Bug#789941: Performance regression: slow sequential reads for some block devices (readpage vs readpages) - patched in 3.18+
Package: linux-image-3.16.0-4-amd64
Version: 3.16.7-ckt11-1
Initially discovered inside a QEMU guest, by doing the following:
# hdparm -t /dev/vda
Under the Wheezy 3.2 kernel:
Timing buffered disk reads: 384 MB in 3.01 seconds = 127.50 MB/sec
Under the Jessie 3.16 kernel:
Timing buffered disk reads: 46 MB in 3.07 seconds = 14.97 MB/sec
After some work swapping kernels, I discovered that this behaviour
exists in the 3.16 and 3.17 kernels, but not in 3.2-3.15 or 3.18+
Watching iostat as the I/O is happening indicated that the 3.16/3.17
guests were performing the I/O in 8-sector (512 bytes per sector)
chunks; in the other kernels, the request sizes were 254 sectors instead.
Some further work identified this patch in 3.18:
https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/diff/?id=447f05bb488bff4282088259b04f47f0f9f76760
diff --git a/fs/block_dev.c b/fs/block_dev.c
index 6d72746..e2f3ad08 100644
--- a/fs/block_dev.c
+++ b/fs/block_dev.c
@@ -304,6 +304,12 @@ static int blkdev_readpage(struct file * file,
struct page * page)
return block_read_full_page(page, blkdev_get_block);
}
+static int blkdev_readpages(struct file *file, struct address_space
*mapping,
+ struct list_head *pages, unsigned nr_pages)
+{
+ return mpage_readpages(mapping, pages, nr_pages, blkdev_get_block);
+}
+
static int blkdev_write_begin(struct file *file, struct address_space
*mapping,
loff_t pos, unsigned len, unsigned flags,
struct page **pagep, void **fsdata)
@@ -1622,6 +1628,7 @@ static int blkdev_releasepage(struct page *page,
gfp_t wait)
static const struct address_space_operations def_blk_aops = {
.readpage = blkdev_readpage,
+ .readpages = blkdev_readpages,
.writepage = blkdev_writepage,
.write_begin= blkdev_write_begin,
.write_end = blkdev_write_end,
It applies cleanly to 3.16 and 3.17; with the patch applied, the
larger request sizes are again seen and performance returns to the
previous level.
Could we apply this to the 3.16 kernel in Jessie?
--
To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/558bfe84.5010...@bytemark.co.uk
Bug#789952: linux-image-4.0.0-2-amd64: kernel 4.0.0-2 with iwlwifi 7260 reports incorrect wifi bit rate
Package: src:linux Version: 4.0.5-1 Severity: normal When using the kernel provided by linux-image-4.0.0-2-amd64, the WiFi bit rate is misreported. The exact (incorrect) bit rate changes on each boot (possibly on the first WiFi connection), but is constant until the computer is rebooted, regardless of WiFi disconnects or reconnects. The issue does not occur when using linux-image-3.16.0-4-amd64. - $ uname -a Linux iset 4.0.0-2-amd64 #1 SMP Debian 4.0.5-1 (2015-06-16) x86_64 GNU/Linux $ dpkg -l | grep firmware-iwl ii firmware-iwlwifi0.44 $ dmesg | grep -i iwl [2.332168] iwlwifi :03:00.0: firmware: failed to load iwlwifi-7260-12.ucode (-2) [2.332212] iwlwifi :03:00.0: Direct firmware load for iwlwifi-7260-12.ucode failed with error -2 [2.332703] iwlwifi :03:00.0: firmware: failed to load iwlwifi-7260-11.ucode (-2) [2.332744] iwlwifi :03:00.0: Direct firmware load for iwlwifi-7260-11.ucode failed with error -2 [2.343878] iwlwifi :03:00.0: firmware: direct-loading firmware iwlwifi-7260-10.ucode [2.344037] iwlwifi :03:00.0: loaded firmware version 23.15.10.0 op_mode iwlmvm [2.389263] iwlwifi :03:00.0: Detected Intel(R) Dual Band Wireless AC 7260, REV=0x144 [2.389337] iwlwifi :03:00.0: L1 Enabled - LTR Enabled [2.389597] iwlwifi :03:00.0: L1 Enabled - LTR Enabled [2.590343] ieee80211 phy0: Selected rate control algorithm 'iwl-mvm-rs' [4.734368] iwlwifi :03:00.0: L1 Enabled - LTR Enabled [4.734629] iwlwifi :03:00.0: L1 Enabled - LTR Enabled $ /sbin/iwconfig eth0 no wireless extensions. wlan0 IEEE 802.11abgn ESSID:MYWIFI Mode:Managed Frequency:5.745 GHz Access Point: 40:16:7E:XX:XX:XX Bit Rate=6 Mb/s Tx-Power=22 dBm Retry short limit:7 RTS thr:off Fragment thr:off Power Management:on Link Quality=60/70 Signal level=-50 dBm Rx invalid nwid:0 Rx invalid crypt:0 Rx invalid frag:0 Tx excessive retries:0 Invalid misc:32 Missed beacon:0 docker0 no wireless extensions. lono wireless extensions. - I copied a 3GB file to/from my network share and repeatedly ran `iwconfig` to sample the reported bit rate. While my download was proceeding at 37 MBps, the reported bit rate was constant at 6 Mbps. I obtained iwlwifi-7260-12.ucode from the linux-firmware git repo master branch, copied it to /lib/firmware, rebooted, and reran the test. The results were the same. I rebooted into kernel 3.16 (which loads iwlwifi-7260-9.ucode despite the fact that iwlwifi-7260-10.ucode and iwlwifi-7260-12.ucode were available in /lib/firmware), the bit rate is correctly reported and fluctuated between about 500 to 700 Mbps. It is important to note that the actual download performance was not affected and was more or less constant throughout my tests. Only the reported bit rate was incorrect. Nevertheless, an incorrect bit rate will confuse people and may cause them to mistakenly believe that their WiFi routers are not working or configured properly. -- Package-specific info: ** Version: Linux version 4.0.0-2-amd64 (debian-kernel@lists.debian.org) (gcc version 4.9.2 (Debian 4.9.2-21) ) #1 SMP Debian 4.0.5-1 (2015-06-16) ** Command line: BOOT_IMAGE=/boot/vmlinuz-4.0.0-2-amd64 root=UUID=1fe17317-4e03-4bc1-8463-015ed156427e ro quiet splash acpi_osi=Linux ** Tainted: O (4096) * Out-of-tree module has been loaded. ** Kernel log: [2.380064] iTCO_wdt: initialized. heartbeat=30 sec (nowayout=0) [2.382153] FAT-fs (sda1): utf8 is not a recommended IO charset for FAT filesystems, filesystem will be case sensitive! [2.389263] iwlwifi :03:00.0: Detected Intel(R) Dual Band Wireless AC 7260, REV=0x144 [2.389337] iwlwifi :03:00.0: L1 Enabled - LTR Enabled [2.389597] iwlwifi :03:00.0: L1 Enabled - LTR Enabled [2.399262] Adding 2103996k swap on /dev/sda2. Priority:-1 extents:1 across:2103996k SSFS [2.399658] intel_rapl: Found RAPL domain package [2.399661] intel_rapl: Found RAPL domain core [2.399663] intel_rapl: Found RAPL domain uncore [2.399664] intel_rapl: Found RAPL domain dram [2.415702] cfg80211: World regulatory domain updated: [2.415705] cfg80211: DFS Master region: unset [2.415706] cfg80211: (start_freq - end_freq @ bandwidth), (max_antenna_gain, max_eirp), (dfs_cac_time) [2.415707] cfg80211: (2402000 KHz - 2472000 KHz @ 4 KHz), (N/A, 2000 mBm), (N/A) [2.415708] cfg80211: (2457000 KHz - 2482000 KHz @ 4 KHz), (N/A, 2000 mBm), (N/A) [2.415709] cfg80211: (2474000 KHz - 2494000 KHz @ 2 KHz), (N/A, 2000 mBm), (N/A) [2.415710] cfg80211: (517 KHz - 525 KHz @ 8 KHz, 16 KHz AUTO), (N/A, 2000 mBm), (N/A) [2.415711] cfg80211: (525 KHz - 533 KHz @ 8 KHz, 16 KHz AUTO), (N/A, 2000 mBm), (0 s) [2.415712] cfg80211: (549 KHz - 573 KHz @ 16 KHz), (N/A, 2000 mBm), (0 s)
Bug#789951: general protection faults in updatedb.mlocat and nfsd
Package: linux-image-3.16.0-4-amd64 Version: 3.16.7-ckt9-3~deb8u1 NFS server stopped responding I looked in the log and found 4 kernel stack traces, the last one occurred at about the time that the NFS problem was noticed. The final stack is similar to this OpenSUSE stack: https://forums.opensuse.org/showthread.php/505638-general-protection-fault?p=2716891 Jun 21 07:35:30 - kernel: general protection fault: [#3] SMP Jun 21 07:35:30 - kernel: Modules linked in: dm_crypt ecb algif_skcipher af_alg cpufreq_conservative cpufreq_stats cpufreq_powersave cpufreq_userspace binfmt_misc 8021q garp stp mrp llc xfrm_user xfrm4_tunnel tunnel4 ipcomp xfrm_ipcomp esp4 ah4 nfsd auth_rpcgss oid_registry nfs_acl nfs lockd fscache sunrpc deflate ctr twofish_generic twofish_x86_64_3way twofish_x86_64 twofish_common camellia_generic camellia_x86_64 serpent_sse2_x86_64 xts serpent_generic lrw gf128mul glue_helper blowfish_generic blowfish_x86_64 blowfish_common cast5_generic cast_common ablk_helper cryptd des_generic cbc cmac xcbc rmd160 sha512_generic sha256_generic hmac crypto_null af_key xfrm_algo radeon evdev kvm_amd kvm pcspkr ttm drm_kms_helper drm i2c_algo_bit edac_mce_amd k10temp edac_core sp5100_tco i2c_piix4 i2c_core tpm_infineon tpm_tis Jun 21 07:35:30 - kernel: tpm button shpchp acpi_cpufreq processor thermal_sys loop fuse parport_pc ppdev lp parport autofs4 ext4 crc16 mbcache jbd2 crc32c_generic btrfs xor raid6_pq dm_mod raid1 hid_generic usbhid hid md_mod sg sd_mod crc_t10dif crct10dif_generic crct10dif_common usb_storage ata_generic ohci_pci tg3 ptp pps_core pata_atiixp libphy mpt2sas raid_class scsi_transport_sas ehci_pci ohci_hcd ehci_hcd ahci libahci libata scsi_mod usbcore usb_common Jun 21 07:35:30 - kernel: CPU: 1 PID: 1866 Comm: updatedb.mlocat Tainted: G D 3.16.0-4-amd64 #1 Debian 3.16.7-ckt9-2 Jun 21 07:35:30 - kernel: Hardware name: HP ProLiant MicroServer, BIOS O41 01/17/2011 Jun 21 07:35:30 - kernel: task: 8802139b75b0 ti: 8801017cc000 task.ti: 8801017cc000 Jun 21 07:35:30 - kernel: RIP: 0010:[811bfac0] [811bfac0] __d_lookup_rcu+0x80/0x160 Jun 21 07:35:30 - kernel: RSP: 0018:8801017cfcb0 EFLAGS: 00010282 Jun 21 07:35:30 - kernel: RAX: bbffdea7 RBX: a1f0ebbaf7044d9d RCX: 000c Jun 21 07:35:30 - kernel: RDX: c901 RSI: 8801017cfe10 RDI: 880101ec7198 Jun 21 07:35:30 - kernel: RBP: 880101ec7198 R08: 400812aad2579179 R09: 88012861f020 Jun 21 07:35:30 - kernel: R10: 0009 R11: R12: a1f0ebbaf7044d95 Jun 21 07:35:30 - kernel: R13: 0009c08012aa R14: 8801017cfe10 R15: 8801017cfd0c Jun 21 07:35:30 - kernel: FS: 7fa7d3f8a700() GS:88021fc8() knlGS: Jun 21 07:35:30 - kernel: CS: 0010 DS: ES: CR0: 80050033 Jun 21 07:35:30 - kernel: CR2: 7f21ebb84a08 CR3: 000124ca9000 CR4: 07e0 Jun 21 07:35:30 - kernel: Stack: Jun 21 07:35:30 - kernel: 8802139b75b0 001a 8802139b75b0 8801017cfd78 Jun 21 07:35:30 - kernel: 8801017cfd68 0040 8800d6f30be0 880101ec7198 Jun 21 07:35:30 - kernel: 8801017cfe00 811b1b7e ff9c 009c25d0 Jun 21 07:35:30 - kernel: Call Trace: Jun 21 07:35:30 - kernel: [811b1b7e] ? lookup_fast+0x3e/0x2b0 Jun 21 07:35:30 - kernel: [811b3635] ? path_lookupat+0x155/0x780 Jun 21 07:35:30 - kernel: [811b3c86] ? filename_lookup+0x26/0xc0 Jun 21 07:35:30 - kernel: [811b7d74] ? user_path_at_empty+0x54/0x90 Jun 21 07:35:30 - kernel: [810eb46e] ? from_kgid_munged+0xe/0x20 Jun 21 07:35:30 - kernel: [811accaa] ? cp_new_stat+0x13a/0x160 Jun 21 07:35:30 - kernel: [811ac856] ? vfs_fstatat+0x46/0x90 Jun 21 07:35:30 - kernel: [811acd2d] ? SYSC_newlstat+0x1d/0x40 Jun 21 07:35:30 - kernel: [811bc47d] ? SyS_poll+0x5d/0xf0 Jun 21 07:35:30 - kernel: [81510e4d] ? system_call_fast_compare_end+0x10/0x15 Jun 21 07:35:30 - kernel: Code: 00 00 00 4c 89 e8 49 c7 c3 ff ff ff ff 48 c1 e8 20 49 89 c2 eb 11 0f 1f 44 00 00 48 8b 1b 48 85 db 0f 84 84 00 00 00 4c 8d 63 f8 8b 43 fc 48 39 6b 10 75 e7 48 83 7b 08 00 74 e0 83 e0 fe f6 45 Jun 21 07:35:30 - kernel: RIP [811bfac0] __d_lookup_rcu+0x80/0x160 Jun 21 07:35:30 - kernel: RSP 8801017cfcb0 Jun 21 07:35:30 - kernel: ---[ end trace 333bfa01d4718bda ]--- Jun 23 07:35:07 - kernel: general protection fault: [#4] SMP Jun 23 07:35:07 - kernel: Modules linked in: dm_crypt ecb algif_skcipher af_alg cpufreq_conservative cpufreq_stats cpufreq_powersave cpufreq_userspace binfmt_misc 8021q garp stp mrp llc xfrm_user xfrm4_tunnel tunnel4 ipcomp xfrm_ipcomp esp4 ah4 nfsd auth_rpcgss oid_registry nfs_acl nfs lockd fscache sunrpc deflate ctr twofish_generic twofish_x86_64_3way twofish_x86_64 twofish_common camellia_generic camellia_x86_64 serpent_sse2_x86_64 xts
Bug#789913: linux-image-3.16.0-4-amd64: Oopses in nouveau_gpuobj_create
Package: src:linux Version: 3.16.7-ckt11-1 Severity: normal Dear Maintainer, -- Package-specific info: ** Version: Linux version 3.16.0-4-amd64 (debian-kernel@lists.debian.org) (gcc version 4.8.4 (Debian 4.8.4-1) ) #1 SMP Debian 3.16.7-ckt11-1 (2015-05-24) ** Command line: BOOT_IMAGE=/vmlinuz-3.16.0-4-amd64 root=/dev/mapper/helix-root ro quiet cgroup_enable=memory ** Tainted: CI (3072) * Module from drivers/staging has been loaded. * Working around severe firmware bug. ** Kernel log: [5.552781] sound hdaudioC0D0:mono: mono_out=0x0 [5.552782] sound hdaudioC0D0:inputs: [5.552784] sound hdaudioC0D0: Front Mic=0x19 [5.552786] sound hdaudioC0D0: Rear Mic=0x18 [5.552788] sound hdaudioC0D0: Line=0x1a [5.561583] input: HDA Digital PCBeep as /devices/pci:00/:00:1b.0/sound/card0/hdaudioC0D0/input5 [5.562315] input: HDA Intel Front Mic as /devices/pci:00/:00:1b.0/sound/card0/input6 [5.562352] input: HDA Intel Rear Mic as /devices/pci:00/:00:1b.0/sound/card0/input7 [5.562384] input: HDA Intel Line as /devices/pci:00/:00:1b.0/sound/card0/input8 [5.562418] input: HDA Intel Line Out as /devices/pci:00/:00:1b.0/sound/card0/input9 [5.562450] input: HDA Intel Front Headphone as /devices/pci:00/:00:1b.0/sound/card0/input10 [5.703551] [drm] Initialized drm 1.1.0 20060810 [5.803819] ppdev: user-space parallel port driver [5.822169] systemd-udevd[433]: renamed network interface eth1 to rename3 [5.857870] nouveau [ DEVICE][:0f:00.0] BOOT0 : 0x0a8c00b1 [5.857874] nouveau [ DEVICE][:0f:00.0] Chipset: GT218 (NVA8) [5.857876] nouveau [ DEVICE][:0f:00.0] Family : NV50 [5.857923] nouveau [ VBIOS][:0f:00.0] checking PRAMIN for image... [5.927265] nouveau [ VBIOS][:0f:00.0] ... appears to be valid [5.927267] nouveau [ VBIOS][:0f:00.0] using image from PRAMIN [5.927436] nouveau [ VBIOS][:0f:00.0] BIT signature found [5.927439] nouveau [ VBIOS][:0f:00.0] version 70.18.89.00.02 [5.927849] nouveau :0f:00.0: irq 76 for MSI/MSI-X [5.927859] nouveau [ PMC][:0f:00.0] MSI interrupts enabled [5.927890] nouveau [ PFB][:0f:00.0] RAM type: DDR3 [5.927892] nouveau [ PFB][:0f:00.0] RAM size: 512 MiB [5.927893] nouveau [ PFB][:0f:00.0]ZCOMP: 960 tags [5.931440] nouveau [VOLT][:0f:00.0] GPU voltage: 90uv [5.958909] nouveau [ PTHERM][:0f:00.0] FAN control: none / external [5.958916] nouveau [ PTHERM][:0f:00.0] fan management: automatic [5.958919] nouveau [ PTHERM][:0f:00.0] internal sensor: yes [5.958941] nouveau [ CLK][:0f:00.0] 03: core 135 MHz shader 270 MHz memory 135 MHz [5.958945] nouveau [ CLK][:0f:00.0] 07: core 405 MHz shader 810 MHz memory 405 MHz [5.958948] nouveau [ CLK][:0f:00.0] 0f: core 520 MHz shader 1230 MHz memory 790 MHz [5.958967] nouveau [ CLK][:0f:00.0] --: core 405 MHz shader 810 MHz memory 405 MHz [5.959235] [TTM] Zone kernel: Available graphics memory: 12360808 kiB [5.959237] [TTM] Zone dma32: Available graphics memory: 2097152 kiB [5.959238] [TTM] Initializing pool allocator [5.959244] [TTM] Initializing DMA pool allocator [5.959254] nouveau [ DRM] VRAM: 512 MiB [5.959256] nouveau [ DRM] GART: 1048576 MiB [5.959258] nouveau [ DRM] TMDS table version 2.0 [5.959260] nouveau [ DRM] DCB version 4.0 [5.959262] nouveau [ DRM] DCB outp 00: 02000360 [5.959263] nouveau [ DRM] DCB outp 01: 02000362 00020010 [5.959265] nouveau [ DRM] DCB outp 02: 028003a6 0f220010 [5.959267] nouveau [ DRM] DCB outp 03: 01011380 [5.959268] nouveau [ DRM] DCB outp 04: 08011382 00020010 [5.959270] nouveau [ DRM] DCB outp 05: 088113c6 0f220010 [5.959271] nouveau [ DRM] DCB conn 00: 00101064 [5.959273] nouveau [ DRM] DCB conn 01: 00202165 [5.985148] SSE version of gcm_enc/dec engaged. [5.989371] alg: No test for __gcm-aes-aesni (__driver-gcm-aes-aesni) [6.007495] [drm] Supports vblank timestamp caching Rev 2 (21.10.2013). [6.007497] [drm] Driver supports precise vblank timestamp query. [6.059339] nouveau [ DRM] MM: using COPY for buffer copies [6.059716] systemd-udevd[427]: renamed network interface eth0 to eth1 [6.063648] iTCO_vendor_support: vendor-support=0 [6.089288] systemd-udevd[433]: renamed network interface eth1 to eth0 [6.140078] iTCO_wdt: Intel TCO WatchDog Timer Driver v1.11 [6.140105] iTCO_wdt: unable to reset NO_REBOOT flag, device disabled by hardware/BIOS [6.161172] nouveau [ DRM] allocated 1600x900 fb: 0x7, bo 880302534c00 [6.161460] fbcon: nouveaufb (fb0) is primary device [6.248913] Console: switching to colour frame buffer device 200x56 [
