Re: 3.3 kernel bug
--- On Mon, 4/2/12, Peter Robinson pbrobin...@gmail.com wrote: On Apr 2, 2012 7:16 PM, Yioryos Asprobounitis mavrot...@yahoo.com wrote: It would appear that I came up an XO-1 3.3 kernel bug (albeit not with an OLPC build yet, but the bug is not there with OLPC 2.6.35 or OLPC 3.1 and identical userspace) Its a regression in fs/namei.c, full_name_hash()/hash_name(), regarding files with 8 character names or multiples of it More info in this thread http://marc.info/?l=linux-kernelm=133244187421796w=2 It is apparently fixed in 3.4-rc1, but I was wondering if it may worth merging in 3.3, which appears to be the next x86 OLPC kernel. I do not really know if it affects anything specific in the XO builds, so just a heads up for now. If it's a general kernel bug fix I would expect it'll get pulled into the 3.3 stable releases Peter I'm afraid did not make into the 3.3.1 stable (from what I can see). This is also true for another bug that I came across (also in relation to the usb wifi XO-1 module) that brakes (!) the flow control when usb-debug is enabled. ( https://lkml.org/lkml/2012/2/2/427 ) This is also fixed in 3.4-rc1 but not present in 3.3.1. Hopefully 3.3.2 :-) ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: 3.3 kernel bug
On Tue, Apr 3, 2012 at 10:00 AM, Yioryos Asprobounitis mavrot...@yahoo.com wrote: --- On Mon, 4/2/12, Peter Robinson pbrobin...@gmail.com wrote: On Apr 2, 2012 7:16 PM, Yioryos Asprobounitis mavrot...@yahoo.com wrote: It would appear that I came up an XO-1 3.3 kernel bug (albeit not with an OLPC build yet, but the bug is not there with OLPC 2.6.35 or OLPC 3.1 and identical userspace) Its a regression in fs/namei.c, full_name_hash()/hash_name(), regarding files with 8 character names or multiples of it More info in this thread http://marc.info/?l=linux-kernelm=133244187421796w=2 It is apparently fixed in 3.4-rc1, but I was wondering if it may worth merging in 3.3, which appears to be the next x86 OLPC kernel. I do not really know if it affects anything specific in the XO builds, so just a heads up for now. If it's a general kernel bug fix I would expect it'll get pulled into the 3.3 stable releases Peter I'm afraid did not make into the 3.3.1 stable (from what I can see). This is also true for another bug that I came across (also in relation to the usb wifi XO-1 module) that brakes (!) the flow control when usb-debug is enabled. ( https://lkml.org/lkml/2012/2/2/427 ) This is also fixed in 3.4-rc1 but not present in 3.3.1. Hopefully 3.3.2 :-) Got a link to the exact email with the patch that fixes that later bug? Peter ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: 3.3 kernel bug
--- On Tue, 4/3/12, Peter Robinson pbrobin...@gmail.com wrote: From: Peter Robinson pbrobin...@gmail.com Subject: Re: 3.3 kernel bug To: Yioryos Asprobounitis mavrot...@yahoo.com Cc: OLPC Devel devel@lists.laptop.org Date: Tuesday, April 3, 2012, 5:04 AM On Tue, Apr 3, 2012 at 10:00 AM, Yioryos Asprobounitis mavrot...@yahoo.com wrote: --- On Mon, 4/2/12, Peter Robinson pbrobin...@gmail.com wrote: On Apr 2, 2012 7:16 PM, Yioryos Asprobounitis mavrot...@yahoo.com wrote: It would appear that I came up an XO-1 3.3 kernel bug (albeit not with an OLPC build yet, but the bug is not there with OLPC 2.6.35 or OLPC 3.1 and identical userspace) Its a regression in fs/namei.c, full_name_hash()/hash_name(), regarding files with 8 character names or multiples of it More info in this thread http://marc.info/?l=linux-kernelm=133244187421796w=2 It is apparently fixed in 3.4-rc1, but I was wondering if it may worth merging in 3.3, which appears to be the next x86 OLPC kernel. I do not really know if it affects anything specific in the XO builds, so just a heads up for now. If it's a general kernel bug fix I would expect it'll get pulled into the 3.3 stable releases Peter I'm afraid did not make into the 3.3.1 stable (from what I can see). This is also true for another bug that I came across (also in relation to the usb wifi XO-1 module) that brakes (!) the flow control when usb-debug is enabled. ( https://lkml.org/lkml/2012/2/2/427 ) This is also fixed in 3.4-rc1 but not present in 3.3.1. Hopefully 3.3.2 :-) Got a link to the exact email with the patch that fixes that later bug? The one in the parenthesis above :-) Peter ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: 3.3 kernel bug
On Apr 2, 2012 7:16 PM, Yioryos Asprobounitis mavrot...@yahoo.com wrote: It would appear that I came up an XO-1 3.3 kernel bug (albeit not with an OLPC build yet, but the bug is not there with OLPC 2.6.35 or OLPC 3.1 and identical userspace) Its a regression in fs/namei.c, full_name_hash()/hash_name(), regarding files with 8 character names or multiples of it More info in this thread http://marc.info/?l=linux-kernelm=133244187421796w=2 It is apparently fixed in 3.4-rc1, but I was wondering if it may worth merging in 3.3, which appears to be the next x86 OLPC kernel. I do not really know if it affects anything specific in the XO builds, so just a heads up for now. If it's a general kernel bug fix I would expect it'll get pulled into the 3.3 stable releases Peter ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
