangup tty on usb disconnect
> >   http://marc.theaimsgroup.com/?l=linux-usb-devel&m=109049411609590&w=2
> 
> Pete, can you take a look at these?

I already did, I just keep procrastinating. Better is an enemy of good.
I wanted to adopt 2.6 style refcounting, because there's a possibility
of oops lurking in 2.4 code. That would move the hangup call into the
close/release path.

The top part of V. Ganesh's patch with a missing initialization you already
have in -rc1. The bottom part Robert's patch covers (modulo the relocation
of the whole fragment).

I'll make a better 2.6 style patch as soon as I can. Maybe today even.

-- Pete

Re: Linux 2.4.28-rc1

[Jeff Garzik]

Marcelo Tosatti wrote:
>>- David M. Wilson: sis900 Wake-on-LAN support
>>  http://marc.theaimsgroup.com/?l=linux-kernel&m=105835662823748&w=2
> 
> 
> Jeff?

Two principal objections:

1) 2.6 first

2) should use the already-present ethtool WOL interface rather than 
inventing your own


>>- Willy Tarreau: MTU fix for tulip driver
>>  http://marc.theaimsgroup.com/?l=linux-kernel&m=109130863303540&w=2
> 
> 
> Jeff?

Looks OK but I would prefer at least a 2.6 version in parallel...

	Jeff

Re: Linux 2.4.28-rc1

[Russell King]

On Tue, Oct 26, 2004 at 06:33:34PM -0200, Marcelo Tosatti wrote:
> > - David Vrabel: TI CardBus PCI interrupt routing fix
> >   http://marc.theaimsgroup.com/?l=linux-kernel&m=108446444125446&w=2
> 
> Looks OK to me who dont have a clue about PCMCIA (it only tries
> to handle failure case, pretty safe).
> 
> rmk, can you take a look at this patch please?

I think this should be rejected, but since I'm not in posession of 2.4
source code anymore, I don't really know.

We did a lot of work in 2.6 to fix these issues properly.  I don't think
a "simple" fix is acceptable.  However, changing "irqmux" to be u32 is
obviously correct whatever.

-- 
Russell King
 Linux kernel    2.6 ARM Linux   - http://www.arm.linux.org.uk/
 maintainer of:  2.6 PCMCIA      - http://pcmcia.arm.linux.org.uk/
                 2.6 Serial core

Re: Linux 2.4.28-rc1

[Roger Luethi]

On Tue, 26 Oct 2004 18:33:34 -0200, Marcelo Tosatti wrote:
> If you have been suddenly CC'ed to this message please search
> your name below - there is something which concerns you.

It might have helped to CC me, too :-/.

> > - Roger Luethi: via-rhine, fix force media
> >   http://marc.theaimsgroup.com/?l=linux-kernel&m=108507431710317&w=2
> 
> Seems fine - Roger?

ACK. It's not a big deal either way, but it should go in if only for
correctness' sake.

Roger

Re: Linux 2.4.28-rc1

[Barry K. Nathan]

On Tue, Oct 26, 2004 at 05:02:44PM +0300, O.Sezer wrote:
> There are many lost/forgotten patches posted here on lkml. Since 2.4.28
> is near and 2.4 is going into "deep maintainance" mode soon, I gathered
> a short list of some of them.  There, sure, are many more of them,  but
> here it goes.
> I think they deserve a re-review and re-consideration for inclusion.
[snip]

Here's another one:
Jason Baron: 2.4.28-pre3 tty/ldisc fixes
http://marc.theaimsgroup.com/?l=linux-kernel&m=109604869516678&w=2

AFAICT the above patch is the fix for:
CAN-2004-0814: Linux terminal layer races
http://marc.theaimsgroup.com/?l=bugtraq&m=109837405025108&w=2

This patch seems to be working fine for me, but I don't know if anyone
else has really tested it at all, nor do I know (one way or the other)
if the security issues are serious enough to apply this for 2.4.28-rc
and not 2.4.29-pre. Also, I'm running on a single-processor system with
no HyperThreading, so if there are any SMP-related issues then I have no
way of experiencing them.

Anyway, since it's a security fix (unless I'm mistaken), I guess it's
worth considering for inclusion...

-Barry K. Nathan <barryn@pobox.com>

Re: Linux 2.4.28-rc1

[Marcelo Tosatti]

On Tue, Oct 26, 2004 at 07:52:22PM -0700, Barry K. Nathan wrote:
> On Tue, Oct 26, 2004 at 05:02:44PM +0300, O.Sezer wrote:
> > There are many lost/forgotten patches posted here on lkml. Since 2.4.28
> > is near and 2.4 is going into "deep maintainance" mode soon, I gathered
> > a short list of some of them.  There, sure, are many more of them,  but
> > here it goes.
> > I think they deserve a re-review and re-consideration for inclusion.
> [snip]
> 
> Here's another one:
> Jason Baron: 2.4.28-pre3 tty/ldisc fixes
> http://marc.theaimsgroup.com/?l=linux-kernel&m=109604869516678&w=2
> 
> AFAICT the above patch is the fix for:
> CAN-2004-0814: Linux terminal layer races
> http://marc.theaimsgroup.com/?l=bugtraq&m=109837405025108&w=2
> 
> This patch seems to be working fine for me, but I don't know if anyone
> else has really tested it at all, nor do I know (one way or the other)
> if the security issues are serious enough to apply this for 2.4.28-rc
> and not 2.4.29-pre. Also, I'm running on a single-processor system with
> no HyperThreading, so if there are any SMP-related issues then I have no
> way of experiencing them.
> 
> Anyway, since it's a security fix (unless I'm mistaken), I guess it's
> worth considering for inclusion...

Hi Barry,

I think this should be applied as well. The next 
-pre makes me more comfortable.

Jason, Alan?

Linux 2.4.28-rc1

[Marcelo Tosatti]

Hi, 

Here goes the first release candidate of v2.4.28.

It contains a small number of changes from -pre4, a couple of libata bugfixes,
a PIIX IDE driver DMA bugfix, USB fixes, and some tmpfs corrections.

Detailed changelog follows.

Summary of changes from v2.4.28-pre4 to v2.4.28-rc1
============================================

<chaus:rz.uni-potsdam.de>:
  o Fix bug in PIIX code where DMA could be turned on without proper hw configuration (bugzilla bug #3473)

Bartlomiej Zolnierkiewicz:
  o libata: PCI IDE legacy mode fix
  o [libata] do not memset() SCSI request buf in a get-reference style function
  o [libata piix] Fix PATA UDMA masks

Benjamin Herrenschmidt:
  o Mikael Pettersson: PowerPC 745x coherency fix

Dave Jones:
  o davej CREDITS update

François Romieu:
  o sata_nv: enable hotplug event on successfull init only
  o sata_nv: wrong failure path and leak
  o sata_nv: housekeeping for goto labels

Herbert Xu:
  o Fix hiddev devfs oops

Hugh Dickins:
  o tmpfs: stop negative dentries
  o tmpfs: fix shmem_file_write return value

Jake Moilanen:
  o PPC64 build break

Jeff Garzik:
  o [libata] add hook, and export functions needed for sata2 drivers
  o [libata] add sata_uli driver for ULi (formerly ALi) SATA

Jens Axboe:
  o scsi io completion bug

Maciej W. Rozycki:
  o "console=" parameter ignored

Marcelo Tosatti:
  o Changed EXTRAVERSION to -rc1

Margit Schubert-While:
  o Add prism54 to MAINTAINERS

Paul Fulghum:
  o serial send_break duration fix

Pete Zaitcev:
  o Crash with cat /proc/bus/usb/devices and disconnect

Özkan Sezer:
  o e1000 driver, gcc-3.4 inlining fix

Re: Linux 2.4.28-rc1

[Jan Engelhardt]

>Hi,
>
>Here goes the first release candidate of v2.4.28.

Hi,

if this one <http://lkml.org/lkml/2004/10/22/74> makes it in, it could also be
considered for the 2.4 tree.


Regards,
Jan Engelhardt
-- 
Gesellschaft für Wissenschaftliche Datenverarbeitung
Am Fassberg, 37077 Göttingen, www.gwdg.de

Re: Linux 2.4.28-rc1

[Marcelo Tosatti]

On Sat, Oct 23, 2004 at 12:14:37AM +0200, Jan Engelhardt wrote:
> >Hi,
> >
> >Here goes the first release candidate of v2.4.28.
> 
> Hi,
> 
> if this one <http://lkml.org/lkml/2004/10/22/74> makes it in, it could also be
> considered for the 2.4 tree.

Sure, looks pretty easy, can Jan Kara or you prepare a backport please?

Re: Linux 2.4.28-rc1

[Jan Engelhardt]

>> >Here goes the first release candidate of v2.4.28.
>>
>> Hi,
>>
>> if this one <http://lkml.org/lkml/2004/10/22/74> makes it in, it could also be
>> considered for the 2.4 tree.
>
>Sure, looks pretty easy, can Jan Kara or you prepare a backport please?

There is no difference, the patch will also apply to 2.4 without problems.



Jan Engelhardt
-- 
Gesellschaft für Wissenschaftliche Datenverarbeitung
Am Fassberg, 37077 Göttingen, www.gwdg.de

Re: Linux 2.4.28-rc1

[Paul Fulghum]

On Fri, 2004-10-22 at 13:59, Marcelo Tosatti wrote:
> Here goes the first release candidate of v2.4.28.

Any chance of getting this in?

-- 
Paul Fulghum
paulkf@microgate.com

>From paulkf@microgate.com Fri Oct  8 13:20:56 2004
Subject: [PATCH] serial receive lockup fix
From: Paul Fulghum <paulkf@microgate.com>
To: Marcelo Tosatti <marcelo.tosatti@cyclades.com>
Cc: linux-kernel <linux-kernel@vger.kernel.org>

Fix lockup caused by serial driver not clearing
receive interrupt if flip buffer becomes full.

Signed-off-by: Paul Fulghum <paulkf@microgate.com>



--- a/drivers/char/serial.c	2004-09-29 09:08:35.000000000 -0500
+++ b/drivers/char/serial.c	2004-09-29 09:09:07.000000000 -0500
@@ -573,8 +573,19 @@
 	do {
 		if (tty->flip.count >= TTY_FLIPBUF_SIZE) {
 			tty->flip.tqueue.routine((void *) tty);
-			if (tty->flip.count >= TTY_FLIPBUF_SIZE)
+			if (tty->flip.count >= TTY_FLIPBUF_SIZE) {
+				/* no room in flip buffer, discard rx FIFO contents to clear IRQ
+				 * *FIXME* Hardware with auto flow control
+				 * would benefit from leaving the data in the FIFO and
+				 * disabling the rx IRQ until space becomes available.
+				 */
+				do {
+					serial_inp(info, UART_RX);
+					icount->overrun++;
+					*status = serial_inp(info, UART_LSR);
+				} while ((*status & UART_LSR_DR) && (max_count-- > 0));
 				return;		// if TTY_DONT_FLIP is set
+			}
 		}
 		ch = serial_inp(info, UART_RX);
 		*tty->flip.char_buf_ptr = ch;

Re: Linux 2.4.28-rc1

[Marcelo Tosatti]

On Sat, Oct 23, 2004 at 08:30:20AM -0500, Paul Fulghum wrote:
> On Fri, 2004-10-22 at 13:59, Marcelo Tosatti wrote:
> > Here goes the first release candidate of v2.4.28.
> 
> Any chance of getting this in?

Oh I missed that, sorry (I've seen it and thought I had
applied).

Well, there it is.

> 
> -- 
> Paul Fulghum
> paulkf@microgate.com
> 
> >From paulkf@microgate.com Fri Oct  8 13:20:56 2004
> Subject: [PATCH] serial receive lockup fix
> From: Paul Fulghum <paulkf@microgate.com>
> To: Marcelo Tosatti <marcelo.tosatti@cyclades.com>
> Cc: linux-kernel <linux-kernel@vger.kernel.org>
> 
> Fix lockup caused by serial driver not clearing
> receive interrupt if flip buffer becomes full.
> 
> Signed-off-by: Paul Fulghum <paulkf@microgate.com>
> 
> 
> 
> --- a/drivers/char/serial.c	2004-09-29 09:08:35.000000000 -0500
> +++ b/drivers/char/serial.c	2004-09-29 09:09:07.000000000 -0500
> @@ -573,8 +573,19 @@
>  	do {
>  		if (tty->flip.count >= TTY_FLIPBUF_SIZE) {
>  			tty->flip.tqueue.routine((void *) tty);
> -			if (tty->flip.count >= TTY_FLIPBUF_SIZE)
> +			if (tty->flip.count >= TTY_FLIPBUF_SIZE) {
> +				/* no room in flip buffer, discard rx FIFO contents to clear IRQ
> +				 * *FIXME* Hardware with auto flow control
> +				 * would benefit from leaving the data in the FIFO and
> +				 * disabling the rx IRQ until space becomes available.
> +				 */
> +				do {
> +					serial_inp(info, UART_RX);
> +					icount->overrun++;
> +					*status = serial_inp(info, UART_LSR);
> +				} while ((*status & UART_LSR_DR) && (max_count-- > 0));
>  				return;		// if TTY_DONT_FLIP is set
> +			}
>  		}
>  		ch = serial_inp(info, UART_RX);
>  		*tty->flip.char_buf_ptr = ch;
> 
>