Performance regression in firmware-iwlwifi 0.38~bpo70+1
A message that is not in the archive (sent off-list or not yet received): 20130709153104.20718ebb@thinkpad
Quoted 8 lines · Sergey Frolov
Hello.
I would like to report a performance regression in the firmware-iwlwifi 0.38~bpo70+1 package.
After installing it and reloading iwlwifi with a new firmware iwlwifi-6000g2a-6.ucode, my adapter "Intel Corporation Centrino Advanced-N 6205 [Taylor Peak] (rev 34)" can manage only about 200 kb/s DL speed tops regardless of power saving options.
This package has version 18.168.6.1 of this firmware.
Quoted 4 lines
Reverting to 0.36+wheezy.1 and reloading module with iwlwifi-6000g2a-5.ucode resolves the issue. The AP is located in the same room, it's an old 2.4Ghz 11g ASUS WL-320gE access point, configured with wpa-psk encryption.
This has version 17.168.5.3.
Quoted 1 line
I am using linux-image-3.9-0.bpo.1-amd64 kernel from backports.
This is based on Linux 3.9.6.
Quoted 2 lines
If you need any more info, please CC me, as I am not subscribed to this list.
Do you recognise this bug? Can you suggest any diagnostic information Sergey can provide?
Ben.
Signature · 2 lines
Ben Hutchings It is easier to write an incorrect program than to understand a correct one.
PiBPbiBUdWUsIDIwMTMtMDctMDkgYXQgMTU6MzEgKzAzMDAsIFNlcmdleSBGcm9sb3Ygd3JvdGU6 DQo+ID4gSGVsbG8uDQo+ID4NCj4gPiBJIHdvdWxkIGxpa2UgdG8gcmVwb3J0IGEgcGVyZm9ybWFu Y2UgcmVncmVzc2lvbiBpbiB0aGUNCj4gPiBmaXJtd2FyZS1pd2x3aWZpIDAuMzh+YnBvNzArMSBw YWNrYWdlLg0KPiA+IEFmdGVyIGluc3RhbGxpbmcgaXQgYW5kIHJlbG9hZGluZyBpd2x3aWZpIHdp dGggYSBuZXcgZmlybXdhcmUNCj4gPiBpd2x3aWZpLTYwMDBnMmEtNi51Y29kZSwgbXkgYWRhcHRl ciAiSW50ZWwgQ29ycG9yYXRpb24gQ2VudHJpbm8NCj4gPiBBZHZhbmNlZC1OIDYyMDUgW1RheWxv ciBQZWFrXSAocmV2IDM0KSIgY2FuIG1hbmFnZSBvbmx5IGFib3V0IDIwMCBrYi9zDQo+ID4gREwg c3BlZWQgdG9wcyByZWdhcmRsZXNzIG9mIHBvd2VyIHNhdmluZyBvcHRpb25zLg0KPiANCj4gVGhp cyBwYWNrYWdlIGhhcyB2ZXJzaW9uIDE4LjE2OC42LjEgb2YgdGhpcyBmaXJtd2FyZS4NCj4gDQo+ ID4gUmV2ZXJ0aW5nIHRvIDAuMzYrd2hlZXp5LjEgYW5kIHJlbG9hZGluZyBtb2R1bGUgd2l0aA0K PiA+IGl3bHdpZmktNjAwMGcyYS01LnVjb2RlIHJlc29sdmVzIHRoZSBpc3N1ZS4gVGhlIEFQIGlz IGxvY2F0ZWQgaW4gdGhlDQo+ID4gc2FtZSByb29tLCBpdCdzIGFuIG9sZCAyLjRHaHogMTFnIEFT VVMgV0wtMzIwZ0UgYWNjZXNzIHBvaW50LA0KPiA+IGNvbmZpZ3VyZWQgd2l0aCB3cGEtcHNrIGVu Y3J5cHRpb24uDQo+IA0KPiBUaGlzIGhhcyB2ZXJzaW9uIDE3LjE2OC41LjMuDQo+IA0KPiA+IEkg YW0gdXNpbmcgbGludXgtaW1hZ2UtMy45LTAuYnBvLjEtYW1kNjQga2VybmVsIGZyb20gYmFja3Bv cnRzLg0KPiANCj4gVGhpcyBpcyBiYXNlZCBvbiBMaW51eCAzLjkuNi4NCj4NCg0KSnVzdCB0byBt YWtlIHN1cmUgSSBwcm9wZXJseSB1bmRlcnN0YW5kIHdoYXQgeW91IGFyZSBzYXlpbmcgaGVyZTog b25seSB0aGUgZmlybXdhcmUgaXMgY2hhbmdlZD8gTm90IHRoZSBkcml2ZXI/DQo=
Quoted 1 line · Emmanuel
Just to make sure I properly understand what you are saying here: only the firmware is changed? Not the driver?
BTW, in my Thinkpad x21e (Intel Wireless-N 1000) I've noticed a very high (~80%+) CPU utilization by the "irq/42-iwlwifi" kernel process with the 3.9.x series kernel (self-built from upstream GIT sources) and 3.10. Not sure in what exact 3.9 version this behaviour started. I'm using the standard 0.36+wheezy.1 (non-free) version of firmware-iwlwifi.
Quoted 9 lines
Quoted 1 line
Just to make sure I properly understand what you are saying here: only the
firmware is changed? Not the driver?
BTW, in my Thinkpad x21e (Intel Wireless-N 1000) I've noticed a very high (~80%+) CPU utilization by the "irq/42-iwlwifi" kernel process with the 3.9.x series kernel (self-built from upstream GIT sources) and 3.10. Not sure in what exact 3.9 version this behaviour started. I'm using the standard 0.36+wheezy.1 (non-free) version of firmware-iwlwifi.
For CPU utilization:
https://git.kernel.org/cgit/linux/kernel/git/iwlwifi/iwlwifi.git/commit/?id=01911dab97cb3e21e640aaca82689acec00ed848
https://git.kernel.org/cgit/linux/kernel/git/iwlwifi/iwlwifi.git/commit/?id=68972c46f2975d3d61f9dc9f311f77bfc8a8b12b