August 2012 - kernelnewbies@kernelnewbies.org

network dead/stuck for kernel 2.6.32-279.2.1
by Nitin Sharma Aug. 18, 2012

Aug. 18, 2012

I had my eth0 die with this trace. restarting network didn't work, had to reboot the box. looks like a 2.6.32-279.2.1 bug or related to the eth0 driver? Any pointers would be appreciated. OS: CentOS release 6.3 (Final) 08:24:29 kernel: ------------[ cut here ]------------ 08:24:29 kernel: WARNING: at net/sched/sch_generic.c:261 dev_watchdog+0x26d/0x280() (Not tainted) 08:24:29 kernel: Hardware name: X8DT6 08:24:29 kernel: NETDEV WATCHDOG: eth0 (e1000e): transmit queue 0 timed out 08:24:29 kernel: Modules linked in: autofs4 sunrpc cpufreq_ondemand acpi_cpufreq freq_table mperf ipv6 microcode ses enclosure sg serio_raw i2c_i801 i2c_core iTCO_wdt iTCO_vendor_support e1000e ioatdma dca i7core_edac edac_core ext4 mbcache jbd2 raid1 sd_mod crc_t10dif pata_acpi ata_generic ata_piix mpt2sas scsi_transport_sas raid_class dm_mirror dm_region_hash dm_log dm_mod [last unloaded: scsi_wait_scan] 08:24:29 kernel: Pid: 0, comm: swapper Not tainted 2.6.32-279.2.1.el6.x86_64 #1 08:24:29 kernel: Call Trace: 08:24:29 kernel: <IRQ> [<ffffffff8106b747>] ? warn_slowpath_common+0x87/0xc0 08:24:29 kernel: [<ffffffff8106b836>] ? warn_slowpath_fmt+0x46/0x50 08:24:29 kernel: [<ffffffff8145972d>] ? dev_watchdog+0x26d/0x280 08:24:29 kernel: [<ffffffff813a7960>] ? rh_timer_func+0x0/0x10 08:24:29 kernel: [<ffffffff813a7140>] ? usb_hcd_poll_rh_status+0x140/0x180 08:24:29 kernel: [<ffffffff8107dc65>] ? internal_add_timer+0xb5/0x110 08:24:29 kernel: [<ffffffff814594c0>] ? dev_watchdog+0x0/0x280 08:24:29 kernel: [<ffffffff8107e897>] ? run_timer_softirq+0x197/0x340 08:24:29 kernel: [<ffffffff810a21d0>] ? tick_sched_timer+0x0/0xc0 08:24:29 kernel: [<ffffffff8102b40d>] ? lapic_next_event+0x1d/0x30 08:24:29 kernel: [<ffffffff81073ec1>] ? __do_softirq+0xc1/0x1e0 08:24:29 kernel: [<ffffffff81096c50>] ? hrtimer_interrupt+0x140/0x250 08:24:29 kernel: [<ffffffff8100c24c>] ? call_softirq+0x1c/0x30 08:24:29 kernel: [<ffffffff8100de85>] ? do_softirq+0x65/0xa0 08:24:29 kernel: [<ffffffff81073ca5>] ? irq_exit+0x85/0x90 08:24:29 kernel: [<ffffffff81505d10>] ? smp_apic_timer_interrupt+0x70/0x9b 08:24:29 kernel: [<ffffffff8100bc13>] ? apic_timer_interrupt+0x13/0x20 08:24:29 kernel: <EOI> [<ffffffff812cd9fe>] ? intel_idle+0xde/0x170 08:24:29 kernel: [<ffffffff812cd9e1>] ? intel_idle+0xc1/0x170 08:24:29 kernel: [<ffffffff81407757>] ? cpuidle_idle_call+0xa7/0x140 08:24:29 kernel: [<ffffffff81009e06>] ? cpu_idle+0xb6/0x110 08:24:29 kernel: [<ffffffff814e446a>] ? rest_init+0x7a/0x80 08:24:29 kernel: [<ffffffff81c21f7b>] ? start_kernel+0x424/0x430 08:24:29 kernel: [<ffffffff81c2133a>] ? x86_64_start_reservations+0x125/0x129 08:24:29 kernel: [<ffffffff81c21438>] ? x86_64_start_kernel+0xfa/0x109 08:24:29 kernel: ---[ end trace fd2257f6319ad101 ]--- 08:24:29 kernel: e1000e 0000:03:00.0: eth0: Reset adapter 08:24:29 kernel: e1000e 0000:03:00.0: eth0: Error reading PHY register 08:24:29 kernel: e1000e: eth0 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: Rx/Tx 08:24:30 abrt-dump-oops: Reported 1 kernel oopses to Abrt 08:24:30 abrtd: Directory 'oops-2012-08-12-08:24:30-6095-0' creation detected 08:24:30 abrtd: Can't open file '/var/spool/abrt/oops-2012-08-12-08:24:30-6095-0/uid': No such file or directory 08:24:37 kernel: Bridge firewalling registered 08:24:39 kernel: e1000e 0000:03:00.0: eth0: Reset adapter 08:24:39 kernel: e1000e 0000:03:00.0: eth0: Error reading PHY register 08:24:39 kernel: e1000e: eth0 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: Rx/Tx 08:24:49 kernel: e1000e 0000:03:00.0: eth0: Reset adapter 08:24:49 kernel: e1000e 0000:03:00.0: eth0: Error reading PHY register 08:24:49 kernel: e1000e: eth0 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: Rx/Tx 08:24:51 abrtd: Sending an email... 08:24:52 abrtd: Email was sent to: root@localhost 08:24:52 abrtd: Duplicate: UUID 08:24:52 abrtd: DUP_OF_DIR: /var/spool/abrt/oops-2012-08-03-09:06:36-13462-0 08:24:52 abrtd: New problem directory /var/spool/abrt/oops-2012-08-12-08:24:30-6095-0, processing 08:24:52 abrtd: Can't open file '/var/spool/abrt/oops-2012-08-12-08:24:30-6095-0/uid': No such file or directory 08:24:59 kernel: e1000e 0000:03:00.0: eth0: Reset adapter 08:24:59 kernel: e1000e 0000:03:00.0: eth0: Error reading PHY register 08:24:59 kernel: e1000e: eth0 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: Rx/Tx 08:25:49 kernel: e1000e 0000:03:00.0: eth0: Reset adapter 08:25:49 kernel: e1000e 0000:03:00.0: eth0: Error reading PHY register Thanks Nitin

2 3

Stackable file systems and NFS
by Ranjan Sinha Aug. 17, 2012

Aug. 17, 2012

Hi, We are writing a stackable file system. While testing this on NFS we came across a situation where sometimes we are getting ASCII NUL bytes in the buffer read from the lower file system (NFS in this case). This happens when NFS client 1 is writing to a file and NFS client 2 is continuously reading from the same file. A simple case like the following also shows the same problem ----------------8<------------------8<----------------------------- while (1) { if (vfs_getattr(filp->f_vfsmnt, filp->f_dentry, &stat) != 0) { printk(KERN_INFO "Error in getting attribute\n"); break; } if (no_inc == 60) { printk(KERN_INFO "File not increased for 1 min. Quitting\n"); break; } inc = stat.size - old_size; if (inc) { old_size = stat.size; no_inc = 0; count = vfs_read(filp, buf, sizeof(buf), &filp->f_pos); if (count < 0) { printk(KERN_INFO "Error in reading file\n"); goto out; } printk(KERN_INFO "File increase = [%llu], total read = [%llu]" " file_len = [%llu] File size from inode = [%llu]\n", inc, count, stat.size, i_size_read(filp->f_dentry->d_inode)); printk(KERN_INFO "----------------------------------------------------\n"); debug_dump("Read buffer", buf, total_count); } else { no_inc++; } } ----------------8<------------------8<----------------------------- We have also noticed that the expected increase (inc) and the size returned in (vfs_read()) is different. A similar program in user space also reproduces this issue. A google search for this issue tells me that others have seen this issue as well. http://stackoverflow.com/questions/6814404/java-inputstream-read-methods-re… Is this a known and expected behaviour of NFS? PS : This is a very old kernel (2.6.18.274.) We still have to support RHEL 5 :) -- Regards, Ranjan

3 11

Debug linker with GDB
by Geraint Yang Aug. 16, 2012

Aug. 16, 2012

Hello everyone. I am debugging the dynamic library linker in kernel with gdb, when the kernel is going to execute a program which uses *.so, the linker will be loaded and load other required libraries before the program is executed, because the load address of ld(linker) is 0x00000000, the kernel will load the linker in a valid address, for example 0x6fee5000, gdb fail to resolve the symbol table. I wonder if there is any method to fix the base address, so gdb can resolve the symbol table correctly. Thanks, Yang

2 1

Android jelly bean mounting problem
by sandeep kumar Aug. 16, 2012

Aug. 16, 2012

Hi all, We are facing one issue, 3>EXT4-fs (mmcblk0p12): VFS: Can't find ext4 filesystem [ 5.943134] EXT4-fs (mmcblk0p12): VFS: Can't find ext4 filesystem <3>fs_mgr: Cannot mount filesystem on /dev/block/platform/msm_sdcc.1/by-num/p12 at /system [ 5.992430] fs_mgr: Cannot mount filesystem on /dev/block/platform/msm_sdcc.1/by-num/p12 at /system <3>init: fs_mgr_mount_all returned an error [ 6.005364] init: fs_mgr_mount_all returned an error <3>EXT4-fs (mmcblk0p15): VFS: Can't find ext4 filesystem [ 6.015346] EXT4-fs (mmcblk0p15): VFS: Can't find ext4 filesystem <3>EXT4-fs (mmcblk0p14): VFS: Can't find ext4 filesystem [ 6.110148] EXT4-fs (mmcblk0p14): VFS: Can't find ext4 filesystem <3>init: cannot find '/system/bin/sh', disabling 'qcom-post-fs' [ 6.230449] init: cannot find '/system/bin/sh', disabling 'qcom-post-fs' Can you help me.. With regards, Sandeep Kumar Anantapalli,

3 3

Adding the cp210x module to the NAO kernel
by Ankur Sinha Aug. 16, 2012

Aug. 16, 2012

Hi folks, I'm using the NAO[1] for my research. It uses a linux 2.6.29.6-rt24 kernel[2] which they seem to have customized to their needs. I need to use a NavChip[3] IMU(inertial measurement unit) with the NAO. The NavChip works out of the box with my Fedora 17 kernel(3.5.1-1.fc17.x86_64). This is what dmesg says when I plug it in: > [71406.521154] usb 2-1.5: new full-speed USB device number 6 using ehci_hcd > [71406.607360] usb 2-1.5: New USB device found, idVendor=10c4, idProduct=ea60 > [71406.607367] usb 2-1.5: New USB device strings: Mfr=1, Product=2, SerialNumber=3 > [71406.607372] usb 2-1.5: Product: CP2102 USB to UART Bridge Controller > [71406.607376] usb 2-1.5: Manufacturer: Silicon Labs > [71406.607379] usb 2-1.5: SerialNumber: 0001 > [71406.772438] usbcore: registered new interface driver cp210x > [71406.772468] USB Serial support registered for cp210x > [71406.772513] cp210x 2-1.5:1.0: cp210x converter detected > [71406.845220] usb 2-1.5: reset full-speed USB device number 6 using ehci_hcd > [71406.931690] usb 2-1.5: cp210x converter now attached to ttyUSB0 > [ankur@ankur 2.6.29.6-rt24-aldebaran-rt]$ However, when I plug it into the NAO, it isn't detected. I looked up, and the kernel they use doesn't seem to have the cp210x module that is required for the NavChip (I think?). The NavChip's processor is an AMD Geode[1] which is x86 compatible according to wiki. Can you please tell me how I can add the required module to the NAO's kernel? I already tried something, but this is my first brush with kernel compilation, so it was probably wrong: I built the cp210x module using the NAO kernel tree: > [ankur@ankur aldebaran-linux-aldebaran-66f8749]$ pwd > /home/ankur/Documents/work/code/NAO/OS/aldebaran-linux-aldebaran-66f8749 > [ankur@ankur aldebaran-linux-aldebaran-66f8749]$ file drivers/usb/serial/cp210x.ko > drivers/usb/serial/cp210x.ko: ELF 32-bit LSB relocatable, Intel 80386, version 1 (SYSV), BuildID[sha1]=0x61274b6a395ca0bbb45729b87283f361c04b396f, not stripped > [ankur@ankur aldebaran-linux-aldebaran-66f8749]$ > I placed this into the drivers/usb/serial directory in the NAO system, but it didn't work (no surprise there!). It gave me an error: "Invalid module format" when I tried modprobe or insmod. The /var/log/messages had this: "no version for struct_module found: kernel tainted". Any hints please? It's probably something simple, just not simple enough for a kernel newbie :) This is the set of modules they ship their kernel with: > [ankur@ankur 2.6.29.6-rt24-aldebaran-rt]$ pwd > /run/media/ankur/OpenNao-system/lib/modules/2.6.29.6-rt24-aldebaran-rt > [ankur@ankur 2.6.29.6-rt24-aldebaran-rt]$ ll > total 409 > lrwxrwxrwx. 1 root root 93 Jun 4 01:23 build -> /opennao-geode/tmp/portage/ald-kernel/opennao-kernel-2.6.29-r41/work/linux-2.6.29-opennao-r41 > drwxr-xr-x. 5 root root 1024 Jun 4 01:23 kernel > -rw-r--r--. 1 root root 59429 Aug 14 11:01 modules.alias > -rw-r--r--. 1 root root 49998 Aug 14 11:01 modules.alias.bin > -rw-r--r--. 1 root root 69 Aug 14 11:01 modules.ccwmap > -rw-r--r--. 1 root root 11117 Aug 14 11:01 modules.dep > -rw-r--r--. 1 root root 13279 Aug 14 11:01 modules.dep.bin > -rw-r--r--. 1 root root 73 Aug 14 11:01 modules.ieee1394map > -rw-r--r--. 1 root root 141 Aug 14 11:01 modules.inputmap > -rw-r--r--. 1 root root 81 Aug 14 11:01 modules.isapnpmap > -rw-r--r--. 1 root root 74 Aug 14 11:01 modules.ofmap > -rw-r--r--. 1 root root 937 Jun 3 17:25 modules.order > -rw-r--r--. 1 root root 2829 Aug 14 11:01 modules.pcimap > -rw-r--r--. 1 root root 85 Aug 14 11:01 modules.seriomap > -rw-r--r--. 1 root root 25281 Aug 14 11:01 modules.symbols > -rw-r--r--. 1 root root 36981 Aug 14 11:01 modules.symbols.bin > -rw-r--r--. 1 root root 199651 Aug 14 11:01 modules.usbmap > lrwxrwxrwx. 1 root root 93 Jun 4 01:23 source -> /opennao-geode/tmp/portage/ald-kernel/opennao-kernel-2.6.29-r41/work/linux-2.6.29-opennao-r41 > drwxr-xr-x. 5 root root 1024 Jun 4 01:23 updates > [ankur@ankur 2.6.29.6-rt24-aldebaran-rt]$ find . -name "*.ko" -not -name "cp210x.ko" -print > ./updates/compat/compat.ko > ./updates/compat/compat_firmware_class.ko > ./updates/net/wireless/cfg80211.ko > ./updates/net/mac80211/mac80211.ko > ./updates/net/rfkill/rfkill_backport.ko > ./updates/drivers/misc/eeprom/eeprom_93cx6.ko > ./updates/drivers/net/wireless/rt2x00/rt2x00pci.ko > ./updates/drivers/net/wireless/rt2x00/rt2800lib.ko > ./updates/drivers/net/wireless/rt2x00/rt2400pci.ko > ./updates/drivers/net/wireless/rt2x00/rt61pci.ko > ./updates/drivers/net/wireless/rt2x00/rt2x00usb.ko > ./updates/drivers/net/wireless/rt2x00/rt2800pci.ko > ./updates/drivers/net/wireless/rt2x00/rt2x00lib.ko > ./updates/drivers/net/wireless/rt2x00/rt2500usb.ko > ./updates/drivers/net/wireless/rt2x00/rt2800usb.ko > ./updates/drivers/net/wireless/rt2x00/rt73usb.ko > ./updates/drivers/net/wireless/rt2x00/rt2500pci.ko > ./updates/drivers/net/wireless/zd1211rw/zd1211rw.ko > ./kernel/net/wireless/cfg80211.ko > ./kernel/net/mac80211/mac80211.ko > ./kernel/net/bluetooth/sco.ko > ./kernel/net/bluetooth/hidp/hidp.ko > ./kernel/net/bluetooth/l2cap.ko > ./kernel/net/bluetooth/bnep/bnep.ko > ./kernel/net/bluetooth/rfcomm/rfcomm.ko > ./kernel/lib/cimarron/cimarron.ko > ./kernel/lib/crc-itu-t.ko > ./kernel/drivers/staging/rt2870/rt3070sta.ko > ./kernel/drivers/misc/eeprom/eeprom_93cx6.ko > ./kernel/drivers/usb/class/cdc-acm.ko > ./kernel/drivers/usb/serial/usbserial.ko > ./kernel/drivers/usb/serial/ftdi_sio.ko > ./kernel/drivers/hid/hid.ko > ./kernel/drivers/net/wireless/zd1211rw/zd1211rw.ko > ./kernel/drivers/net/wireless/rtl818x/rtl8187.ko > ./kernel/drivers/hwmon/lm90.ko > ./kernel/drivers/scsi/scsi_wait_scan.ko > ./kernel/drivers/media/video/lxv4l2/lxv4l2.ko > ./kernel/drivers/i2c/busses/scx200_acb.ko > ./kernel/drivers/i2c/busses/i2c-serial.ko > ./kernel/drivers/bluetooth/btusb.ko > ./kernel/drivers/bluetooth/bfusb.ko > ./kernel/drivers/bluetooth/bcm203x.ko > ./kernel/drivers/bluetooth/hci_vhci.ko > [ankur@ankur 2.6.29.6-rt24-aldebaran-rt]$ Please let me know if you need any more info, and I'll provide it. The kernel goes on a USB flash drive that needs to be inserted into the NAO to make it boot. They provide a utility that formats the flash drive as such: > [ankur@ankur ~]$ cd /run/media/ankur/ > [ankur@ankur ankur]$ ll > total 5 > drwxr-xr-x. 7 root root 4096 Aug 6 22:28 OpenNao-data > drwxr-xr-x. 17 root root 1024 Aug 14 10:50 OpenNao-system > [ankur@ankur ankur]$ mount | egrep OpenNao > /dev/sdb1 on /run/media/ankur/OpenNao-system type ext3 (rw,nosuid,nodev,relatime,seclabel,data=ordered,uhelper=udisks2) > /dev/sdb2 on /run/media/ankur/OpenNao-data type ext3 (rw,nosuid,nodev,relatime,seclabel,data=ordered,uhelper=udisks2) > [ankur@ankur ankur]$ > [1] http://en.wikipedia.org/wiki/Nao_(robot) [2] https://github.com/aldebaran/linux-aldebaran/ [3] http://www.intersense.com/pages/16/16/ [4] http://en.wikipedia.org/wiki/Geode_(processor) -- Thanks, Warm regards, Ankur: "FranciscoD" Please only print if necessary. Looking to contribute to Fedora? Look here: https://fedoraproject.org/wiki/Fedora_Join_SIG http://fedoraproject.org/wiki/User:Ankursinha http://dodoincfedora.wordpress.com/

2 5

invoking zcache w/o using a workload
by Asim Aug. 16, 2012

Aug. 16, 2012

Hi, I was playing around with zcache and fiddling with different compression algorithms. Furthermore, I was integrating some checksum algorithms for cleancache for my own project goals. Unfortunately, I need to run a workload to trigger zcache. Is there a way I can trigger zcache_get/put (specifically clean cache) with a simple unit test (as opposed to running a workload) ? This will greatly reduce my test/build time. I was wondering if there is an existing unit test-suite. Or I could write a simple kernel module to trigger put/get but I was wondering how to go about invoking these operations. My guess is calling cleancache_get/free using a page off (get_free_page) will not invoke zcache. Any suggestions will be useful. Thanks, Asim

2 5

Understanding and using socket RTT
by Vikash Kumar Aug. 14, 2012

Aug. 14, 2012

Hi, I am writing an application in kernel which needs to make decisions based on RTT of the socket open with another host. So, I was looking at TCP code to find out how can I get that information from it. Whatever little I have understood from the code, it looks like tcp_sock->srtt (smoothed round trip time) is the one I can more or less trust to be the current rtt of the socket and that this value is in jiffies. However I got confused due to the comment on its declaration. u32 srtt; /* smoothed round trip time << 3 */ Questions I have 1) Does this mean that the value stored in this variable is already left shifted by 3 or it should be left shifted by 3 before using it ? Or the value is simply the computed rtt in jiffies and it has no relation with the left shift operation ? 2) Is the SRTT correct value to look at for rtt or should another metric / value be used for this? Regards, Vikash Kumar http://VikashKumar.com/

1 0

Re:
by Vivek Panwar Aug. 13, 2012

Aug. 13, 2012

Thanks a lot . On Mon, Aug 13, 2012 at 4:31 PM, Fan Yang <lljyangfan(a)gmail.com> wrote: > > > 2012/8/13 Vivek Panwar <vivekpanwar2008(a)gmail.com> > >> HI, >> >> How KGDB works internally or how a gdb stub works internally with source >> code for x86_64 arch. Can some one please suggest me the good docs for the >> same.I am very new to this so want know whole detailed internally working >> procedure for KGDB. >> >> Thanks >> Vivek >> >> _______________________________________________ >> Kernelnewbies mailing list >> Kernelnewbies(a)kernelnewbies.org >> http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies >> >> >

2 1

(no subject)
by Vivek Panwar Aug. 13, 2012

Aug. 13, 2012

HI, How KGDB works internally or how a gdb stub works internally with source code for x86_64 arch. Can some one please suggest me the good docs for the same.I am very new to this so want know whole detailed internally working procedure for KGDB. Thanks Vivek

2 1

Re: How to allocate hugepage in kernel module?
by aurum spark Aug. 12, 2012

Aug. 12, 2012

Hi J.Hwan Kim, Before we talk about APIs let's understand few things about requirement 1. How much contiguous memory do you require for you module ? 2. Should it be DMAble ? 3. Are you going to use that as shared memory with any other card or processor ? So there are certain API which can provide you huge chunk of contiguous memory. 1. One of them as you are using is get_free_pages(). Try suggestions by others to get rid of error. 2. Another thing you can use is dma_alloc_coherent(). But there is limit on this and you should consider that other modules will also required DMAble memory. 3. Depending on the kernel version you are using you can do bootmem_reserve() API during initialization code of kernel and you can keep this memory reserved for your module. Pass that as platform data. 4. You can try passing say mem=768M parameter through kernel commandline(assuming you have 1G of memory). That will force kernel to use 768M and you can use rest of the 256M memory by doing ioremap in your module. Regards, ~/Aurum

1 0