On 3/28/2024 11:33 AM, Andrew Jones wrote:
> On Thu, Mar 28, 2024 at 10:21:10AM -0500, Pavan Kumar Paluri wrote:
>> In some cases, KUT guest might fail to exit boot services due to a
>> possible memory map update that might have taken place between
>> efi_get_memory_map() and efi_exit_boot_services() calls. As per UEFI
>> spec 2.10 (Section 7.4.6 EFI_BOOT_SERVICES.ExitBootServices()), we need
>> to keep trying to update the memory map and calls to exit boot
>> services as long as case status is EFI_INVALID_PARAMETER. Keep freeing
>> the old memory map before obtaining new memory map via
>> efi_get_memory_map() in case of exit boot services failure.
>>
>> Signed-off-by: Pavan Kumar Paluri <papaluri@xxxxxxx>
>> ---
>> lib/efi.c | 34 ++++++++++++++++++++--------------
>> 1 file changed, 20 insertions(+), 14 deletions(-)
>>
>> diff --git a/lib/efi.c b/lib/efi.c
>> index 8a74a22834a4..d2569b22b4f2 100644
>> --- a/lib/efi.c
>> +++ b/lib/efi.c
>> @@ -406,8 +406,8 @@ efi_status_t efi_main(efi_handle_t handle, efi_system_table_t *sys_tab)
>> efi_system_table = sys_tab;
>>
>> /* Memory map struct values */
>> - efi_memory_desc_t *map = NULL;
>> - unsigned long map_size = 0, desc_size = 0, key = 0, buff_size = 0;
>> + efi_memory_desc_t *map;
>> + unsigned long map_size, desc_size, key, buff_size;
>> u32 desc_ver;
>>
>> /* Helper variables needed to get the cmdline */
>> @@ -446,13 +446,6 @@ efi_status_t efi_main(efi_handle_t handle, efi_system_table_t *sys_tab)
>> efi_bootinfo.mem_map.key_ptr = &key;
>> efi_bootinfo.mem_map.buff_size = &buff_size;
>>
>> - /* Get EFI memory map */
>> - status = efi_get_memory_map(&efi_bootinfo.mem_map);
>> - if (status != EFI_SUCCESS) {
>> - printf("Failed to get memory map\n");
>> - goto efi_main_error;
>> - }
>> -
>> #ifdef __riscv
>> status = efi_get_boot_hartid();
>> if (status != EFI_SUCCESS) {
>> @@ -461,11 +454,24 @@ efi_status_t efi_main(efi_handle_t handle, efi_system_table_t *sys_tab)
>> }
>> #endif
>>
>> - /*
>> - * Exit EFI boot services, let kvm-unit-tests take full control of the
>> - * guest
>> - */
>> - status = efi_exit_boot_services(handle, &efi_bootinfo.mem_map);
>> + status = EFI_INVALID_PARAMETER;
>> + while (status == EFI_INVALID_PARAMETER) {
>> + /* Get EFI memory map */
>
> I tried to get rid of this comment since it states the exact same thing
> as the function name below does.
>
Will make the change.
>> + status = efi_get_memory_map(&efi_bootinfo.mem_map);
>> + if (status != EFI_SUCCESS) {
>> + printf("Failed to get memory map\n");
>> + goto efi_main_error;
>> + }
>> + /*
>> + * Exit EFI boot services, let kvm-unit-tests take full
>> + * control of the guest.
>> + */
>> + status = efi_exit_boot_services(handle,
>> + &efi_bootinfo.mem_map);
>
> We have 100 char lines (and that's just a soft limit) so this would look
> better sticking out.
>
Will do.
>> + if (status == EFI_INVALID_PARAMETER)
>> + efi_free_pool(*efi_bootinfo.mem_map.map);
>> + }
>> +
>> if (status != EFI_SUCCESS) {
>> printf("Failed to exit boot services\n");
>> goto efi_main_error;
>> --
>> 2.34.1
>>
>
> Besides the nits,
>
> Reviewed-by: Andrew Jones <andrew.jones@xxxxxxxxx>
>
> (A general comment for the series is that we're on v3 but there's no
> changelog anywhere. Please use cover letters for a series and then
> put the changelog in the cover letter.)
>
Ah, I missed out on that. I will include a cover-letter for v4 and also
plan to drop patches 3 & 4 and send them separately as they aren't very
relevant to UEFI fixes and for easier upstreaming.
Thanks,
Pavan
> Thanks,
> drew
