Skip to contentSkip to conversation

fdt: fix get_next_memory_node()

PATCH In review Marek Vasut 3 participants 5 messages Sep 9, 2018 – Sep 10, 2018 Library code

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 fdt: fix get_next_memory_node()

The get_next_memory_node() always sets mem to -1 , which is incorrect, because then every iteration of memory bank parsing will start from the first memory bank instead of the previous one.

On systems with 1 memory bank defined in DT and CONFIG_NR_DRAM_BANKS=4 , like ie. r8a77965-salvator-x , this will result in U-Boot incorrectly reporting four identical memory banks with the same memory configuration.

Fix this by setting mem to startoffset value, which restores the behavior before the fixed patch was applied.

Cc Jens Wiklander <[email protected]>
Cc Simon Glass <[email protected]>
Cc Tom Rini <[email protected]>
Fixes 452bc121027d ("fdt: fix fdtdec_setup_memory_banksize()")
---
 lib/fdtdec.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)
lib/fdtdec.c +1 -3
index a208589c48..b2ec12e9b5 100644
@@ -1182,10 +1182,8 @@ int fdtdec_setup_mem_size_base(void)
 
 #if defined(CONFIG_NR_DRAM_BANKS)
 
-static int get_next_memory_node(const void *blob, int startoffset)
+static int get_next_memory_node(const void *blob, int mem)
 {
-	int mem = -1;
-
 	do {
 		mem = fdt_node_offset_by_prop_value(gd->fdt_blob, mem,
 						    "device_type", "memory", 7);
Signature · 1 line
2.18.0
#1 lore raw

PATCH 1/1 fdt: fix get_next_memory_node()

Quoted 8 lines · Marek Vasut

The get_next_memory_node() always sets mem to -1 , which is incorrect, because then every iteration of memory bank parsing will start from the first memory bank instead of the previous one.

On systems with 1 memory bank defined in DT and CONFIG_NR_DRAM_BANKS=4 , like ie. r8a77965-salvator-x , this will result in U-Boot incorrectly reporting four identical memory banks with the same memory configuration.

Quoted 8 lines

Fix this by setting mem to startoffset value, which restores the behavior before the fixed patch was applied.

Cc Jens Wiklander <[email protected]>
Cc Simon Glass <[email protected]>
Cc Tom Rini <[email protected]>
Fixes 452bc121027d ("fdt: fix fdtdec_setup_memory_banksize()")

Before
arch_number = 0x00000000
boot_params = 0x00000000

DRAM bank   = 0x00000000
-> start    = 0x00000000
-> size     = 0x10000000
DRAM bank   = 0x00000001
-> start    = 0x10000000
-> size     = 0x40000000
DRAM bank   = 0x00000002
-> start    = 0x50000000
-> size     = 0x30000000
DRAM bank   = 0x00000003
-> start    = 0x00000000
-> size     = 0x10000000

after
arch_number = 0x00000000
boot_params = 0x00000000

DRAM bank   = 0x00000000
-> start    = 0x00000000
-> size     = 0x10000000
DRAM bank   = 0x00000001
-> start    = 0x10000000
-> size     = 0x40000000
DRAM bank   = 0x00000002
-> start    = 0x50000000
-> size     = 0x30000000

Thanks,
Michal

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: OpenPGP digital signature
URL: <http://lists.denx.de/pipermail/u-boot/attachments/20180910/08b8dd21/attachment.sig>

#2 lore raw

PATCH 1/1 fdt: fix get_next_memory_node()

Quoted 41 lines · Michal Simek
Quoted 8 lines · Marek Vasut

The get_next_memory_node() always sets mem to -1 , which is incorrect, because then every iteration of memory bank parsing will start from the first memory bank instead of the previous one.

On systems with 1 memory bank defined in DT and CONFIG_NR_DRAM_BANKS=4 , like ie. r8a77965-salvator-x , this will result in U-Boot incorrectly reporting four identical memory banks with the same memory configuration.

Quoted 8 lines

Fix this by setting mem to startoffset value, which restores the behavior before the fixed patch was applied.

Cc Jens Wiklander <[email protected]>
Cc Simon Glass <[email protected]>
Cc Tom Rini <[email protected]>
Fixes 452bc121027d ("fdt: fix fdtdec_setup_memory_banksize()")

Before
arch_number = 0x00000000
boot_params = 0x00000000

DRAM bank   = 0x00000000
-> start    = 0x00000000
-> size     = 0x10000000
DRAM bank   = 0x00000001
-> start    = 0x10000000
-> size     = 0x40000000
DRAM bank   = 0x00000002
-> start    = 0x50000000
-> size     = 0x30000000
DRAM bank   = 0x00000003
-> start    = 0x00000000
-> size     = 0x10000000

after
arch_number = 0x00000000
boot_params = 0x00000000

Quoted 9 lines
DRAM bank   = 0x00000000
-> start    = 0x00000000
-> size     = 0x10000000
DRAM bank   = 0x00000001
-> start    = 0x10000000
-> size     = 0x40000000
DRAM bank   = 0x00000002
-> start    = 0x50000000
-> size     = 0x30000000

That's on a system with three banks, so it was broken before this patch?

Signature · 2 lines
Best regards,
Marek Vasut
#3 lore raw

PATCH 1/1 fdt: fix get_next_memory_node()

Quoted 52 lines · Marek Vasut
Quoted 41 lines · Michal Simek
Quoted 8 lines · Marek Vasut

The get_next_memory_node() always sets mem to -1 , which is incorrect, because then every iteration of memory bank parsing will start from the first memory bank instead of the previous one.

On systems with 1 memory bank defined in DT and CONFIG_NR_DRAM_BANKS=4 , like ie. r8a77965-salvator-x , this will result in U-Boot incorrectly reporting four identical memory banks with the same memory configuration.

Quoted 8 lines

Fix this by setting mem to startoffset value, which restores the behavior before the fixed patch was applied.

Cc Jens Wiklander <[email protected]>
Cc Simon Glass <[email protected]>
Cc Tom Rini <[email protected]>
Fixes 452bc121027d ("fdt: fix fdtdec_setup_memory_banksize()")

Before
arch_number = 0x00000000
boot_params = 0x00000000

DRAM bank   = 0x00000000
-> start    = 0x00000000
-> size     = 0x10000000
DRAM bank   = 0x00000001
-> start    = 0x10000000
-> size     = 0x40000000
DRAM bank   = 0x00000002
-> start    = 0x50000000
-> size     = 0x30000000
DRAM bank   = 0x00000003
-> start    = 0x00000000
-> size     = 0x10000000

after
arch_number = 0x00000000
boot_params = 0x00000000

Quoted 9 lines
DRAM bank   = 0x00000000
-> start    = 0x00000000
-> size     = 0x10000000
DRAM bank   = 0x00000001
-> start    = 0x10000000
-> size     = 0x40000000
DRAM bank   = 0x00000002
-> start    = 0x50000000
-> size     = 0x30000000
Quoted 1 line

That's on a system with three banks, so it was broken before this patch?

yes that's on 3 bank system and as is visible above there are 4 banks listed and also all of them were calculated together. That's why system with 2GB of memory had surprisingly 2.3GiB.

Before with my 3 bank setup
DRAM: 2.3 GiB
After
DRAM: 2 GiB

With one bank this is even more tricky because
DRAM: 8 GiB
is reported but relocation works properly.

Definitely this is regression and it should go to this release.

Thanks,
Michal

Signature · 14 lines
Michal Simek, Ing. (M.Eng), OpenPGP -> KeyID: FE3D1F91
w: www.monstr.eu p: +42-0-721842854
Maintainer of Linux kernel - Xilinx Microblaze
Maintainer of Linux kernel - Xilinx Zynq ARM and ZynqMP ARM64 SoCs
U-Boot custodian - Xilinx Microblaze/Zynq/ZynqMP SoCs


-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: OpenPGP digital signature
URL: <http://lists.denx.de/pipermail/u-boot/attachments/20180910/f337ae88/attachment.sig>
#4 lore raw
In reply to Marek Vasut
Quoted 11 lines · Marek Vasut

The get_next_memory_node() always sets mem to -1 , which is incorrect, because then every iteration of memory bank parsing will start from the first memory bank instead of the previous one.

On systems with 1 memory bank defined in DT and CONFIG_NR_DRAM_BANKS=4 , like ie. r8a77965-salvator-x , this will result in U-Boot incorrectly reporting four identical memory banks with the same memory configuration.

Fix this by setting mem to startoffset value, which restores the behavior before the fixed patch was applied.

Quoted 6 lines
Cc Jens Wiklander <[email protected]>
Cc Simon Glass <[email protected]>
Cc Tom Rini <[email protected]>
Fixes 452bc121027d ("fdt: fix fdtdec_setup_memory_banksize()")

Applied to u-boot/master, thanks!

Signature · 8 lines
Tom
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: not available
URL: <http://lists.denx.de/pipermail/u-boot/attachments/20180910/aba7698e/attachment.sig>