SLOB breaks Crypto
in [Kernel]
|
Prev: [RELEASE] LTTng 0.214 for kernel 2.6.33.4
Next: vlynq: make whole Kconfig-menu dependant on architecture
From: FUJITA Tomonori on 19 May 2010 08:10 On Wed, 19 May 2010 12:40:36 +0100 David Woodhouse <dwmw2(a)infradead.org> wrote: > On Wed, 2010-05-19 at 13:32 +0200, Geert Uytterhoeven wrote: > > Instead of having (different) defaults in sl[aou]b, perhaps we should > > just remove the defaults completely, to ensure all architectures set > > ARCH_SLAB_MINALIGN to the correct value? > > What is 'correct'? The architecture sets it to the minimum value that it > can cope with, according to its own alignment constraints (and DMA/cache > constraints, in the case of ARCH_KMALLOC_MINALIGN). IIRC, not all the architectures do that; ARCH_KMALLOC_MINALIGN doesn't mean "DMA-safe" alignment currently. -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo(a)vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
From: David Woodhouse on 19 May 2010 08:30 On Wed, 2010-05-19 at 21:02 +0900, FUJITA Tomonori wrote: > On Wed, 19 May 2010 12:40:36 +0100 > David Woodhouse <dwmw2(a)infradead.org> wrote: > > > On Wed, 2010-05-19 at 13:32 +0200, Geert Uytterhoeven wrote: > > > Instead of having (different) defaults in sl[aou]b, perhaps we should > > > just remove the defaults completely, to ensure all architectures set > > > ARCH_SLAB_MINALIGN to the correct value? > > > > What is 'correct'? The architecture sets it to the minimum value that it > > can cope with, according to its own alignment constraints (and DMA/cache > > constraints, in the case of ARCH_KMALLOC_MINALIGN). > > IIRC, not all the architectures do that; ARCH_KMALLOC_MINALIGN doesn't > mean "DMA-safe" alignment currently. Surely those architectures that have alignment constraints for DMA but which don't set ARCH_KMALLOC_MINALIGN are just buggy -- it _does_ mean that. -- David Woodhouse Open Source Technology Centre David.Woodhouse(a)intel.com Intel Corporation -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo(a)vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
From: FUJITA Tomonori on 19 May 2010 08:40 On Wed, 19 May 2010 13:19:45 +0100 David Woodhouse <dwmw2(a)infradead.org> wrote: > On Wed, 2010-05-19 at 21:02 +0900, FUJITA Tomonori wrote: > > On Wed, 19 May 2010 12:40:36 +0100 > > David Woodhouse <dwmw2(a)infradead.org> wrote: > > > > > On Wed, 2010-05-19 at 13:32 +0200, Geert Uytterhoeven wrote: > > > > Instead of having (different) defaults in sl[aou]b, perhaps we should > > > > just remove the defaults completely, to ensure all architectures set > > > > ARCH_SLAB_MINALIGN to the correct value? > > > > > > What is 'correct'? The architecture sets it to the minimum value that it > > > can cope with, according to its own alignment constraints (and DMA/cache > > > constraints, in the case of ARCH_KMALLOC_MINALIGN). > > > > IIRC, not all the architectures do that; ARCH_KMALLOC_MINALIGN doesn't > > mean "DMA-safe" alignment currently. > > Surely those architectures that have alignment constraints for DMA but > which don't set ARCH_KMALLOC_MINALIGN are just buggy -- it _does_ mean > that. Well, I thought so but seems that there isn't such agreement: http://kerneltrap.org/mailarchive/linux-kernel/2010/5/12/4568960 -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo(a)vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
From: David Woodhouse on 19 May 2010 08:50 On Wed, 2010-05-19 at 21:26 +0900, FUJITA Tomonori wrote: > > > Surely those architectures that have alignment constraints for DMA > but > > which don't set ARCH_KMALLOC_MINALIGN are just buggy -- it _does_ > mean > > that. > > Well, I thought so but seems that there isn't such agreement: > > http://kerneltrap.org/mailarchive/linux-kernel/2010/5/12/4568960 How strange. That thread has a fair amount of misinformation from people who should know better. -- dwmw2 -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo(a)vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
From: Pekka Enberg on 19 May 2010 09:00
On Wed, May 19, 2010 at 2:46 PM, Herbert Xu <herbert(a)gondor.apana.org.au> wrote: > On Wed, May 19, 2010 at 12:16:45PM +0100, David Woodhouse wrote: >> >> I don't mind. Feel free to apply them to slab.git, but be aware that >> Herbert wanted to see a patch fixing sparc32 ARCH_SLAB_MINALIGN before >> the crypto one is applied. >> >> Although arguably SLOB was broken on sparc32 even before the crypto >> patch -- so perhaps that fix shouldn't _have_ to go in first? > > Well prior to this crypto on sparc32 did work with SLAB/SLUB. �If > you change it without including the sparc pach, then sparc32 would > be broken regardless of which allocator you used. OK, I'll pick up David's patches and just wait for sparc changes to hit Linus' tree first. Herbert, do I have your ACK for the crypto patches? -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo(a)vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/ |