Re: SiS 5513 rev 0x015597/5598 freeze diff
On Wed, Jan 04, 2012 at 02:55:04PM +1100, Jonathan Gray wrote: On Mon, Jan 02, 2012 at 04:35:39PM -0500, Loganaden Velvindron wrote: Attaching any CDROM/DVDRIVE causes this error message when mouting a disk. cd0(pciide0:0:1): timeout type: atapi c_bcount : 2048 c_skip : 0 The machine then freezes and a hard reboot is necessary. This issue has been reported by 2 users in the past: http://archives.neohapsis.com/archives/openbsd/2005-11/0782.html `` cd0(pciide0:0:0): timeout type: atapi c_bcount: 0 c_skip: 0 cd0(pciide0:0:0): timeout type: atapi c_bcount: 32 A workaround was to disable pciide: UKC disable pciide* 59 pciide* disabled UKC quit Continuing... Looking closely, these problems occur with this variant of the SiS 5513 EIDE chipset: pciide0 at pci0 dev 2 function 5 SiS 5513 EIDE rev 0x01: 5597/5598: DMA, chann el 0 wired to compatibility, channel 1 wired to compatibility atapiscsi0 at pciide0 channel 0 drive 1 scsibus0 at atapiscsi0: 2 targets cd0 at scsibus0 targ 0 lun 0: ATAPI, CD-ROM 52X, 172A ATAPI 5/cdrom removable cd0(pciide0:0:1): using PIO mode 4, DMA mode 2 By disabling dma, the issue goes away. Looking at pciide.c shows that there are 2 variants of the 5597/5598: The one with rev 0xd0 which support UDMA and the previous versions with no support for UltraDMA. To be more precise UDMA exists but it's buggy. It appears that rev 0x01 has a buggy DMA support. Falling back to PIO works fine. The diff disables DMA silently but not enabling DMA capabilities on rev 0x01 of 5597/5598. Disabling DMA is quite a large hammer, it seems more likely the timings need to be adjusted or some config space io needs to happen. I can't see any revision specific errata in the FreeBSD/Linux code. Chris (chris@) tracked down the problem. According to: http://svn.netlabs.org/repos/xata/ATA/tags/v1.8.11/doc/ChipList.txt Variants below 0xd0 have a problem using UDMA. Mine is hooked to this: pcib0 at pci0 dev 2 function 0 SiS 968 ISA rev 0x01 What's needed is to add it to the sis_hostbr table. pciide0 at pci0 dev 2 function 5 SiS 5513 EIDE rev 0x01: 968: DMA, channel 0 wired to compatibility, channel 1 wired to compatibility avail mem = 2042195968 (1947MB) pcib0 at pci0 dev 2 function 0 SiS 968 ISA rev 0x01 pciide0 at pci0 dev 2 function 5 SiS 5513 EIDE rev 0x01: 968: DMA, channel 0 wired to compatibility, channel 1 wired to compatibility (Huge thanks to Chris :-)) Here's the diff: Index: src/sys/dev/pci/pciide.c === RCS file: /cvs/src/sys/dev/pci/pciide.c,v retrieving revision 1.336 diff -u -p -r1.336 pciide.c --- src/sys/dev/pci/pciide.c4 Jan 2012 03:38:59 - 1.336 +++ src/sys/dev/pci/pciide.c14 Jan 2012 13:50:30 - @@ -4966,7 +4966,8 @@ static struct sis_hostbr_type { {PCI_PRODUCT_SIS_962, 0x00, 6, 962, SIS_TYPE_133NEW}, {PCI_PRODUCT_SIS_963, 0x00, 6, 963, SIS_TYPE_133NEW}, {PCI_PRODUCT_SIS_964, 0x00, 6, 964, SIS_TYPE_133NEW}, - {PCI_PRODUCT_SIS_965, 0x00, 6, 965, SIS_TYPE_133NEW} + {PCI_PRODUCT_SIS_965, 0x00, 6, 965, SIS_TYPE_133NEW}, + {PCI_PRODUCT_SIS_968, 0x00, 6, 968, SIS_TYPE_133NEW} }; static struct sis_hostbr_type *sis_hostbr_type_match;
Re: SiS 5513 rev 0x015597/5598 freeze diff
On Sat, Jan 14, 2012 at 08:56:58AM -0500, Loganaden Velvindron wrote: On Wed, Jan 04, 2012 at 02:55:04PM +1100, Jonathan Gray wrote: On Mon, Jan 02, 2012 at 04:35:39PM -0500, Loganaden Velvindron wrote: Attaching any CDROM/DVDRIVE causes this error message when mouting a disk. cd0(pciide0:0:1): timeout type: atapi c_bcount : 2048 c_skip : 0 The machine then freezes and a hard reboot is necessary. This issue has been reported by 2 users in the past: http://archives.neohapsis.com/archives/openbsd/2005-11/0782.html `` cd0(pciide0:0:0): timeout type: atapi c_bcount: 0 c_skip: 0 cd0(pciide0:0:0): timeout type: atapi c_bcount: 32 A workaround was to disable pciide: UKC disable pciide* 59 pciide* disabled UKC quit Continuing... Looking closely, these problems occur with this variant of the SiS 5513 EIDE chipset: pciide0 at pci0 dev 2 function 5 SiS 5513 EIDE rev 0x01: 5597/5598: DMA, chann el 0 wired to compatibility, channel 1 wired to compatibility atapiscsi0 at pciide0 channel 0 drive 1 scsibus0 at atapiscsi0: 2 targets cd0 at scsibus0 targ 0 lun 0: ATAPI, CD-ROM 52X, 172A ATAPI 5/cdrom removable cd0(pciide0:0:1): using PIO mode 4, DMA mode 2 By disabling dma, the issue goes away. Looking at pciide.c shows that there are 2 variants of the 5597/5598: The one with rev 0xd0 which support UDMA and the previous versions with no support for UltraDMA. To be more precise UDMA exists but it's buggy. It appears that rev 0x01 has a buggy DMA support. Falling back to PIO works fine. The diff disables DMA silently but not enabling DMA capabilities on rev 0x01 of 5597/5598. Disabling DMA is quite a large hammer, it seems more likely the timings need to be adjusted or some config space io needs to happen. I can't see any revision specific errata in the FreeBSD/Linux code. Chris (chris@) tracked down the problem. According to: http://svn.netlabs.org/repos/xata/ATA/tags/v1.8.11/doc/ChipList.txt Variants below 0xd0 have a problem using UDMA. Mine is hooked to this: pcib0 at pci0 dev 2 function 0 SiS 968 ISA rev 0x01 What's needed is to add it to the sis_hostbr table. pciide0 at pci0 dev 2 function 5 SiS 5513 EIDE rev 0x01: 968: DMA, channel 0 wired to compatibility, channel 1 wired to compatibility avail mem = 2042195968 (1947MB) pcib0 at pci0 dev 2 function 0 SiS 968 ISA rev 0x01 pciide0 at pci0 dev 2 function 5 SiS 5513 EIDE rev 0x01: 968: DMA, channel 0 wired to compatibility, channel 1 wired to compatibility (Huge thanks to Chris :-)) Ah yes, I'd forgotten the horrors of the sis part of pciide. I wish they'd stop making chipsets already. Which Chris by the way? Chris Cappuccio? Anyway, I'll likely commit the diff with the addition of 966 tomorrow, (the 967 doesn't seem to exist). Index: pciide.c === RCS file: /cvs/src/sys/dev/pci/pciide.c,v retrieving revision 1.336 diff -u -p -r1.336 pciide.c --- pciide.c4 Jan 2012 03:38:59 - 1.336 +++ pciide.c14 Jan 2012 14:12:40 - @@ -4966,7 +4966,9 @@ static struct sis_hostbr_type { {PCI_PRODUCT_SIS_962, 0x00, 6, 962, SIS_TYPE_133NEW}, {PCI_PRODUCT_SIS_963, 0x00, 6, 963, SIS_TYPE_133NEW}, {PCI_PRODUCT_SIS_964, 0x00, 6, 964, SIS_TYPE_133NEW}, - {PCI_PRODUCT_SIS_965, 0x00, 6, 965, SIS_TYPE_133NEW} + {PCI_PRODUCT_SIS_965, 0x00, 6, 965, SIS_TYPE_133NEW}, + {PCI_PRODUCT_SIS_966, 0x00, 6, 966, SIS_TYPE_133NEW}, + {PCI_PRODUCT_SIS_968, 0x00, 6, 968, SIS_TYPE_133NEW} }; static struct sis_hostbr_type *sis_hostbr_type_match;
Re: SiS 5513 rev 0x015597/5598 freeze diff
Yep, Chris Cappuccio. He says that there are still issues with the SiS code. Instead of defaulting to SiS 5597/5598, perhaps it would have been better not to attach as it would have made the problem more obvious. Also, I sent a mail to a guy who works on SiS chipset support in Linux. No reply up to now. I was looking for the datasheets. According to the URL chris sent me, the 966 should also work, but I can't test it. Thanks for the upcoming commit, I'll be able to upgrade painlessly to 5.1 :-) On Sat, Jan 14, 2012 at 6:21 PM, Jonathan Gray j...@goblin.cx wrote: On Sat, Jan 14, 2012 at 08:56:58AM -0500, Loganaden Velvindron wrote: On Wed, Jan 04, 2012 at 02:55:04PM +1100, Jonathan Gray wrote: On Mon, Jan 02, 2012 at 04:35:39PM -0500, Loganaden Velvindron wrote: Attaching any CDROM/DVDRIVE causes this error message when mouting a disk. cd0(pciide0:0:1): timeout type: atapi c_bcount : 2048 c_skip : 0 The machine then freezes and a hard reboot is necessary. This issue has been reported by 2 users in the past: http://archives.neohapsis.com/archives/openbsd/2005-11/0782.html `` cd0(pciide0:0:0): timeout type: atapi c_bcount: 0 c_skip: 0 cd0(pciide0:0:0): timeout type: atapi c_bcount: 32 A workaround was to disable pciide: UKC disable pciide* 59 pciide* disabled UKC quit Continuing... Looking closely, these problems occur with this variant of the SiS 5513 EIDE chipset: pciide0 at pci0 dev 2 function 5 SiS 5513 EIDE rev 0x01: 5597/5598: DMA, chann el 0 wired to compatibility, channel 1 wired to compatibility atapiscsi0 at pciide0 channel 0 drive 1 scsibus0 at atapiscsi0: 2 targets cd0 at scsibus0 targ 0 lun 0: ATAPI, CD-ROM 52X, 172A ATAPI 5/cdrom removable cd0(pciide0:0:1): using PIO mode 4, DMA mode 2 By disabling dma, the issue goes away. Looking at pciide.c shows that there are 2 variants of the 5597/5598: The one with rev 0xd0 which support UDMA and the previous versions with no support for UltraDMA. To be more precise UDMA exists but it's buggy. It appears that rev 0x01 has a buggy DMA support. Falling back to PIO works fine. The diff disables DMA silently but not enabling DMA capabilities on rev 0x01 of 5597/5598. Disabling DMA is quite a large hammer, it seems more likely the timings need to be adjusted or some config space io needs to happen. I can't see any revision specific errata in the FreeBSD/Linux code. Chris (chris@) tracked down the problem. According to: http://svn.netlabs.org/repos/xata/ATA/tags/v1.8.11/doc/ChipList.txt Variants below 0xd0 have a problem using UDMA. Mine is hooked to this: pcib0 at pci0 dev 2 function 0 SiS 968 ISA rev 0x01 What's needed is to add it to the sis_hostbr table. pciide0 at pci0 dev 2 function 5 SiS 5513 EIDE rev 0x01: 968: DMA, channel 0 wired to compatibility, channel 1 wired to compatibility avail mem = 2042195968 (1947MB) pcib0 at pci0 dev 2 function 0 SiS 968 ISA rev 0x01 pciide0 at pci0 dev 2 function 5 SiS 5513 EIDE rev 0x01: 968: DMA, channel 0 wired to compatibility, channel 1 wired to compatibility (Huge thanks to Chris :-)) Ah yes, I'd forgotten the horrors of the sis part of pciide. I wish they'd stop making chipsets already. Which Chris by the way? Chris Cappuccio? Anyway, I'll likely commit the diff with the addition of 966 tomorrow, (the 967 doesn't seem to exist). Index: pciide.c === RCS file: /cvs/src/sys/dev/pci/pciide.c,v retrieving revision 1.336 diff -u -p -r1.336 pciide.c --- pciide.c4 Jan 2012 03:38:59 - 1.336 +++ pciide.c14 Jan 2012 14:12:40 - @@ -4966,7 +4966,9 @@ static struct sis_hostbr_type { {PCI_PRODUCT_SIS_962, 0x00, 6, 962, SIS_TYPE_133NEW}, {PCI_PRODUCT_SIS_963, 0x00, 6, 963, SIS_TYPE_133NEW}, {PCI_PRODUCT_SIS_964, 0x00, 6, 964, SIS_TYPE_133NEW}, - {PCI_PRODUCT_SIS_965, 0x00, 6, 965, SIS_TYPE_133NEW} + {PCI_PRODUCT_SIS_965, 0x00, 6, 965, SIS_TYPE_133NEW}, + {PCI_PRODUCT_SIS_966, 0x00, 6, 966, SIS_TYPE_133NEW}, + {PCI_PRODUCT_SIS_968, 0x00, 6, 968, SIS_TYPE_133NEW} }; static struct sis_hostbr_type *sis_hostbr_type_match; -- Brightest day, Blackest night, No bug shall escape my sight, And those who worship evil's mind, be wary of my powers, puffy lantern's light !
Re: vmmap replacement -- please test
On Sat, Jan 14, 2012 at 8:42 AM, Ariane van der Steldt ari...@stack.nl wrote: Everybody thanks for testing this, it's been a tremendous help to me. As far as I'm concerned, this diff will be commited once we unlock after release (in a coordinated manner ofcourse, since this is uvm we're talking about). Great news. I'm using your diff on my amd64 laptop since 15 days and no problems. Thanks a lot for your work, David
Re: vmmap replacement -- please test
Ariane van der Steldt wrote [2012-01-14 08:42+0100]: As far as I'm concerned, this diff will be commited once we unlock after release (in a coordinated manner ofcourse, since this is uvm we're talking about). It's about time too, ofcourse: 64 revisions of the same diff is alot. :D -- Ariane I too cannot report anything to bring you forward (amd64). --steffen