Fwd: Can kldload trigger pci bus rescan?
It would be cool if pccard and usb also reprobed when kldload ran. The usb case is slightly more complex --- having (say) uscanner claim something that ugen is currently claiming. ___ freebsd-hackers@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Can kldload trigger pci bus rescan?
In message: [EMAIL PROTECTED] Zaphod Beeblebrox [EMAIL PROTECTED] writes: : It would be cool if pccard and usb also reprobed when kldload ran. The usb : case is slightly more complex --- having (say) uscanner claim something that : ugen is currently claiming. PC Card has done so for years, OLDCARD and NEWCARD. In fact, you can even load and unload the bridge driver (cbb today, pcic in yesteryear) and the pccard bus driver. pccard works very hard to make sure that these things work. usb does also, but not in a manner that's useful. It does work with usb for simple devices when ugen isn't loaded into the kernel. However, the brain dead way that usb dollops out subdevices makes fixing it to work like rest of the system difficult... The port from NetBSD didn't take the time to make the device model match FreeBSD, but rather just shoe-horned it into the system. Warner ___ freebsd-hackers@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to [EMAIL PROTECTED]
Can kldload trigger pci bus rescan?
Hi guys: I tried to use kldload to load our HBA driver. But the driver's pci probe function can not find the HBA card! Does it mean that kldload can not trigger pci bus rescan? If so, what should I do for triggering pci bus rescan after loading our pci driver? Thank you for your help! Hong ___ freebsd-hackers@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Can kldload trigger pci bus rescan?
On Wednesday 26 April 2006 17:17, [EMAIL PROTECTED] wrote: I tried to use kldload to load our HBA driver. But the driver's pci probe function can not find the HBA card! Does it mean that kldload can not trigger pci bus rescan? If so, what should I do for triggering pci bus rescan after loading our pci driver? It will reprobe.. Well, all of the probe functions in the kernel will be called again. -- Daniel O'Connor software and network engineer for Genesis Software - http://www.gsoft.com.au The nice thing about standards is that there are so many of them to choose from. -- Andrew Tanenbaum GPG Fingerprint - 5596 B766 97C0 0E94 4347 295E E593 DC20 7B3F CE8C pgpKZKi45FocR.pgp Description: PGP signature
Re: Can kldload trigger pci bus rescan?
In message: [EMAIL PROTECTED]@promisechina.com [EMAIL PROTECTED] writes: : I tried to use kldload to load our HBA driver. But the driver's pci probe : function can not find the HBA card! Does it mean that kldload can not : trigger pci bus rescan? If so, what should I do for triggering pci bus : rescan after loading our pci driver? kldload of a driver with a pci attachment will cause all pci busses to reprobe/attach their unattached children automatically. When you say that the driver's probe function can not find the HBA card, what do you mean? That none of the devices presented to it are accepted? Warner ___ freebsd-hackers@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Can kldload trigger pci bus rescan?
[EMAIL PROTECTED] wrote: I tried to use kldload to load our HBA driver. But the driver's pci probe function can not find the HBA card! By this do you mean that when you have the card in the system, FreeBSD booted and you kldload the driver, you don't see kernel messages showing the appropriate attach messages? Is anything printed? Does it mean that kldload can not trigger pci bus rescan? If so, what should I do for triggering pci bus rescan after loading our pci driver? As Daniel and Warner already stated, the PCI rescan is automatic. This is more likely an identification or driver mismatch problem. Please provide the following: - The exact make and model of the HBA card. - The exact version of FreeBSD you're using. - Which driver you're trying to use. This information will allow us to nail down possible reasons why the attach is not occuring. ___ freebsd-hackers@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Can kldload trigger pci bus rescan?
[EMAIL PROTECTED] wrote: Hi guys: My situation is: (1) We are porting our HBA driver (Promise FastTrak TX4310 SoftRaid5) from Linux to FreeBSD(6.0). Our HBA driver can work normally under Linux. (2) At this time, I just wrote a sample driver code (see sr5.c in the attached file) to check the driver's probe function in FreeBSD. (3) After I successfully compiled the codes, I run make load (see Makefile and load.info in the attached file) for testing. (4) Our probe function was called indeed, but no matching pci device was found! (Yes, I have plugged our HBA card in the system). (5) Our HBA card information can be found by running pciconf (see pciconf.info in the attached file). The only two storage devices in the pciconf output are: [EMAIL PROTECTED]:7:1: class=0x01018a card=0x74411022 chip=0x74411022 rev=0x04 hdr=0x00 vendor = 'Advanced Micro Devices (AMD)' device = 'AMD-768 EIDE Controller' class= mass storage subclass = ATA [EMAIL PROTECTED]:4:0: class=0x010400 card=0x3515105a chip=0x3515105a rev=0x02 hdr=0x00 vendor = 'Promise Technology Inc' class= mass storage subclass = RAID The first looks like the ATA controller on the motherboard. The second is your Promise ATA RAID card. Both devices are already attached by FreeBSD's ATA framework. You will need to patch FreeBSD's ATA framework to not detect your card. The attached patch should do the trick (it removes the corresponding PCI ID for your card). I tested the patch to the point that it compiles, but I can't guarantee this will prevent the attach or even result in a bootable kernel, so test carefully. To apply the patch, do the following with root privileges: cd /usr/src # Replace /usr/src with the base of your src tree. patch /path/to/ata_no_PDC40719.diff If you have ata compiled into your kernel, rebuild your kernel. If you're loading the ATA modules, you can get away with just rebuilding the ata modules: cd /usr/src/sys/modules/ata make clean make depend make make install --- sys/dev/ata/ata-chipset.c.old Fri Oct 14 02:34:50 2005 +++ sys/dev/ata/ata-chipset.c Wed Apr 26 22:14:14 2006 @@ -2409,7 +2409,6 @@ { ATA_PDC40518, 0, PRMIO, PRSATA2, ATA_SA150, Promise PDC40518 }, { ATA_PDC40519, 0, PRMIO, PRSATA2, ATA_SA150, Promise PDC40519 }, { ATA_PDC40718, 0, PRMIO, PRSATA2, ATA_SA300, Promise PDC40718 }, - { ATA_PDC40719, 0, PRMIO, PRSATA2, ATA_SA300, Promise PDC40719 }, { 0, 0, 0, 0, 0, 0}}; char buffer[64]; uintptr_t devid = 0; --- sys/dev/ata/ata-pci.h.old Fri Oct 14 02:34:50 2005 +++ sys/dev/ata/ata-pci.h Wed Apr 26 22:14:45 2006 @@ -213,7 +213,6 @@ #define ATA_PDC405180x3d18105a #define ATA_PDC405190x3519105a #define ATA_PDC407180x3d17105a -#define ATA_PDC407190x3515105a #define ATA_PDC206170x6617105a #define ATA_PDC206180x6626105a #define ATA_PDC206190x6629105a ___ freebsd-hackers@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to [EMAIL PROTECTED]