Bug 6194
| Summary: | ehci_hcd errors & usb device freeze during copying files to the device | ||
|---|---|---|---|
| Product: | Drivers | Reporter: | Jaroslav Sykora (jaroslav.sykora) |
| Component: | USB | Assignee: | David Brownell (dbrownell) |
| Status: | REJECTED UNREPRODUCIBLE | ||
| Severity: | normal | CC: | greg, protasnb, stern, tsg-swdev |
| Priority: | P2 | ||
| Hardware: | i386 | ||
| OS: | Linux | ||
| Kernel Version: | 2.6.16-rc5 | Subsystem: | |
| Regression: | --- | Bisected commit-id: | |
| Bug Depends on: | |||
| Bug Blocks: | 5089 | ||
| Attachments: | dmesg without ehci (unloaded after boot) - the device works fine | ||
|
Description
Jaroslav Sykora
2006-03-08 13:38:14 UTC
david, fun ehci error messages... Here are some comments that might not be relevant. From your log:
SCSI device sda: 509248 2048-byte hdwr sectors (1043 MB)
....
sd 2:0:0:0: SCSI error: return code = 0x50000
end_request: I/O error, dev sda, sector 521532
First of all, is that size correct? It's rather unusual for a flash disk to
claim to have 2048-byte sectors, and it's also unusual (though certainly not
impossible) for one to have a gigabyte of storage. Do you know for a fact
whether the player actually has 1 GB or only 256 MB?
Second, note the sector number where the error occurred. If we assume that the
device's sectors really are only 512 bytes, then the error occurred when the
system tried to read well beyond the last sector. (I think the sector value in
that error message is expressed using 512-byte logical sectors.)
Of course, none of this explains why the device works okay when ehci-hcd isn't
loaded.
The device is Creative Zen Nano Plus 1GB - this beast: http://creative.com/products/product.asp?category=213&subcategory=214&product=12720 It has really 1GB unless they invented some new kind of compression ;-) Created attachment 7547 [details]
dmesg without ehci (unloaded after boot) - the device works fine
It's as if the drive wedges itself in the middle of some request, and then stops responding to a USB reset sequence. I can't think of anything Linux can do to make it respond again. We've certainly seen firmware errors (in USB peripherals) that only show up at high speed. If we had a handle on what we're doing to trigger that error, maybe we could avoid it ... but we have no such handle. I have seen similar problems. One thing that may help is to try to attach the disk to the motherboard directly (to USB ports on the motherboard in the back of the computer) without using any USB cable or USB ports on front panel. I solved my problem by that. The weird thing is that some flash-disks work fine in any slot, some not. Maybe it has something to do with bug 5894? The problem identified in Bug 6722 might be the cause here (is 100MB transfer done after exactly 2 seconds??). 480Mbits/sec (60MB/s) max. USB 2.0 transfer rate would strongly suggest that... Any update on this problem? Thanks. Since no activity in this bug for a while closing it. Please reopen if problem confirmed with recent kernel. |