IOATDMA: driver oopses on request_irq()

Previous thread: [mm patch] i915: fix invalid opcode exception on cpus without clflush by Kyle McMartin on Thursday, January 17, 2008 - 7:52 pm. (4 messages)

Next thread: Merge Candidate tree for the 2.6 kernel by James Bottomley on Thursday, January 17, 2008 - 8:03 pm. (1 message)
To: <shannon.nelson@...>
Cc: linux-kernel <linux-kernel@...>
Date: Thursday, January 17, 2008 - 8:00 pm

https://bugzilla.redhat.com/show_bug.cgi?id=311991
[still present in 2.6.23.8]

[Drivers really need to actually be ready to receive interrupts when
they call request_irq(). We keep finding drivers that aren't because
we enable CONFIG_DEBUG_SHIRQ]

kernel version: 2.6.22.7-85.fc7debug
modprobe ioatdma during bootup sez:

[ 142.497353] Module ioatdma cannot be unloaded due to unsafe usage in
drivers/dma/ioatdma.c:829
[ 142.497906] ACPI: PCI Interrupt 0000:00:08.0[A] -> GSI 16 (level, low) -> IRQ 16
[ 142.497945] Unable to handle kernel NULL pointer dereference at
0000000000000003 RIP:
[ 142.498002] [<ffffffff8835adc0>] :ioatdma:ioat_do_interrupt+0xd/0x50
[ 142.498770] PGD 10fc46067 PUD 114a56067 PMD 0
[ 142.499156] Oops: 0000 [1] SMP
[ 142.499498] last sysfs file: /block/sr0/removable
[ 142.499753] CPU 3
[ 142.500049] Modules linked in: ioatdma dell_rbu nfsd exportfs lockd nfs_acl
ipmi_devintf ipmi_si ipmi_msghandler hidp l2cap bluetooth sunr
pc xt_length ipt_TOS xt_state xt_tcpudp ipt_REJECT ipt_LOG xt_limit
iptable_mangle iptable_nat nf_nat nf_conntrack_ipv4 nf_conntrack nfnetlin
k iptable_filter ip_tables x_tables ppp_deflate zlib_deflate ppp_synctty
ppp_async crc_ccitt ppp_generic slhc bridge dm_mirror dm_multipath d
m_mod video sbs button dock battery ac tcp_westwood kvm_intel kvm sr_mod cdrom
ata_generic ata_piix serio_raw libata bnx2 rtc_cmos sg joydev
usb_storage shpchp megaraid_sas sd_mod scsi_mod ext3 jbd mbcache ehci_hcd
ohci_hcd uhci_hcd
[ 142.505129] Pid: 5239, comm: modprobe Not tainted 2.6.22.7-85.fc7debug #1
[ 142.505892] RSP: 0018:ffff81010ff83c48 EFLAGS: 00010096
[ 142.506146] RAX: 0000000000001214 RBX: 0000000000000202 RCX: ffffffff8835b0f5
[ 142.506405] RDX: 0000000000000000 RSI: ffff81012cc11700 RDI: 00000000000008f4
[ 142.506663] RBP: ffff81012cc11700 R08: ffff81012cc11700 R09: ffff81012cc11700
[ 142.506919] R10: 000000000000005a R11: ffffffff811f2016 R12: 0000000000000020
[ 142.507176] R13: ffff81012cc11700 R14: ffffffff8...

To: Chuck Ebbert <cebbert@...>
Cc: linux-kernel <linux-kernel@...>
Date: Thursday, January 17, 2008 - 8:11 pm

Chuck - We did some significant rework on the ioatdma driver last
summer. Unfortunately it didn't make it into 2.6.23, but it is
available in 2.6.24 as well as out-of-core on our SourceForge site.
Please try the newer code - I believe it will fix your problem.

Thanks,
sln
======================================================================
Mr. Shannon Nelson LAN Access Division, Intel Corp.
Shannon.Nelson@intel.com I don't speak for Intel
(503) 712-7659 Parents can't afford to be squeamish.
--

Previous thread: [mm patch] i915: fix invalid opcode exception on cpus without clflush by Kyle McMartin on Thursday, January 17, 2008 - 7:52 pm. (4 messages)

Next thread: Merge Candidate tree for the 2.6 kernel by James Bottomley on Thursday, January 17, 2008 - 8:03 pm. (1 message)