RE: 53c400 driver

2000-11-22 Thread Igmar Palsenberg
On Tue, 21 Nov 2000, Dunlap, Randy wrote: > JE's UHCI driver (drivers/usb/uhci.[hc]) uses > nested_lock() and nested_unlock() for this. > Maybe it could help. I may should solve the nested spinlock issue.. It however doesn't solve the 100Kb+ pile of spaghetti the code is. I think I'll just

RE: 53c400 driver

2000-11-22 Thread Igmar Palsenberg
On Tue, 21 Nov 2000, Dunlap, Randy wrote: JE's UHCI driver (drivers/usb/uhci.[hc]) uses nested_lock() and nested_unlock() for this. Maybe it could help. I may should solve the nested spinlock issue.. It however doesn't solve the 100Kb+ pile of spaghetti the code is. I think I'll just start

RE: 53c400 driver

2000-11-21 Thread Dunlap, Randy
iews of my employer.| --- > -Original Message- > From: Alan Cox [mailto:[EMAIL PROTECTED]] > Sent: Tuesday, November 21, 2000 3:48 PM > To: [EMAIL PROTECTED] > Cc: [EMAIL PROTECTED] > Subject: Re: 53c400 driver > > &g

Re: 53c400 driver

2000-11-21 Thread Alan Cox
> 53c400a non-PNP still lock this system hard. It starts barking about a > busy SCSI bus, and then I can fsck again. > > To Alan : How hard is it to get thing beast (53c400 and family) to be SMP > safe ?? Or is it better to start over again ? The problem is that the code takes spinlocks

53c400 driver

2000-11-21 Thread Igmar Palsenberg
Hi, Support for some 53c400 cards is still bad (the non-PnP), so I'll start fixing this. I'll be my fist kernel job, so please spare me :)) Issues : 53c400a non-PNP still lock this system hard. It starts barking about a busy SCSI bus, and then I can fsck again. To Alan : How hard is it to

53c400 driver

2000-11-21 Thread Igmar Palsenberg
Hi, Support for some 53c400 cards is still bad (the non-PnP), so I'll start fixing this. I'll be my fist kernel job, so please spare me :)) Issues : 53c400a non-PNP still lock this system hard. It starts barking about a busy SCSI bus, and then I can fsck again. To Alan : How hard is it to

Re: 53c400 driver

2000-11-21 Thread Alan Cox
53c400a non-PNP still lock this system hard. It starts barking about a busy SCSI bus, and then I can fsck again. To Alan : How hard is it to get thing beast (53c400 and family) to be SMP safe ?? Or is it better to start over again ? The problem is that the code takes spinlocks recursively.

RE: 53c400 driver

2000-11-21 Thread Dunlap, Randy
of my employer.| --- -Original Message- From: Alan Cox [mailto:[EMAIL PROTECTED]] Sent: Tuesday, November 21, 2000 3:48 PM To: [EMAIL PROTECTED] Cc: [EMAIL PROTECTED] Subject: Re: 53c400 driver 53c400a non-PNP still lock this system hard