diff for duplicates of <20161028154245.GA19454@kroah.com> diff --git a/a/1.txt b/N1/1.txt index 306be1b..e83746f 100644 --- a/a/1.txt +++ b/N1/1.txt @@ -1,18 +1,18 @@ On Fri, Oct 28, 2016 at 08:36:34AM -0700, Michael Zoran wrote: > On Fri, 2016-10-28 at 11:31 -0400, Greg KH wrote: > > On Fri, Oct 28, 2016 at 08:16:51AM -0700, Michael Zoran wrote: -> > > The conversion to dma_map_sg left a few loose ends.??This change +> > > The conversion to dma_map_sg left a few loose ends. This change > > > ties up those loose ends. > > > > > > 1. Settings the DMA mask is mandatory on 64 bit even though it -> > > is optional on 32 bit.??Set the mask so that DMA space is always +> > > is optional on 32 bit. Set the mask so that DMA space is always > > > in the lower 32 bit range of data on both platforms. > > > > > > 2. The scatterlist was not completely initialized correctly. > > > Initialize the list with sg_init_table so that DMA works correctly > > > if scatterlist debugging is enabled in the build configuration. > > > -> > > 3. The error paths in create_pagelist were not consistent.??Make +> > > 3. The error paths in create_pagelist were not consistent. Make > > > them all consistent by calling a helper function called > > > cleanup_pagelistinfo to cleanup regardless of what state the > > > pagelist @@ -20,7 +20,7 @@ On Fri, Oct 28, 2016 at 08:36:34AM -0700, Michael Zoran wrote: > > > > > > 4. create_pagelist and free_pagelist had a very large amount of > > > duplication in computing offsets into a single allocation of memory -> > > in the DMA area.??Introduce a new structure called the pagelistinfo +> > > in the DMA area. Introduce a new structure called the pagelistinfo > > > that is appened to the end of the allocation to store necessary > > > information to prevent duplication of code and make cleanup on > > > errors @@ -37,13 +37,13 @@ On Fri, Oct 28, 2016 at 08:36:34AM -0700, Michael Zoran wrote: > > > it > > > were submitted piecemeal. > > -> > No, it's harder.??A patch should only do one type of thing, this +> > No, it's harder. A patch should only do one type of thing, this > > patch > > has to be reviewed thinking of 4 different things all at once, making > > it > > much more difficult to do so. > > -> > We write patches to be read easily, and make them easy to review.??We +> > We write patches to be read easily, and make them easy to review. We > > don't write them in a way to be easy for the developer to create :) > > > > Can you please break this up into a patch series? diff --git a/a/content_digest b/N1/content_digest index 855fa3a..c333afa 100644 --- a/a/content_digest +++ b/N1/content_digest @@ -1,27 +1,37 @@ "ref\020161028151651.3430-1-mzoran@crowfest.net\0" "ref\020161028153124.GA29906@kroah.com\0" "ref\01477668994.12378.1.camel@crowfest.net\0" - "From\0gregkh@linuxfoundation.org (Greg KH)\0" - "Subject\0[PATCH] staging: vc04_services: tie up loose ends with dma_map_sg conversion\0" + "From\0Greg KH <gregkh@linuxfoundation.org>\0" + "Subject\0Re: [PATCH] staging: vc04_services: tie up loose ends with dma_map_sg conversion\0" "Date\0Fri, 28 Oct 2016 11:42:45 -0400\0" - "To\0linux-arm-kernel@lists.infradead.org\0" + "To\0Michael Zoran <mzoran@crowfest.net>\0" + "Cc\0devel@driverdev.osuosl.org" + daniels@collabora.com + swarren@wwwdotorg.org + lee@kernel.org + linux-kernel@vger.kernel.org + eric@anholt.net + noralf@tronnes.org + linux-rpi-kernel@lists.infradead.org + popcornmix@gmail.com + " linux-arm-kernel@lists.infradead.org\0" "\00:1\0" "b\0" "On Fri, Oct 28, 2016 at 08:36:34AM -0700, Michael Zoran wrote:\n" "> On Fri, 2016-10-28 at 11:31 -0400, Greg KH wrote:\n" "> > On Fri, Oct 28, 2016 at 08:16:51AM -0700, Michael Zoran wrote:\n" - "> > > The conversion to dma_map_sg left a few loose ends.??This change\n" + "> > > The conversion to dma_map_sg left a few loose ends.\302\240\302\240This change\n" "> > > ties up those loose ends.\n" "> > > \n" "> > > 1. Settings the DMA mask is mandatory on 64 bit even though it\n" - "> > > is optional on 32 bit.??Set the mask so that DMA space is always\n" + "> > > is optional on 32 bit.\302\240\302\240Set the mask so that DMA space is always\n" "> > > in the lower 32 bit range of data on both platforms.\n" "> > > \n" "> > > 2. The scatterlist was not completely initialized correctly.\n" "> > > Initialize the list with sg_init_table so that DMA works correctly\n" "> > > if scatterlist debugging is enabled in the build configuration.\n" "> > > \n" - "> > > 3. The error paths in create_pagelist were not consistent.??Make\n" + "> > > 3. The error paths in create_pagelist were not consistent.\302\240\302\240Make\n" "> > > them all consistent by calling a helper function called\n" "> > > cleanup_pagelistinfo to cleanup regardless of what state the\n" "> > > pagelist\n" @@ -29,7 +39,7 @@ "> > > \n" "> > > 4. create_pagelist and free_pagelist had a very large amount of\n" "> > > duplication in computing offsets into a single allocation of memory\n" - "> > > in the DMA area.??Introduce a new structure called the pagelistinfo\n" + "> > > in the DMA area.\302\240\302\240Introduce a new structure called the pagelistinfo\n" "> > > that is appened to the end of the allocation to store necessary\n" "> > > information to prevent duplication of code and make cleanup on\n" "> > > errors\n" @@ -46,13 +56,13 @@ "> > > it\n" "> > > were submitted piecemeal.\n" "> > \n" - "> > No, it's harder.??A patch should only do one type of thing, this\n" + "> > No, it's harder.\302\240\302\240A patch should only do one type of thing, this\n" "> > patch\n" "> > has to be reviewed thinking of 4 different things all at once, making\n" "> > it\n" "> > much more difficult to do so.\n" "> > \n" - "> > We write patches to be read easily, and make them easy to review.??We\n" + "> > We write patches to be read easily, and make them easy to review.\302\240\302\240We\n" "> > don't write them in a way to be easy for the developer to create :)\n" "> > \n" "> > Can you please break this up into a patch series?\n" @@ -75,4 +85,4 @@ "\n" greg k-h -08ac8bfa432655e8f815e07c25dee2f7ac66de6fd72dfe7b7d0bffb2e475d9aa +8964a2368689028c7eac2703b2f3300bd0d6b9e125a0be52725b65b182dbafb0
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.