Bug 28523 - llseek has inconsistent semantics between smp and enterprise 2.2.16 kernels
Summary: llseek has inconsistent semantics between smp and enterprise 2.2.16 kernels
Keywords:
Status: CLOSED NOTABUG
Alias: None
Product: Red Hat Linux
Classification: Retired
Component: kernel
Version: 7.0
Hardware: i386
OS: Linux
medium
high
Target Milestone: ---
Assignee: Michael K. Johnson
QA Contact: Brock Organ
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2001-02-21 00:09 UTC by Georg Nikodym
Modified: 2007-04-18 16:31 UTC (History)
1 user (show)

Fixed In Version:
Doc Type: Bug Fix
Doc Text:
Clone Of:
Environment:
Last Closed: 2001-02-21 00:09:20 UTC
Embargoed:

Attachments (Terms of Use)

Description Georg Nikodym 2001-02-21 00:09:17 UTC 
llseek has inconsistent semantics between smp and enterprise 2.2.16 kernels

Running /usr/bin/nm on ARM cross-compiled binaries works fine on the
2.2.16-smp kernel (or any other kernel) and a fully updated RH7.0 system.

Using the -enterprise kernel, nm fails with:

/usr/bin/nm: conftest.o: Value too large for defined data type

Using strace, I see that it's because _llseek() returned EOVERFLOW.  Fair
enough.

On the other kernels, this same _llseek() returned EINVAL and nm itself
went on to
complete successfully.

While it may be argued that nm has a bug, the semantics of a system call
probably shouldn't
be variable...
Comment 1 Michael K. Johnson 2001-02-21 18:53:31 UTC 
The semantics are more strictly specified by the LFS standard, which
is implemented only in the enterprise kernel.
Note You need to log in before you can comment on or make changes to this bug.