On Fri, 16 May 2014 17:04:32 +0100 Ian Campbell <i...@hellion.org.uk> wrote:
> On Fri, 2014-05-16 at 18:05 +0300, Siarhei Siamashka wrote: > > The FEL SPL entry point at the address 0x2000 starts execution > > in the ARM mode. So we need to make sure that this particular > > code is also compiled for the ARM mode. The rest of the SPL can > > be compiled as Thumb2 and benefit from the code size reduction. > > What is the saving like? For A13-OLinuXinoM_FEL_config and gcc version 4.8.2 (Gentoo 4.8.2-r1 p1.4-ssptest, pie-0.5.9-ssptest) == before == $ armv7a-hardfloat-linux-gnueabi-size u-boot-spl text data bss dec hex filename 11010 372 16 11398 2c86 u-boot-spl == after == $ armv7a-hardfloat-linux-gnueabi-size u-boot-spl text data bss dec hex filename 8762 372 16 9150 23be u-boot-spl Having extra SRAM space headroom is nice. Also the unification and the use of Thumb2 code everywhere means that any possible compiler bugs are likely to affect both regular and FEL builds in the same way. Should I send a v2 patch with an updated commit message? -- Best regards, Siarhei Siamashka -- You received this message because you are subscribed to the Google Groups "linux-sunxi" group. To unsubscribe from this group and stop receiving emails from it, send an email to linux-sunxi+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.