[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-devel] [PATCH] fix acpi regression
From: |
Wen Congyang |
Subject: |
Re: [Qemu-devel] [PATCH] fix acpi regression |
Date: |
Tue, 12 Apr 2011 17:08:18 +0800 |
User-agent: |
Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.9.1.9) Gecko/20100413 Fedora/3.0.4-2.fc13 Thunderbird/3.0.4 |
At 04/12/2011 04:48 PM, Isaku Yamahata Write:
> On Tue, Apr 12, 2011 at 01:13:15PM +0800, Wen Congyang wrote:
>> This bug is introduced by commit 23910d3f.
>
> Oh, Thanks. Good catch. I agree with the first hunk.
> But what is the second hunk for? The GPE STS register is W1C.
> (NULL check and 0xff masking is okay. it's a bit redundant, though)
I found this bug when I hotplug a nic.
I do not know about ACPI.
The second hunk is the same before commit 23910d3f.
I will read ACPI spec to confirm it and update this patch.
Thanks
>
>>From ACPI4 spec 4.7.4.1.1.
> can only be cleared by software writing a "1" to its respective bit position.
>
> thanks,
>
>>
>> Signed-off-by: Wen Congyang <address@hidden>
>>
>> ---
>> hw/acpi.c | 10 +++-------
>> 1 files changed, 3 insertions(+), 7 deletions(-)
>>
>> diff --git a/hw/acpi.c b/hw/acpi.c
>> index e372474..790bd3b 100644
>> --- a/hw/acpi.c
>> +++ b/hw/acpi.c
>> @@ -355,7 +355,7 @@ static uint8_t *acpi_gpe_ioport_get_ptr(ACPIGPE *gpe,
>> uint32_t addr)
>> if (addr < gpe->len / 2) {
>> cur = gpe->sts + addr;
>> } else if (addr < gpe->len) {
>> - cur = gpe->en + addr;
>> + cur = gpe->en + addr - gpe->len / 2;
>> } else {
>> abort();
>> }
>> @@ -369,12 +369,8 @@ void acpi_gpe_ioport_writeb(ACPIGPE *gpe, uint32_t
>> addr, uint32_t val)
>>
>> addr -= gpe->blk;
>> cur = acpi_gpe_ioport_get_ptr(gpe, addr);
>> - if (addr < gpe->len / 2) {
>> - /* GPE_STS */
>> - *cur = (*cur) & ~val;
>> - } else if (addr < gpe->len) {
>> - /* GPE_EN */
>> - *cur = val;
>> + if (cur != NULL) {
>> + *cur = val & 0xff;
>> } else {
>> abort();
>> }
>> --
>> 1.7.1
>>