| Message ID | 1447273397-22833-1-git-send-email-festevam@gmail.com |
|---|---|
| State | Superseded |
| Headers | show |
Hi Fabio, On 11 November 2015 at 13:23, Fabio Estevam <festevam@gmail.com> wrote: > From: Fabio Estevam <fabio.estevam@freescale.com> > > Commit 5ba534d247d418 ("arm: Switch 32-bit ARM to using generic global_data > setup") causes malloc() to fail in SPL. > > The reason is that the GD_MALLOC_BASE is not passed anymore. > > Restore the code that passes malloc base so that we can have > malloc working in SPL code again. > > Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com> > --- > arch/arm/lib/crt0.S | 4 ++++ > 1 file changed, 4 insertions(+) > > diff --git a/arch/arm/lib/crt0.S b/arch/arm/lib/crt0.S > index 80548eb..d620126 100644 > --- a/arch/arm/lib/crt0.S > +++ b/arch/arm/lib/crt0.S > @@ -87,6 +87,10 @@ ENTRY(_main) > mov sp, r0 > > mov r0, #0 > +#if defined(CONFIG_SYS_MALLOC_F_LEN) > + sub sp, sp, #CONFIG_SYS_MALLOC_F_LEN > + str sp, [r9, #GD_MALLOC_BASE] > +#endif > bl board_init_f > > #if ! defined(CONFIG_SPL_BUILD) > -- > 1.9.1 > Thanks for digging into this. But this should be set up in board_init_f_mem(): #if defined(CONFIG_SYS_MALLOC_F) && \ (!defined(CONFIG_SPL_BUILD) || !defined(CONFIG_SYS_SPL_MALLOC_START)) top -= CONFIG_SYS_MALLOC_F_LEN; gd->malloc_base = top; #endif Is it possible that the #ifdef logic is wrong for your board? Regardds, Simon
Hello Fabio, On Wed, 11 Nov 2015 18:23:17 -0200, Fabio Estevam <festevam@gmail.com> wrote: > From: Fabio Estevam <fabio.estevam@freescale.com> > > Commit 5ba534d247d418 ("arm: Switch 32-bit ARM to using generic global_data > setup") causes malloc() to fail in SPL. > > The reason is that the GD_MALLOC_BASE is not passed anymore. This is the explanation of the SPL fail, but not the /cause/ of it. IOW, why is GD_MALLOC_BASE not passed any more, and what function or routine is it not passed any more to? > Restore the code that passes malloc base so that we can have > malloc working in SPL code again. > > [...] > +#if defined(CONFIG_SYS_MALLOC_F_LEN) > + sub sp, sp, #CONFIG_SYS_MALLOC_F_LEN > + str sp, [r9, #GD_MALLOC_BASE] > +#endif NAK, as this only papers over the actual issue. Board_init_f_mem should have set the malloc base in GD. Therefore, rather than doing it again later, we must determine why it was not properly done earlier.11111 Can you give me the toolchain version, board name and commit ID that I could use to reproduce the *faulty* build and check the generated code? Amicalement,
diff --git a/arch/arm/lib/crt0.S b/arch/arm/lib/crt0.S index 80548eb..d620126 100644 --- a/arch/arm/lib/crt0.S +++ b/arch/arm/lib/crt0.S @@ -87,6 +87,10 @@ ENTRY(_main) mov sp, r0 mov r0, #0 +#if defined(CONFIG_SYS_MALLOC_F_LEN) + sub sp, sp, #CONFIG_SYS_MALLOC_F_LEN + str sp, [r9, #GD_MALLOC_BASE] +#endif bl board_init_f #if ! defined(CONFIG_SPL_BUILD)