I guess maybe i have a problem with math then: the name plate on the drive says LBA = 320173056. Cyl = 16383, heads=16, sectors=63. that output from the driver ( et al ) says thus:
Nov 11 10:18:23 lx kernel: hda: 60030432 sectors (30736 MB) w/2048KiB Cache, CHS=59554/16/63 Nov 11 10:18:23 lx kernel: hdc: 35239680 sectors (18043 MB) w/1966KiB Cache, CHS=34960/16/63 Nov 11 10:18:23 lx kernel: hde: 320173056 sectors (163929 MB) w/2048KiB Cache, CHS=317632/255/63, UDMA(33) The number of sectors of hde & the name plate match. The number of sectors/trk match. The number of heads & cyl's dont match. using 255 isnt so bad, but it does not match. MY MATHEMATICAL PROBLEM is that when u multiply 63 * 255 * 317632 u get something like 5102758080. If the # of heads were 16 then cyl matches. Fdisk says that there are 255 heads, 63 sectors, 19929 cylinders. Multiplying out 63*255*19929 = 320159385. 320173056 - 320159385 = 13671 difference in count ( 63*255=16065 ) There is some 6.9mb missing. maybe you can point out where my math has gone wrong? /gat As for the driver itself, it thinks its a udma(33) rather that the udma(133) ? HDA goes just as fast as HDE using hdparm -t ( both drives are ( now ) on the same ATA/133 controller). cached blocks on HDE can go as fast as 128mb/sec. uncached is around 28-29mb/sec. I was also expecting some log message to say PIO mode 6 enabled. reading raw the whole disk ( dd if=/dev/hde of=/dev/null bs=1024k ) yields a throughput of 28mb/sec. a tar (cf /dev/null .) took some 90min to process 40gb of storage, which is a throughput of some 7mb/sec. Ouch. [EMAIL PROTECTED] wrote: > > GAT, > > I am correct in the math, I wrote the rules in the ATA/ATAPI-6 document > that describes the content in words 100-103 in the IDENTIFY page of the > drive. Trust that I got it correct, and I did not screw up what I > created. > > Cheers, > > Andre Hedrick > The Linux ATA/IDE guy > > On Thu, 8 Nov 2001, Uncle George wrote: > > > I think that at this moment that the mathematics of your cchhss > > calculations are in need of some attention, as well as the capacity > > numbers of the drive are in doubt. > > > > "hde: 320173056 sectors (163929 MB)" > > > > the marketeers, who believe in base 10, say 160gb. The professionals, > > who believe in base 2, say its 152gb. Is there a third group who > > believes its 163gb? I really need to take out the drive and see what the > > numbers really are! > > /gat > > > > It also appears that one can create a file > 2gig on an intel system. > > which is interesting. Its also interesting that 'cat' gets a seg fault > > at 32gig. its also interesting that write() will fail at a 2 gig > > boundary. lotsa interesting things, which were of mo interest to me for > > some time, as i was for a long time on a 64bit machine where such > > limitations did not happen. > > > > as for me, i dont support the private group that creates midi/sound > > device drivers for linux . They charge for their drivers. They also seem > > to have signed exclusive disclosure arrangements with the manufacturers, > > which prevents people like me from gaining the necessary infoormation > > about their device(s). Out of that comes a public group that does the > > same thing, but under a more 'linux' orientation. But this is politics, > > and generally outside of my political peripheral vision, by which i dont > > see too clearly the NDA agreements between XFREE & various video card > > manufacturers. > > > > > > BTW, minor point - the 20269 is not listed on the make config of the > > kernel build. > > > > [EMAIL PROTECTED] wrote: > > > > > > GAT, > > > > > > Device: Maxtor 4G160H8 Serial Number: G802RYZE > > > LBA 0 DMA Read Test = 53.74 MB/Sec (4.65 Seconds) > > > Outer Diameter Sequential DMA Read Test = 34.73 MB/Sec (7.20 Seconds) > > > Inner Diameter Sequential DMA Read Test = 17.82 MB/Sec (14.03 Seconds) > > > > > > Corrected rates for under reporting: > > > > > > LBA 0 DMA Read Test = 72.55 MB/Sec > > > Outer Diameter Sequential DMA Read Test = 46.89 MB/Sec > > > Inner Diameter Sequential DMA Read Test = 24.06 MB/Sec > > > > > > And yes I am the sole creator of ATA in Linux, had it not been for me, > > > the subsystem would not be the success it is today. > > > > > > And if something were to happen to me, there are a few trusted souls that > > > have access to the goods. So do not think that all is lost. > > > > > > 63 == 64 if 0 == 1 > > > 255 == 256 if 0 == 1 > > > > > > You forget that zero is a counting number. > > > > > > Cheers, > > > > > > Andre Hedrick > > > The Linux ATA/IDE guy > > > > > > On Wed, 7 Nov 2001, Uncle George wrote: > > > > > > > reading from the web site one gets to understand ( albeit erroneously ) > > > > that the major ide disk folks are supporting the open src of the tech > > > > specs to their hardware. > > > > In reality, it is linux-ide et al that is supporting the disk > > > > manufactures in their quest to sell their products. I would have thought > > > > that 'support' would have been the manufactures release of the necessary > > > > info to build the software ( by anyone in the world ). After all, if u > > > > were to disappear either volentarily, or involentarily, there willo be > > > > no one else that will have the docs to continue on. Oh well. > > > > > > > > There seem to be some mathematical inconsistencies in the kernel > > > > recognoition of the 160gig maxtor drive > > > > > > > > Nov 7 18:43:36 kodak kernel: hde: 320173056 sectors (163929 MB) > > > > w/2048KiB Cache, CHS=317632/255/63, UDMA(133) > > > > > > > > The cardboard box says 160mb ( decimal ) but in reality it is somewhere > > > > near 152gb ( base 2 ). Which also appears to match the 'internal drive > > > > label' as seen by their bios at boottime. > > > > > > > > mke2fs works, but has strange numbers. > > > > > > > > speed is about 33mb/sec, not exactly near 133mb/sec > > > > > > > > later i will take out the drive to read off the real numbers. as the > > > > drive specs are not in the PDF file! and see if the cchhss match in any > > > > fashion > > > > /gat > > > > > > > > BTW i'm also surious as to why 63, and not 64 sectors/trk or ( 256 > > > > tracks/cyl ) is block 0, or track 0 non-existant ? > > > > > > > > [EMAIL PROTECTED] wrote: > > > > > > > > > > GAT, > > > > > > > > > > I do not understand the question? > > > > > I am now self employed and I support this project so I am seeking funds > > > > > to offset development costs. > > > > > > > > > > I do not publish the Documents I have obtained to develop code. > > > > > Those are aquired under various NDA's. > > > > > > > > > > Regards, > > > > > > > > > > Andre Hedrick > > > > > The Linux ATA/IDE guy > > > > > > > > > > On Wed, 7 Nov 2001, Uncle George wrote: > > > > > > > > > > > AND where is the tech programming specs by the organizations that > > > > > > support linux ? > > > > > > /gat > > > > > > > > > > > > should that not be kept somewhere locally ? > > > > > > > > > > > > > > > > > > [EMAIL PROTECTED] wrote: > > > > > > > > > > > > > > Hey GAT, > > > > > > > > > > > > > > Other Web site. > > > > > > > > > > > > > > http://www.linuxdiskcert.org/ > > > > > > > > > > > > > > Cheers, > > > > > > > > > > > > > > Andre Hedrick > > > > > > > The Linux ATA/IDE guy > > > > > > > > > > > > > > On Wed, 7 Nov 2001, Uncle George wrote: > > > > > > > > > > > > > > > in kernel 2.4.14 there is support of the pdc20268(?), but not the > > > > > > > > pdc20269. Nor does there seem to be support for PIO mode 6 ( did i miss > > > > > > > > a number while java'ng ? ) > > > > > > > > html pages says that there is. > > > > > > > > > > > > > > > > There does not seem to be any linux support from promise for this kard >( > > > > > > > > as stated on their web site ), and certainly not in text, src, or > > > > > > > > document format. There also does not seem to be any tech docs on this > > > > > > > > this ( linux-ide.org ) site. > > > > > > > > /gat > > > > > > > > > > > > > > > > > > > > _______________________________________________ Redhat-devel-list mailing list [EMAIL PROTECTED] https://listman.redhat.com/mailman/listinfo/redhat-devel-list
