Skip to contentSkip to conversation

x86/boot/e820: Re-enable fallback if e820 table is empty

PATCH Superseded David Gow 2 participants 5 messages Apr 15 – Apr 16 x86

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore

David Gow

author
#0 lore raw

PATCH 1/1 x86/boot/e820: Re-enable fallback if e820 table is empty

In commit 157266edcc56 ("x86/boot/e820: Simplify append_e820_table() and remove restriction on single-entry tables"), the check that the number of entries in the e820 table was removed. The
intention was to support single-entry maps, but by removing the check
entirely, we also skip the fallback (to, e.g., the BIOS 88h function).

This means that if no E820 map is passed in from the bootloader (which is the case on some bootloaders, like linld), we end up with an empty memory map, and the kernel fails to boot (either by deadlocking on OOM, or by failing to allocate the real mode trampoline, or similar).

Re-instate the check in append_e820_table(), but only check nr_entries is non-zero. This allows e820__memory_setup_default() to fall back to other memory size sources, and doesn't affect e820__memory_setup_extended(), as the latter ignores the return value from append_e820_table().

Tested against linld, and the kernel boots again fine.

Fixes 157266edcc56 ("x86/boot/e820: Simplify append_e820_table() and remove restriction on single-entry tables")
---
 arch/x86/kernel/e820.c | 4 ++++
 1 file changed, 4 insertions(+)
arch/x86/kernel/e820.c +4 -0
index 2a9992758933..7dfd4bdc95fb 100644
@@ -450,6 +450,10 @@ __init static int append_e820_table(struct boot_e820_entry *entries, u32 nr_entr
 {
 	struct boot_e820_entry *entry = entries;
 
+	/* If there aren't any entries, we'll want to fall-back to another source. */
+	if (!nr_entries)
+		return -1;
+
 	while (nr_entries) {
 		u64 start = entry->addr;
 		u64 size  = entry->size;
Signature · 1 line
2.53.0
#1 lore raw
Quoted 1 line · David Gow

In commit 157266edcc56 ("x86/boot/e820: Simplify append_e820_table() and remove restriction on single-entry tables"),

Note, it's fine to wrap this line in the commit message, it's only requirement to tags to be "1 tag — 1 line".

Quoted 17 lines

the check that the number of entries in the e820 table was removed. The intention was to support single-entry maps, but by removing the check entirely, we also skip the fallback (to, e.g., the BIOS 88h function).

This means that if no E820 map is passed in from the bootloader (which is the case on some bootloaders, like linld), we end up with an empty memory map, and the kernel fails to boot (either by deadlocking on OOM, or by failing to allocate the real mode trampoline, or similar).

Re-instate the check in append_e820_table(), but only check nr_entries is non-zero. This allows e820__memory_setup_default() to fall back to other memory size sources, and doesn't affect e820__memory_setup_extended(), as the latter ignores the return value from append_e820_table().

Tested against linld, and the kernel boots again fine.

Fixes 157266edcc56 ("x86/boot/e820: Simplify append_e820_table() and remove restriction on single-entry tables")

...

arch/x86/kernel/e820.c __init static int append_e820_table(struct boot_e820_entry *entries, u32 nr_entr · line 455

Quoted 5 lines
arch/x86/kernel/e820.c +3 -0
 	struct boot_e820_entry *entry = entries;
 
+	/* If there aren't any entries, we'll want to fall-back to another source. */
+	if (!nr_entries)
+		return -1;

Can we return -ENODATA or -ENOENT here?

Signature · 2 lines
With Best Regards,
Andy Shevchenko
author
#2 lore raw

Le 15/04/2026 à 4:51 PM, Andy Shevchenko a écrit :

Quoted 6 lines · David Gow
Quoted 1 line

In commit 157266edcc56 ("x86/boot/e820: Simplify append_e820_table() and remove restriction on single-entry tables"),

Note, it's fine to wrap this line in the commit message, it's only requirement to tags to be "1 tag — 1 line".

Thanks -- I vaguely recall checkpatch complaining about this at some point, though I could've been imagining it...

Quoted 20 lines
Quoted 17 lines

the check that the number of entries in the e820 table was removed. The intention was to support single-entry maps, but by removing the check entirely, we also skip the fallback (to, e.g., the BIOS 88h function).

This means that if no E820 map is passed in from the bootloader (which is the case on some bootloaders, like linld), we end up with an empty memory map, and the kernel fails to boot (either by deadlocking on OOM, or by failing to allocate the real mode trampoline, or similar).

Re-instate the check in append_e820_table(), but only check nr_entries is non-zero. This allows e820__memory_setup_default() to fall back to other memory size sources, and doesn't affect e820__memory_setup_extended(), as the latter ignores the return value from append_e820_table().

Tested against linld, and the kernel boots again fine.

Fixes 157266edcc56 ("x86/boot/e820: Simplify append_e820_table() and remove restriction on single-entry tables")

...

Quoted 8 lines

arch/x86/kernel/e820.c __init static int append_e820_table(struct boot_e820_entry *entries, u32 nr_entr · line 455

Quoted 5 lines
  	struct boot_e820_entry *entry = entries;
arch/x86/kernel/e820.c +3 -0
+	/* If there aren't any entries, we'll want to fall-back to another source. */
+	if (!nr_entries)
+		return -1;

Can we return -ENODATA or -ENOENT here?

Sounds like a good idea. The previous code (like this) used -1 for both "there are no entires" and "there are entries which look bogus". Maybe we should replace the latter with -EINVAL, too.

-- David

#3 lore raw
Quoted 2 lines · David Gow

Le 15/04/2026 à 4:51 PM, Andy Shevchenko a écrit :

Quoted 1 line · David Gow

...

Quoted 10 lines
Quoted 6 lines

arch/x86/kernel/e820.c __init static int append_e820_table(struct boot_e820_entry *entries, u32 nr_entr · line 455

Quoted 4 lines
  	struct boot_e820_entry *entry = entries;
arch/x86/kernel/e820.c +3 -0
+	/* If there aren't any entries, we'll want to fall-back to another source. */
+	if (!nr_entries)
+		return -1;

Can we return -ENODATA or -ENOENT here?

Sounds like a good idea. The previous code (like this) used -1 for both "there are no entires" and "there are entries which look bogus". Maybe we should replace the latter with -EINVAL, too.

Maybe, but is it related to (coupled with) this change? If not, we can do it later if needed.

Signature · 2 lines
With Best Regards,
Andy Shevchenko
author
#4 lore raw

Le 15/04/2026 à 10:26 PM, Andy Shevchenko a écrit :

Quoted 13 lines · David Gow
Quoted 2 lines

Le 15/04/2026 à 4:51 PM, Andy Shevchenko a écrit :

Quoted 1 line · David Gow

...

Quoted 7 lines
Quoted 6 lines

arch/x86/kernel/e820.c __init static int append_e820_table(struct boot_e820_entry *entries, u32 nr_entr · line 455

Quoted 4 lines
   	struct boot_e820_entry *entry = entries;
arch/x86/kernel/e820.c +3 -0
+	/* If there aren't any entries, we'll want to fall-back to another source. */
+	if (!nr_entries)
+		return -1;

Can we return -ENODATA or -ENOENT here?

Quoted 7 lines
Quoted 3 lines

Sounds like a good idea. The previous code (like this) used -1 for both "there are no entires" and "there are entries which look bogus". Maybe we should replace the latter with -EINVAL, too.

Maybe, but is it related to (coupled with) this change? If not, we can do it later if needed.

I think it's probably related enough: append_e820_table() only has the one other error return. That being said, I don't think it matters too much either way: there's exactly one caller which checks the return value, and it only cares that it's nonzero. We could just replace the whole thing with a bool. (Or, if we wanted to be more brave, with the count of entries added, though that'd complicate the caller a bit if we wanted to preserve the same behaviour.)

Regardless, I've sent out v2 here:
https://lore.kernel.org/all/[email protected]/

Cheers,
-- David