PATA: UDMA settings in Alans patch-2.6.17-rc4-ide1.gz better than with 2.6.17-rc6-mm2
The current 2.6.17-rc6-mm2 tree sets UDMA speed for my IDE harddisk to 33, whereas Alans patch-2.6.17-rc4-ide1.gz patch has it at 100. This results in hdparm -T /dev/sda measurements on -mm2 roughly halved from 60+ MB/sec to 30+ MB/sec.
Measurements were in both cases taken on kernels based off 2.6.17-rc6, with Alans -rc4 patch forward-fitted.
The remaining three ATAPI devices get UDMA/33 in both setups.
Alans ide patch for 2.6.17-rc4 conflicts with this post-rc4 change in drivers/scsi/libata-core.c:
Following patch can be used in -rc[56] after applying Alans patch in order to apply the failed hunk and also keep the post-rc4 patch from git:
drivers/scsi/libata-core.c +2 -2
@@ -3736,14 +3736,14 @@ static void atapi_packet_task(void *_dat*/ spin_lock_irqsave (&ap->host_set->lock, flags); ap->flags &= ~ATA_FLAG_NOINTR ; -ata_data_xfer (ap, qc->cdb, qc->dev->cdb_len,1 ); + ap->ops->data_xfer (ap, qc->dev, qc->cdb, qc->dev->cdb_len,1 );ata_altstatus (ap);/* flush */ if (qc->tf.protocol ==ATA_PROT_ATAPI_DMA ) ap->ops->bmdma_start (qc);/* initiate bmdma */ spin_unlock_irqrestore (&ap->host_set->lock, flags); }else { -ata_data_xfer (ap, qc->cdb, qc->dev->cdb_len,1 ); + ap->ops->data_xfer (ap, qc->dev, qc->cdb, qc->dev->cdb_len,1 );ata_altstatus (ap);/* flush */ /* PIO commands are handled by polling */
dmesg section from 2.6.17-rc6 with Alans -rc4 patch fitted:
-----------------------------------------------------------pata_via 0000:00:0f.1: version 0.1.9
ACPI: PCI Interrupt 0000:00:0f.1[A] -> GSI 20 (level, low) -> IRQ 17
PCI: VIA IRQ fixup for 0000:00:0f.1, from 255 to 1
ata3: PATA max UDMA/133 cmd 0x1F0 ctl 0x3F6 bmdma 0xFC00 irq 14
ata3: dev 0 cfg 49:2f00 82:346b 83:7f01 84:4003 85:3c69 86:3c01 87:4003 88:203f
ata3: dev 0 ATA-7, max UDMA/100, 312581808 sectors: LBA48
ata3: dev 1 cfg 49:0f00 82:0000 83:0000 84:0000 85:0000 86:0000 87:0000 88:0407
ata3: dev 1 ATAPI, max UDMA/33via_do_set_mode: Mode=12 ast broken=Y udma=133 mul=4
t.act8b = 3, t.rec8b = 1, t.active = 3, t.recover = 1
FIT t.act8b = 2, t.rec8b = 0, t.active = 2, t.recover = 0via_do_set_mode: Mode=12 ast broken=Y udma=133 mul=4
t.act8b = 3, t.rec8b = 1, t.active = 3, t.recover = 1
FIT t.act8b = 2, t.rec8b = 0, t.active = 2, t.recover = 0via_do_set_mode: Mode=69 ast broken=Y udma=133 mul=4
t.act8b = 3, t.rec8b = 1, t.active = 3, t.recover = 1
FIT t.act8b = 2, t.rec8b = 0, t.active = 2, t.recover = 0via_do_set_mode: Mode=66 ast broken=Y udma=133 mul=4
t.act8b = 3, t.rec8b = 1, t.active = 3, t.recover = 1
FIT t.act8b = 2, t.rec8b = 0, t.active = 2, t.recover = 0
ata3: dev 0 configured for UDMA/100
ata3: dev 1 configured for UDMA/33
scsi2 : pata_via
isa bounce pool size: 16 pages
Vendor: ATA Model: SAMSUNG SP1614N Rev: TM10 Type: Direct-Access ANSI SCSI revision: 05 Vendor: TOSHIBA Model: DVD-ROM SD-M1612 Rev: X004 Type: CD-ROM ANSI SCSI revision: 05
ata4: PATA max UDMA/133 cmd 0x170 ctl 0x376 bmdma 0xFC08 irq 15
ata4: dev 0 cfg 49:0f00 82:0000 83:0000 84:0000 85:0000 86:0000 87:0000 88:0407
ata4: dev 0 ATAPI, max UDMA/33
ata4: dev 1 cfg 49:0f00 82:421c 83:0000 84:0000 85:0000 86:0000 87:0000 88:0407
ata4: dev 1 ATAPI, max UDMA/33via_do_set_mode: Mode=12 ast broken=Y udma=133 mul=4
t.act8b = 3, t.rec8b = 1, t.active = 3, t.recover = 1
FIT t.act8b = 2, t.rec8b = 0, t.active = 2, t.recover = 0via_do_set_mode: Mode=12 ast broken=Y udma=133 mul=4
t.act8b = 3, t.rec8b = 1, t.active = 3, t.recover = 1
FIT t.act8b = 2, t.rec8b = 0, t.active = 2, t.recover = 0via_do_set_mode: Mode=66 ast broken=Y udma=133 mul=4
t.act8b = 3, t.rec8b = 1, t.active = 3, t.recover = 1
FIT t.act8b = 2, t.rec8b = 0, t.active = 2, t.recover = 0via_do_set_mode: Mode=66 ast broken=Y udma=133 mul=4
t.act8b = 3, t.rec8b = 1, t.active = 3, t.recover = 1
FIT t.act8b = 2, t.rec8b = 0, t.active = 2, t.recover = 0
ata4: dev 0 configured for UDMA/33
ata4: dev 1 configured for UDMA/33
scsi3 : pata_via
Vendor: PLEXTOR Model: CD-R PX-W4012A Rev: 1.06 Type: CD-ROM ANSI SCSI revision: 05 Vendor: HL-DT-ST Model: DVDRAM GSA-4163B Rev: A103 Type: CD-ROM ANSI SCSI revision: 05
Cheers,
Arne
Thanks for the report. This is expected at the moment. The current core code picks speeds once per interface and keeps master/slave at the same speed. My patches fix that but before it can go mainstream I need to get the driver patches in and be sure no existing driver will break when I do it.
Alan