From mboxrd@z Thu Jan 1 00:00:00 1970 From: Marian Balakowicz Date: Wed, 07 May 2008 10:18:49 +0200 Subject: [U-Boot-Users] [PATCH] [new uImage] Fix build errors when CONFIG_LOGBUFFER and CONFIG_FIT are enabled In-Reply-To: <20080507071336.786642477F@gemini.denx.de> References: <20080507071336.786642477F@gemini.denx.de> Message-ID: <482165E9.1000401@semihalf.com> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: u-boot@lists.denx.de Wolfgang Denk wrote: > In message <20080507070353.7403.39121.stgit@hekate.izotz.org> you wrote: >> Recent modifcations to LOGBUFFER handling code were incorrecly >> introduced to fit_check_kernel() routine during >> "Merge branch 'new-image' of git://www.denx.de/git/u-boot-testing", >> commit 27f33e9f45ef7f9685cbdc65066a1828e85dde4f. >> >> This patch cleans up this merge issue. > > Ummm... but don't you break logbuffer support that way? > >> Signed-off-by: Marian Balakowicz >> --- >> >> common/cmd_bootm.c | 11 ----------- >> common/image.c | 4 ---- >> 2 files changed, 0 insertions(+), 15 deletions(-) >> >> >> diff --git a/common/cmd_bootm.c b/common/cmd_bootm.c >> index 44f6b9f..0d67132 100644 >> --- a/common/cmd_bootm.c >> +++ b/common/cmd_bootm.c >> @@ -433,17 +433,6 @@ static int fit_check_kernel (const void *fit, int os_noffset, int verify) >> } >> show_boot_progress (105); >> >> -#ifdef CONFIG_LOGBUFFER >> -#ifndef CONFIG_ALT_LB_ADDR >> - kbd=gd->bd; >> - /* Prevent initrd from overwriting logbuffer */ >> - if (initrd_high < (kbd->bi_memsize-LOGBUFF_LEN-LOGBUFF_OVERHEAD)) >> - initrd_high = kbd->bi_memsize-LOGBUFF_LEN-LOGBUFF_OVERHEAD; >> - debug ("## Logbuffer at 0x%08lX ", kbd->bi_memsize-LOGBUFF_LEN); >> -#else >> - debug ("## Logbuffer at 0x%08lX ", CONFIG_ALT_LB_ADDR); >> -#endif >> -#endif > > Well, it's easy to remove the code here, but if serves a purpose, so > it cannot be simply removed without adding it somewhere else. Or am I > missing something? This piece of the code was removed earlier with one of the Kumar's patches that added LMB support for booting allocations. Then it got incorrectly merged back in a wrong place, which this patch intends to fix. As to the check itself: - it seem bogus, it does not properly check for initrd and logbuffer areas overlaps - it does not take into account the recent "effective" memory size limitation introduced for ppc I'll post a separate, preliminary patch that will attempt to implement the check in a proper way. Cheers, Marian