2002-06-23 20:39:45

by Tigran Aivazian

Subject: [patch-2.5.24] microcode tidy-up + observation

Dear Linus,

First of all, I attached a small patch, tested under 2.5.24:

o changes KERN_ERR priority for harmless messages to KERN_INFO as
they are not really errors and shouldn't pollute the console.
(microcode.c)

o if a write(2) with too large a buffer happens (such as would cause
BUG() in vmalloc()) we should return -ENOMEM rather than -EINVAL and
that silently (currently the driver is almost encouraging absurd lengths)
(microcode.c)

o old reference to my webpage removed and another one updated.
(CREDITS/MAINTAINERS)

Observation: I noticed that all smp_num_cpus in the driver have been
replaced by NR_CPUS. I assume this was done for the CPU hotplug support,
although if smp_num_cpus reflects the "current" number of online cpus, I
think, the driver would have worked as is and avoided wasting 64k-epsilon
of memory.

Regards
Tigran

diff -X dontdiff -urN linux-2.5.24-orig/CREDITS linux-2.5.24/CREDITS
--- linux-2.5.24-orig/CREDITS Thu Jun 20 23:53:46 2002
+++ linux-2.5.24/CREDITS Sun Jun 23 21:11:22 2002
@@ -39,7 +39,7 @@

N: Tigran A. Aivazian
E: [email protected]
-W: http://www.ocston.org/~tigran
+W: http://www.moses.uklinux.net/patches
D: BFS filesystem
D: Intel IA32 CPU microcode update support
D: Various kernel patches
diff -X dontdiff -urN linux-2.5.24-orig/MAINTAINERS linux-2.5.24/MAINTAINERS
--- linux-2.5.24-orig/MAINTAINERS Thu Jun 20 23:53:54 2002
+++ linux-2.5.24/MAINTAINERS Sun Jun 23 21:10:28 2002
@@ -255,7 +255,6 @@
P: Tigran A. Aivazian
M: [email protected]
L: [email protected]
-W: http://www.ocston.org/~tigran/patches/bfs
S: Maintained

BLOCK LAYER
diff -X dontdiff -urN linux-2.5.24-orig/arch/i386/kernel/microcode.c linux-2.5.24/arch/i386/kernel/microcode.c
--- linux-2.5.24-orig/arch/i386/kernel/microcode.c Thu Jun 20 23:53:47 2002
+++ linux-2.5.24/arch/i386/kernel/microcode.c Sun Jun 23 21:17:50 2002
@@ -260,7 +260,7 @@

if (microcode[i].rev < rev) {
spin_unlock_irqrestore(&microcode_update_lock, flags);
- printk(KERN_ERR
+ printk(KERN_INFO
"microcode: CPU%d not 'upgrading' to earlier revision"
" %d (current=%d)\n", cpu_num, microcode[i].rev, rev);
return;
@@ -268,7 +268,7 @@
/* notify the caller of success on this cpu */
req->err = 0;
spin_unlock_irqrestore(&microcode_update_lock, flags);
- printk(KERN_ERR
+ printk(KERN_INFO
"microcode: CPU%d already at revision"
" %d (current=%d)\n", cpu_num, microcode[i].rev, rev);
return;
@@ -337,10 +337,9 @@
sizeof(struct microcode));
return -EINVAL;
}
- if ((len >> PAGE_SHIFT) > num_physpages) {
- printk(KERN_ERR "microcode: too much data (max %ld pages)\n", num_physpages);
- return -EINVAL;
- }
+ if ((len >> PAGE_SHIFT) > num_physpages)
+ return -ENOMEM;
+
down_write(&microcode_rwsem);
if (!mc_applied) {
mc_applied = kmalloc(NR_CPUS*sizeof(struct microcode),

2002-06-23 22:55:44

by Tigran Aivazian

Subject: Re: [patch-2.5.24] microcode tidy-up + observation

On Sun, 23 Jun 2002, Tigran Aivazian wrote:
> Observation: I noticed that all smp_num_cpus in the driver have been
> replaced by NR_CPUS. I assume this was done for the CPU hotplug support,
> although if smp_num_cpus reflects the "current" number of online cpus, I
> think, the driver would have worked as is and avoided wasting 64k-epsilon
> of memory.

Ok, I can see now that smp_num_cpus doesn't exist anymore on i386 but is
still there on other architectures.