Bug 511 - XFree86-xfs-3.3.3-1: xfs coredumps at first request.
Summary: XFree86-xfs-3.3.3-1: xfs coredumps at first request.
Keywords:
Status: CLOSED CURRENTRELEASE
Alias: None
Product: Red Hat Linux
Classification: Retired
Component: XFree86
Version: 5.1
Hardware: i386
OS: Linux
high
high
Target Milestone: ---
Assignee: Preston Brown
QA Contact:
URL:
Whiteboard:
: 849 908 (view as bug list)
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 1998-12-17 22:34 UTC by Aleksey Nogin
Modified: 2008-05-01 15:37 UTC (History)
1 user (show)

Fixed In Version:
Doc Type: Bug Fix
Doc Text:
Clone Of:
Environment:
Last Closed: 1999-01-18 22:02:52 UTC
Embargoed:


Attachments (Terms of Use)

Description Aleksey Nogin 1998-12-17 22:34:34 UTC
Hi,

I tried to upgrade XFree86 with the latest updates from
updates.redhat.com and I have a problem - xfs coredumps as
soon as X server tries to access it. When I downgraded back
to 3.3.2.3-18 (leaving 3.3.3-1 libs and server), everything
returned back to normal.

My configuration: RH 5.1 + updates, glibc-2.0.7-19, Kernel
2.1.130 (SMP).

Here is the tail of the strace output (showing reading of
the last two font directories on startup and everything
after that):

open("/usr/X11R6/lib/X11/fonts/cyrillic/100dpi/fonts.dir",
O_RDONLY) = 5
fstat(5, {st_mode=0, st_size=0, ...})   = 0
fstat(5, {st_mode=0, st_size=0, ...})   = 0
mmap(0, 4096, PROT_READ|PROT_WRITE,
MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x400cb000
read(5, "60\ncrox6t.pcf.gz -cronyx-times-"..., 4096) = 4096
brk(0x810d000)                          = 0x810d000
brk(0x810e000)                          = 0x810e000
brk(0x810f000)                          = 0x810f000
brk(0x8110000)                          = 0x8110000
read(5, "gz -cronyx-times-bold-r-normal--"..., 4096) = 351
brk(0x8111000)                          = 0x8111000
read(5, "", 4096)                       = 0
read(5, "", 4096)                       = 0
close(5)                                = 0
munmap(0x400cb000, 4096)                = 0
open("/usr/X11R6/lib/X11/fonts/cyrillic/100dpi/fonts.alias",
O_RDONLY) = -1 ENOENT (No such file or directory)
brk(0x8113000)                          = 0x8113000
open("/usr/u/nuprl/fonts/bdf/fonts.dir", O_RDONLY) = 5
fstat(5, {st_mode=0, st_size=0, ...})   = 0
fstat(5, {st_mode=0, st_size=0, ...})   = 0
mmap(0, 4096, PROT_READ|PROT_WRITE,
MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x400cb000
read(5, "5\nnuprl-20.pcf nuprl-20\nnuprl-"..., 4096) = 100
read(5, "", 4096)                       = 0
read(5, "", 4096)                       = 0
close(5)                                = 0
munmap(0x400cb000, 4096)                = 0
open("/usr/u/nuprl/fonts/bdf/fonts.alias", O_RDONLY) = -1
ENOENT (No such file or directory)
open("/var/log/xfs-errors.log", O_WRONLY|O_CREAT|O_TRUNC,
0666) = 5
dup2(5, 2)                              = 2
close(5)                                = 0
gettimeofday({913930856, 761814}, NULL) = 0
select(128, [3 4], NULL, NULL, {599, 740000} <unfinished
...>
--- SIGCONT (Continued) ---
<... select resumed> )                  = -1 ENOSYS
(Function not implemented)
gettimeofday({913930886, 168388}, NULL) = 0
accept(3, {sin_family=AF_INET, sin_port=htons(3227),
sin_addr=inet_addr("128.84.248.55")}, [16]) = 5
setsockopt(5, IPPROTO_TCP1, [1], 4)     = 0
getsockname(5, {sin_family=AF_INET, sin_port=htons(7100),
sin_addr=inet_addr("128.84.248.55")}, [16]) = 0
getpeername(5, {sin_family=AF_INET, sin_port=htons(3227),
sin_addr=inet_addr("128.84.248.55")}, [16]) = 0
fcntl(5, F_SETFL, O_RDONLY|O_NONBLOCK)  = 0
gettimeofday({913930886, 169834}, NULL) = 0
gettimeofday({913930886, 170023}, NULL) = 0
select(128, [3 4 5], NULL, NULL, {570, 331000}) = 1 (in [5],
left {570, 340000})
gettimeofday({913930886, 170791}, NULL) = 0
--- SIGSEGV (Segmentation fault) ---
+++ killed by SIGSEGV +++

Comment 1 David Lawrence 1998-12-17 22:41:59 UTC
This has been assigned to a developer for further review.

Comment 2 Aleksey Nogin 1998-12-17 22:45:59 UTC
I forgot to add: bug is 100% reproducable. I run startx (with xfs
being the only item in my FontPath), it tries to connect to xfs and
resolve "fixed" and xfs immediatelly crashes...

Comment 3 Preston Brown 1998-12-17 23:15:59 UTC
Please make sure that the path in your xfs.conf is correct.  If there
are any directories listed in that file that are not present, xfs
coredumps.

Comment 4 Aleksey Nogin 1998-12-17 23:27:59 UTC
When I had unexisting directories in the config file, xfs was politely
telling me about that (in xfs-errors) and exiting, not dumping core. I
just double-checked - all directories do exist. If you want, I can
send you the rest of the strace log where xfs succesfully reads all
font.dir files.

Comment 5 Patrick J. LoPresti 1998-12-18 21:22:59 UTC
I can confirm that this bug exists.  The fs/config file shipped with
XFree86-3.3.3 refers to directories which do not exist, but xfs merely
complains politely about those.
When you fix the path in fs/config, xfs crashes upon receiving its
first request.  Downgrading to 3.3.2.3 fixes the problem.

Comment 6 Jeff Johnson 1999-01-17 21:07:59 UTC
*** Bug 511 has been marked as a duplicate of this bug. ***

Running RedHat 5.1, upgraded to the December XFree86
updates including XFree86-xfs-3.3.3-1.i386.rpm.  Now if
I try to use xfs, it segfaults as X tries to start up,
making X die shortly afterward.  strace on xfs revealed
nothing useful.  Since I'm running kernel 2.2.0pre7, I
verified that it also had problems on 2.1.130, which it
does.  I'm using the SVGA server with an S3 Virge on a
Cyrix 6x86/166+.  To work around the problem, I have
stopped using the font server, but I'd like to be able
to use it.



------- Additional Comments From ayn2  01/16/99 21:55 -------
This is a duplicate of #511

Comment 7 Preston Brown 1999-01-18 22:02:59 UTC
this is fixed in a forthcoming errata release of XFree86 3.3.3.1.

Comment 8 Jeff Johnson 1999-01-21 16:12:59 UTC
*** Bug 908 has been marked as a duplicate of this bug. ***

New font server from XFree86-xfs-3.3.3-1 crashes and
coredumps when connection to it is attempted.


Note You need to log in before you can comment on or make changes to this bug.